0:04

Welcome back friends. This

0:16

is the changelog. I'm Adams Dekoviak

0:19

and this week on the changelog

0:21

we're talking to Nadia Adelayo, founder

0:23

of Storygraph. Nadia started out as

0:25

a one-woman dev and product team

0:27

and has had to adjust and

0:29

maneuver along the way to becoming

0:31

the Amazon free alternative to Goodreads.

0:34

We talk about the importance of

0:36

customer research, the iterative nature of

0:38

customer research and what it takes

0:40

to synthesize and analyze the findings

0:42

to guide product development, the technical

0:45

challenges and learnings she's faced while building

0:47

Storygraph. For example, at several points they

0:49

faced challenges in handling an influx of

0:52

users and had to re-architect the system.

0:54

We also talk about the business model

0:56

of Storygraph and how they generate revenue

0:59

through plus subscriptions and partnerships with publishers

1:01

for book giveaways. A

1:03

huge, massive thank you to our

1:05

friends and our partners at fly.io,

1:08

the home of changelog.com. With Fly

1:10

you can launch your app near

1:12

your users. Fly transforms containers into

1:15

micro VMs that run on their

1:17

hardware in 30 plus regions on

1:19

six continents and Antarctica is

1:21

coming soon. Big wink. Launch

1:24

an app for free at fly.io. What's

1:38

up friends? This episode of the changelog

1:40

is brought to you by our friends

1:42

over at Vercel. And I'm

1:44

here with Lee Robinson, VP of product. Lee,

1:47

I know you know the tagline for

1:49

Vercel, develop, preview, ship, which has been

1:52

perfect, but now there's more after the

1:54

ship process. You have to worry about

1:56

security, observability, and other parts of just

1:59

running an application. What's

2:01

the story there? What's beyond shipping for

2:03

Vercel? Yeah, you know, when I'm

2:05

building my side projects or when I'm building my

2:08

personal site, it often looks like develop preview ship,

2:10

you know, I try out some new features, I

2:12

tried a new framework, I'm just hacking around with

2:14

something on the weekends. Everything looks good. Great. I

2:16

ship it. I'm done. But as we talk to

2:19

more customers, as we've grown as a company, as

2:21

we've added new products, there's a lot

2:23

more to the product portfolio of Vercel

2:25

nowadays to help pass that experience. So

2:27

when you're building larger, more complex products,

2:30

and when you're working with larger teams,

2:32

you want to have more features, more

2:34

functionality. So, tangibly, what that means is

2:36

features like our Vercel firewall product to

2:38

help you be safe and to have

2:41

that layer of security features like our

2:43

logging and observability tools so you can

2:45

understand and observe your application and production,

2:48

understand if there's errors, understand if things

2:50

are running smoothly and get alerted on

2:52

those. And also then really an expansion

2:54

of our integration suite as well too,

2:56

because you might already be using a

2:58

tool like a Datadog or you might

3:00

already be using a tool at the

3:02

end of this software development lifecycle that

3:04

you want to integrate with to continue

3:07

to scale and secure and observe your

3:09

application and we try to fit into

3:11

those as well too. So we've kind

3:13

of continued to bolster and improve the

3:15

last mile of delivery. That

3:17

sounds amazing. So who's using the Vercel platform

3:20

like that? Can you share some names? Yeah,

3:22

I'm thrilled that we have some

3:25

amazing customers like Under Armour, Nintendo,

3:27

Washington Post, Zapier, who use

3:30

Vercel's frontend cloud to not only help

3:32

scale their infrastructure, scale their business and

3:34

their product, but then also enable their

3:36

team of many developers to be able

3:39

to iterate on their products really quickly

3:41

and take their ideas and build the

3:43

next great thing. Very cool.

3:45

With your configuration for over 35 frameworks,

3:48

Vercel's frontend cloud makes it easy for

3:50

any team to deploy their apps. Today

3:52

you can get started with a 14 day free trial

3:55

of Vercel Pro or get a

3:57

customized enterprise demo from their team.

4:00

Visit vercel.com/change log

4:02

pod to get

4:04

started. That's vercel.com/change

4:07

log pod here

4:36

with Nadia Otto IO founder and

4:38

CEO of Storygraph welcome to the

4:40

show. Thanks for having me. We

4:43

are excited to have you not

4:45

just founder not just CEO but

4:47

pretty much solo dev

4:49

right one women dev

4:51

shop building this what I

4:53

will describe as like a Goodreads competitor

4:55

or killer of all the killer. Killer of

4:58

all the killer. She

5:01

wasn't ready for that one. She's like, well, but

5:03

it sounds good. Let's go with it. We didn't

5:05

say alternative. That's the first time alternative.

5:07

That's because we're being polite. Very

5:10

British of me. Very British. This

5:14

is a really cool thing. Before we get into

5:16

the details. I like this idea. This might not

5:18

be your idea, but I like it anyways, which

5:20

is there are cool web

5:23

products that people love.

5:26

And then they have so much success that they

5:28

eventually get purchased by big tech. And

5:31

then over time, people love them less. Let's just

5:33

say that that's a trend of which Goodreads seems

5:35

to be an instance of the trend. And

5:38

I just feel like that's a place where

5:40

entrepreneurs can just like look at

5:42

that situation and then be like, Hmm, what

5:44

if there was something like

5:47

this, but good again. And

5:49

you could like have a bunch of founders

5:51

just go after that like meetup.com. Just throw that

5:54

one out there. But let's talk about you. Let's

5:56

talk about Storygraph. I'm just thinking

5:58

somebody should disrupt meetup.com. That's

6:00

my point there. But is that what you were

6:02

thinking with Goodreads? Like, oh, I used to love

6:04

Goodreads and now I don't anymore. And so let's

6:06

build something. It's actually pretty funny because

6:09

it wasn't like that at all. At

6:11

all? Okay. In fact, okay, maybe at

6:13

all is, you

6:16

know, a bit far stretching it. Okay, maybe

6:18

a little bit like that, but mostly not.

6:20

A little bit. But I

6:22

purposely, for a long

6:24

time, I even thought the notion that

6:27

I was building a Goodreads alternative. Mainly

6:30

because when I first, I

6:32

thought I had used a Goodreads since 2012.

6:35

And I loved it. There was a time where it was my

6:37

favorite app. The thing is, I didn't use

6:40

a lot of, I was very simple with my usage.

6:42

I found the book I was reading, I marked

6:44

it, I added a star rating. I

6:47

rarely even added text to my reviews.

6:49

So for me, Goodreads did everything I

6:51

needed. It wasn't when I

6:53

got into tech, so I went to a software bootcamp

6:55

in end of 2013. So

6:59

as I was becoming a developer and getting more

7:01

immersed in the tech world, I

7:03

did start to notice or

7:05

think, wait a minute, Goodreads

7:07

has not changed once since

7:09

I've been using it. You know, when you

7:11

get into tech and you learn about how

7:13

websites and apps are developed and you notice with

7:16

other products that you use, oh, that's the new

7:18

design. That's the new feature. I would pay

7:20

more attention to release notes, things like that.

7:22

There was never anything with Goodreads. So I

7:24

do remember thinking, it's a bit strange that

7:26

it's not developed at all, especially since it's owned

7:28

by Amazon. I would think that, but I

7:30

wasn't angry about it. I wasn't mad about

7:32

it because it did what I needed it

7:34

to do. And then

7:36

fast forward to 2019

7:40

when I started working on this product

7:42

in earnest, originally, it

7:45

was just one of my side projects

7:48

and it wasn't even a Goodreads alternative.

7:50

It was a Goodreads companion app.

7:53

Oh, so I was hooking into the

7:55

Goodreads API to pull

7:57

out, to basically create a nice dashboard. of

8:00

all of your shelves and you could see

8:02

your percent through of how much you'd

8:04

read of each of the shelves. That's what it was because

8:06

Goodreads didn't do that. So it wasn't even meant to be

8:10

an alternative. But then when I started

8:12

doing customer research, because I was

8:14

having such a blast and I was like,

8:17

I'd love to build something in books full

8:19

time. Let me see if there is a

8:21

need or a pain point. All

8:23

of the pain points I was getting from

8:26

the people I was speaking to were all

8:28

Goodreads related. So now I'm putting together what

8:30

would I need to build and it starts

8:32

looking a little bit originally just a replacement

8:34

of the recommendations portion and

8:37

then eventually a more fully fledged alternative.

8:39

But for the longest time I told

8:41

myself, especially because throughout the

8:43

first year I was genuinely solo, like

8:45

my co-founder Rob hadn't come into the

8:47

picture yet. So I was genuinely solo

8:50

and I thought to myself, don't

8:52

just change the product or product to

8:54

be a Goodreads alternative for several reasons.

8:57

One being it's very easy to get intimidated

8:59

by that, especially when I've got nothing right

9:01

now and it's just me and I'm building

9:04

an alternative to an Amazon product that's

9:06

huge and has dozens of developers, if

9:08

not more. So that's one. Two, it's

9:10

going to end up being very limiting for

9:13

my product vision because I would just end

9:15

up looking at Goodreads and trying to do

9:17

what they did but better, which isn't compelling

9:20

enough. And so I

9:22

would rather, I wanted to look at it as

9:24

I'm building a new product in the book space

9:27

for avid readers or maybe to get

9:29

people into reading and I will follow

9:31

the pain points and see

9:33

what comes out of that. And

9:36

it wasn't until we got to a certain size

9:38

that it made sense to be like, we're a

9:40

Goodreads alternative because at that point it was like

9:42

the core of the product, we had our niche,

9:44

which was the moods and the pace. The

9:47

core of the product had been fleshed out and

9:49

so at this point now we need to advertise

9:51

ourselves as a Goodreads alternative so that people see

9:54

it and go, oh yeah, I don't like Goodreads, I'll

9:56

try this thing. A lot of. direction

10:00

we can go there, especially with the

10:02

alternative section of it. I'm curious about the, and

10:04

I do want to dive into that as well,

10:07

but I'm curious about the

10:09

customer research aspect because in

10:12

the pre-call you mentioned like the skip that part,

10:14

and I know that I'm definitely one of those

10:16

people because I just want to get to the

10:18

thing, right? I want to microwave this product into

10:20

existence and suddenly be hitting product market fit and

10:22

being amazing, right? Like that's what we all want.

10:25

Can we just get to the millions quickly? Jeez.

10:30

And that's not how it works. And

10:32

sometimes I think even with

10:34

like what we've done here with the change over

10:36

the years, we just we've found our way

10:38

by finding what we love about what we want

10:40

to build. And we don't know that until

10:42

we talk to the people that we intend to

10:44

serve, right? Like a product isn't just simply you

10:47

creating something and people loving it. It's you

10:49

understanding what they want and giving them

10:51

what they want. So how did you find out how

10:53

to give them what they want? So I

10:55

was super lucky in that my

10:58

first and only real job in software

11:00

was at Pivotal, Pivotal Labs. And

11:03

so I worked there for a year and a

11:05

half and I met like people who

11:07

were at the top of their

11:09

field in software engineering, but also

11:11

product management and design. So from

11:14

that I'd already picked up the

11:16

importance of customer

11:18

interviews, approaches to do

11:20

customer research, things like that. So I already

11:22

had a face there. And then

11:24

I'd also read, you know, the standard, I'd

11:27

read the Lean Startup. I'd

11:29

read The Mum Test by Robert

11:31

Patrick. So I'd also read a

11:33

few books that helps me be

11:35

very cognizant of the need

11:38

and importance of customer feedback and

11:40

iterating on that, but also the

11:42

pitfalls that people fall into. So

11:45

when I had that first week of

11:47

working on my side project and loving

11:49

it, oh sorry, also

11:51

before that I'd also had experiences of

11:53

when customer research was not done well

11:55

in prior products or I

11:58

ran a business for a year and a bit before.

12:00

before starting this. And

12:02

I saw what happens when

12:05

customer research isn't made

12:07

important or when the whole team

12:09

is not bought into that. And so I remember

12:11

that when I was getting excited about building something

12:13

in the book space, my first thing was stop,

12:16

pause, gather everything you know about customer

12:19

research. This has to be important because

12:21

the worst thing, especially because I didn't

12:23

have a full-time job and I was just living off

12:26

of my savings. I had five years of runway. At

12:28

the point at which I started this, I

12:30

had four years of runway. So it was very vital that I didn't waste my

12:32

time. So from the beginning, I had

12:35

the fight, the urge to just build it and hope that

12:37

everyone would love it. So I immediately

12:39

was like, right, I need to do customer research.

12:41

And the things that I knew then were, I

12:44

need to have a clear hypothesis of

12:47

what I'm doing. I need to

12:49

have a script because

12:51

a script is what stops you from going

12:54

off on tangents, accidentally asking

12:56

leading questions, things like

12:58

that. You know before you go

13:00

into it that if you get the answers to your

13:02

questions, then you'll be able to figure out the

13:05

results of your hypothesis. And

13:07

I knew that I needed to have an effective synthesis

13:09

method for a way to analyze what

13:12

I'd heard. And so my

13:14

first round of research I

13:16

did, the very first round was

13:19

demoing actually the product that I'd already built,

13:22

which since then I very rarely

13:24

do product demos and interviews

13:26

now because I just think it's

13:28

kind of leading, you know, you're showing someone something and

13:30

you're kind of like, would you use this? You know,

13:32

do you like this? And actually it's

13:35

very easy for people to just say, yeah,

13:37

sure, I would. And actually I was, that's

13:39

what basically happened in the first round of

13:41

interviews I did. And through

13:43

my learning of from the

13:45

mom test, I learned that that was people being

13:48

polite. And actually I needed

13:50

to put a product aside and just

13:52

figure out pain points with average readers.

13:54

And so my first round of research was well,

13:58

I think it was just more discovery phase of. are

14:00

that do average readers have any pain

14:02

points? And after I interviewed

14:04

about, I don't know, five to six

14:06

people or so, I was able to

14:08

go through everything they said. I record all

14:11

my interviews, I watch them back, I take out

14:13

snapshots of thoughts, phrases, things like that,

14:15

and group them into themes. And there

14:18

was a key theme of, I

14:20

don't have one place to get consistent,

14:23

high quality recommendations. So

14:25

then I take that piece. And

14:28

I do, I think, what's the next round? And

14:30

the next round is, well, how do people try

14:33

to figure out how people currently find books?

14:35

So what are their current pain points with

14:37

recommendations now? Listen, I do that round. And

14:39

then I basically, I get to the point

14:41

after doing a few rounds of, ultimately,

14:43

people want to find books based on mood, sentiment,

14:46

vibe, and there's no way to do that. And

14:48

it's like, then I keep going until I get

14:50

to the point where I'm like, here's

14:52

a very basic feature set. And

14:54

this basic feature set is people

14:57

can key in the mood that they're looking for.

14:59

And then they get a book recommendation, you know,

15:01

and then I so I built that very basic

15:03

version. And then I let people

15:05

use it, I go back to the people I interviewed, I

15:07

put it in front of them. And

15:09

then a week later, I book in calls with

15:11

them again. And I asked them, like,

15:14

did you log in to this

15:16

app? Tell me more about what I

15:18

just try and ask questions about their behavior, as opposed

15:20

to like, did you like it? Or you know what

15:22

I mean, just really trying to basically treat

15:24

it quite scientifically as much as

15:26

possible. And so it's just

15:28

through this like iterative process of always having a

15:31

hypothesis or a key question that I'm trying to

15:33

answer, interviewing about five to

15:35

eight people, synthesizing that breaking down

15:37

their learning and just using that to guide

15:39

the next step and just continually doing that.

15:41

And I continue to do that to this

15:43

day, is what has a

15:47

big factor of our success in that

15:49

we are building what people want. Those

15:53

questions you answered or asked seemed very open

15:55

ended versus closed. Like, do you like it

15:58

is yes or no? You

16:00

can just get a yes or no or

16:02

some version of politeness, right? But

16:04

if you say explain to me how you feel

16:06

when you do X You

16:09

get an explanation not a

16:11

yes or no, which even

16:13

with a demo I wondered if you felt like

16:15

you were almost kind of selling it right because

16:18

you're demoing it it's almost like would

16:20

you use this that's kind of damaging

16:22

to the process because you're They

16:24

might feel like you're you're done and you're

16:27

looking for users not feedback and let you

16:29

know Unless that's actually what you asked for

16:31

but the open-ended questions makes a lot of sense

16:33

because you're gonna get a lot more from that

16:36

versus simply a Yes, or

16:38

no closed-ended kind of question. Yeah,

16:40

and so yeah, I did I after that first interview

16:42

with the demo I stopped doing demos and

16:45

I would say that the questions are The

16:47

power of having them be more open-ended so they

16:50

might say I might say I would say they're

16:52

kind of in the middle ground Between yes and

16:54

no and completely open ending like how do you

16:56

feel? I guess it might be like how

16:58

would you describe this product or what's

17:00

your biggest pain point or and they

17:02

could say anything I think that's why

17:04

when you do your synthesis and you

17:06

break down How did each person answer

17:08

this question and despite them saying different

17:11

things if you can say

17:13

wow these three people Ultimately said the same

17:15

thing. That's a key insight because the question

17:18

was open and yet they all

17:20

Express a pain point with recommendations or

17:22

you know what I mean? And that's

17:24

why you gravitate towards those things

17:26

that a handful of people have also the

17:29

same thing despite the question being very open

17:32

with Questions like those small

17:34

sample sizes right five to eight people

17:36

per round so

17:38

it sounds like since I was thinking like how do

17:40

you synthesize open-ended questions because I Have

17:43

a similar situation not on a product

17:45

but on a survey we do for

17:47

front-end feud We're trying to get open-ended

17:49

questions. This is like family feud So

17:52

the question is like, you know, how do you feel

17:54

about AI? Well, we

17:56

want to be able to categorize those and actually

17:58

synthesize them down into groups And we

18:00

can't put multiple choice because that ruins the whole game.

18:02

Like you have to guess what the audience was going

18:04

to say. And so it's a text input, but I'm

18:07

trying to group by in my

18:09

SQL query eventually, you know. And

18:11

that's difficult. And so I've learned tricks

18:13

like say like in a word, how do you feel about

18:16

this? And now they're going to give me one word and

18:18

I can actually work with that better. But

18:20

this is a sample size of 100 to 200 people.

18:22

And with five to eight, I think you can just

18:24

read them. And like in your brain, be like, these

18:27

are actually saying the same thing. So you don't have

18:29

to actually do any data normalization. But I wonder how

18:31

you handle biases

18:33

in participants. Is it

18:35

always somebody new? I mean, because if you're

18:37

asking the same person, they're the easiest person

18:39

to go back to, right? Because they already

18:41

know the process and maybe they provide a

18:43

good feedback. How do you handle that? Because

18:45

small, especially when you're starting, there's very few

18:47

users or it's all your friends. This

18:49

is the hard part with questions like would you use this

18:51

because your friends want to have your back and be like,

18:53

I would, but that might not be true. How

18:56

do you deal with that? That's actually a key

18:58

part of the process, which I neglected

19:00

to mention. Part of

19:02

when you're setting up a round is

19:04

also deciding who are you interviewing? What

19:06

qualifies them? The first round

19:08

were just friends, people I knew. But

19:10

to mix it up a bit, I

19:12

also called the end book

19:15

bloggers from Instagram just

19:17

to have some strangers and things

19:20

like that. So I did that. The earliest

19:22

rounds were a mix of friends

19:25

and strangers from the book community. So

19:28

that was kind of adding some variety. Plus

19:30

maybe a little grassroots marketing, right? You get

19:32

some enthusiasts to know about the storygraph through

19:34

that. Yes. And then I had people obviously

19:36

following the account as well, which I was just building

19:39

up. And again, no product. I was just posting my

19:41

own book reviews of what I was reading. And

19:43

then DMing people saying, I'm actually working

19:46

on a product. There

19:48

was a lot of that in the early days, lots

19:50

of Twitter and Instagram DM. So

19:52

that's one. But then I also

19:54

had this landing page up where I

19:56

said, we're building a tool

19:58

for avid readers. I put

20:01

a subtitle being like, if

20:03

you read 52 books a

20:05

year and you're interested in this, put your

20:07

email here. The reason why I said that is

20:09

because I thought the longest I'd want to wait

20:11

for feedback on anything, like a recommendation, is a

20:13

week. If I'm getting someone who's reading 52

20:15

books a year, if I recommend a

20:18

book to them, they'll hopefully read

20:20

it in a week and then I can interview

20:22

them or just check in, send them a survey

20:24

like, was this actually a good recommendation to see

20:26

if the product was working? Since

20:29

we've been bigger, depending on the research

20:31

round, I'll have different criteria for who

20:33

I want to interview. For

20:36

example, there was one point where I felt stuck

20:38

in product development. I wasn't sure what to focus on.

20:40

I was seeing now and again people tweeting

20:43

and saying things like, oh,

20:45

I love Storygraph. I use it every day. We

20:47

had not thought the app to be used

20:49

every day. I was like, what

20:51

is it about these people who use it every day? What

20:53

is Storygraph doing for them? What

20:56

are their behaviors that make them use it every

20:58

day? Can we use that to learn

21:00

about what are the most compelling parts of the app? I

21:03

just did a tweet and Instagram

21:05

posts saying, we're looking for people who log

21:07

into the app at least once a day.

21:09

That was that. At

21:12

the end of 2022, I was doing a big redesign. Now

21:19

I need to make sure that the redesign

21:21

works across desktop and mobile. Also,

21:24

I needed to make sure that

21:26

it handled all of our common UX

21:28

pain points that made people drop off

21:30

or struggle to use it, but also

21:32

didn't upset the people who were happily

21:34

using it and had a great time.

21:37

I actually had six cohorts.

21:41

The matrix was desktop and mobile. Then

21:44

it was predominantly desktop, predominantly

21:46

mobile. Then it was

21:48

use it very happy, no issues. Use

21:51

it but struggle. Stop choosing

21:53

it because I couldn't figure out

21:55

the UX. I had six groups

21:57

and I sent out a Google form. I

22:00

put that everywhere. I put that on Instagram. I put that

22:02

on Twitter. I put that in my newsletter. Then

22:05

when I was sorting out my research rounds,

22:07

I would just filter that and be like,

22:09

right, I'm trying to test the mobile design

22:11

now. So mobile, blah, blah, blah, blah. We

22:13

made sure that basically by the time we

22:15

implemented the design, it had

22:17

passed the research rounds for all

22:19

those categories. So a key,

22:21

key part of customer research is talking

22:24

to the right people. And

22:26

also another thing is at the beginning of

22:28

my script, I always start

22:30

by saying, I'm doing a

22:33

customer research round. And for this round, I'm

22:35

speaking to insert identity here, because it always

22:37

gives the person a chance to go, that's

22:39

not me. So in case for any reason

22:41

they click the wrong box

22:44

or whatever, it's just like, no.

22:47

I sometimes get people reply to

22:49

my asks and say, I

22:51

don't use the app every day, but I would love to

22:53

talk with you for half an hour. And

22:56

I'm like, yeah, I'm sure you would. No. Because

23:01

I'm really stressed, like, this is who you

23:03

must be. I haven't actually had a

23:05

case with someone yet booking

23:08

and say, Oh, I'm not that person. But I

23:10

want to talk to you. But I have had

23:12

people email to check to be

23:14

like, I'm not that person. But can I and

23:16

I'm like, no, that's very not useful to me.

23:18

And I will end an interview. I have actually

23:20

I think I have ended an interview maybe once

23:23

or twice where it's kind of been like, Oh, I don't think this will

23:25

be valuable for either of us. So

23:28

I rarely interview the same person

23:30

twice unless they're part of a

23:32

specific I had specific very structured

23:34

alpha and beta rounds. But beyond

23:36

that, it's always different people. You're

23:40

doing all this yourself? Mostly more

23:43

recently, I have Abby who works

23:45

for us part time, I

23:48

kind of taught her or showed

23:50

her how I did my synthesis and kind

23:52

of pass that on to her.

23:54

So she has been helping me with a synthesis

23:56

style, but I still do all the interviews. And

23:59

I'm probably at the point where,

24:01

especially because it's a script, I will probably get to the

24:03

point where I'll be able to hand that off

24:05

to Abby, but there's no need right now. And I think

24:08

as I'm the Dev, it's working very

24:10

well for us, but I'm so close to our

24:12

customers and so close to the research. And

24:15

so close to the social media and so

24:17

close to everything.

24:20

You're very close. Yeah,

24:22

I run the Instagram and Twitter. And the

24:24

funniest comments I get are when

24:26

I see, I get comments like

24:28

in the app or the Play Store reviews, I

24:31

will be people like, the founder is so

24:33

nice. Sometimes she comes on Instagram and does

24:35

an Instagram story. And it's always me. It's

24:37

just that sometimes I'll do a video and

24:40

I'll be like, Hi, everyone, this is what's

24:42

happening. But it's always me. The

24:45

other funny thing in the Dev side is

24:47

people who assume I'm

24:49

like just a social media person

24:52

and they'll say something and they'll say, pass

24:55

this onto your Dev team. Or but they'll talk

24:57

at me and they'll often be a Dev themselves

25:00

and they'll try and explain something to me

25:02

as in this person probably will not understand

25:04

what I'm saying. So let me explain it

25:07

very clearly and slowly so that then they

25:09

can pass it on to the Dev team.

25:12

Or they'll say, can you pass me on to

25:14

someone who can explain this thing and

25:16

I'll be like, oh, I can explain it because I

25:18

built it. So like, go ahead. And

25:21

they're always like very surprised. Well,

25:25

it's surprising to have one person do all that. That's for

25:27

sure. Oh, it is. Yeah, for

25:29

sure. Yeah. How long you been

25:31

doing it and how often do

25:33

you do it and tell us a little bit

25:36

about the growth story because this started as a

25:38

side project. Very seems like very successful growth so

25:40

far. I mean, many of us

25:42

sell a Dev probably envious of the traction

25:44

that you've gotten because it's enviable. Tell the

25:47

folks how it's been going. Yeah.

25:49

So I started this. It

25:51

was 2019 was that year where it

25:54

was a side project. So I entered

25:56

January of 2019 with four years of

25:58

runway. And I'd always been

26:01

entrepreneurial and always wanted to do my

26:03

own thing. And I was coming off the back

26:05

of two failed co-founder partnerships.

26:08

So I was like, I'm just gonna go by myself.

26:11

I don't need anybody. I have runway.

26:13

I don't need to get a job. I'm

26:15

just gonna build stuff. And so

26:17

I just started working on that. I explained how I

26:20

got to the point where I was like, right, this

26:22

is interesting to me. That was the founder product fit

26:24

side. So I got the founder product fit very

26:26

early. And that's what made me properly say,

26:28

let me do customer research properly. Let me

26:30

start up an Instagram account. Let me see

26:32

what I can do here. So

26:35

after three months of just customer research. So after

26:37

I first showed that demo to people and I

26:40

thought, no, you can't be showing a demo to people. You

26:42

do this properly. I put the product away. I

26:44

spent three months doing customer research. That was just

26:47

talking to people. And I got to the point

26:49

where I said, okay, I have enough to build

26:51

an alpha. And the alpha

26:53

was literally just a personal recommendation service.

26:56

And so it was a whole

26:58

basic Rails app, but all

27:00

it was was really sending emails to me. And

27:02

I would like send emails back essentially. So someone

27:04

would shut out a form and it'd be like,

27:06

your recommendation will come soon. I'd get an email

27:08

saying, this person needs a book that fits this,

27:11

this, this. I'd spend however long it

27:13

took to try to find a unique find. And then

27:15

I would do something in Rails

27:17

admin or on the console or

27:19

something to just generate an email

27:22

being like, your recommendation is ready.

27:26

No, actually I think I had a backend in the app where

27:28

I would put it in and then it would send an email

27:30

to them to check for their recommendation.

27:33

So that was that. And then I did

27:35

that for about a month or two. And

27:37

then I felt like I'd exhausted all of

27:39

my learnings from the alpha. And

27:41

here's the thing, remember how I said, I was

27:43

looking for people who read a book

27:46

a week because then they would read the

27:48

recommendation. Well, what happened? I check in with

27:50

the people and they'd say, the

27:53

recommendation you gave me looks great. Unfortunately,

27:55

I've got to finish the book I'm reading right

27:57

now. And then I've got the stack. bedside

28:00

table, and that's when the

28:02

pain point transitioned from I

28:05

need a recommendation to how do I know what to

28:07

pick up next? Because actually I do have a lot

28:09

of choice and I want to make sure that I'm

28:12

always picking up the next best book for me. And

28:14

that's when I realized, okay, I need to build

28:16

a more fully fleshed out app because

28:19

just a recommendation service on its own is

28:21

not going to stand because people, you know,

28:24

they get a recommendation but they're not going to read it for months. And

28:27

really they've got they need help getting through

28:29

that pile that's been sitting by their bed

28:32

for months and months and months. So

28:34

that's when I spent a couple months

28:36

just heads down building a beta, still

28:38

just me completely one, like

28:40

solo. And at the

28:43

same time that I said I'm going to

28:45

start building a beta, I started a newsletter

28:47

because my worry was I'm going to spend

28:49

two months building a more fully fleshed out

28:51

product. And I'm not going to talk to

28:53

anyone. I'm going to lose all the momentum

28:55

from all the people I've spoken to over

28:57

the last nine months. And I

28:59

also don't want to build something and then have nobody ready

29:01

to try it when it's ready. So I

29:03

started a newsletter and it was 100 people

29:06

who I'd basically interviewed or

29:08

had heard of the product or something like that. And

29:11

I sent I just started sending a weekly it

29:13

was like a weekly as one

29:15

of my friends said, it's like it's your weekly

29:17

stand up and you are just like sending

29:20

out the email because it's just you doing stand

29:22

up. And I was like, yeah, it's holding me

29:24

accountable. But it was the healing of momentum. That's

29:26

just what I'm doing this week. And also it

29:28

was a much it was a bit like, I

29:30

have to send this email every Monday, I can't

29:32

get to next Monday and have done nothing. So

29:35

it was a very, it was it really kept

29:37

me accountable to just keep doing stuff each week.

29:39

So I, so I had something interesting to say.

29:42

So two months launch the beta. And

29:44

that was just general like sharing it

29:46

on Twitter, things like that. We got

29:48

to 100 users at the beginning of

29:50

2020. Then I was talking about it more.

29:54

Some books to grab has got started talking about it

29:57

more. When I say books to grammars, it's the term

29:59

of like Bloggers on

30:01

Instagram, it's called Bookstagram. Bookstagram.

30:05

Yeah. And so Bookstagrammers

30:07

would talk about it. And we got to a thousand users

30:09

in June of 2020. And also,

30:11

we're in the midst of a pandemic. There's a

30:13

whole story there. I live alone as well. So

30:16

there's a whole piece of going through this product,

30:18

growing and growing, and just in

30:21

the midst of a pandemic, stuck in my flat,

30:23

just doing this thing. Actually, you know

30:25

what? It was actually, obviously,

30:28

the pandemic was not great. So it was

30:30

that we had a pandemic. But it

30:33

worked out really well for me in terms

30:35

of that I didn't have to, you know, I

30:38

listened to a lot of how I built this. And

30:40

you always hear stories of like, people saying, I became

30:42

such a bad friend. I missed

30:44

weddings and missed parties. I canceled late all

30:46

the time. And that

30:49

really, even though I'd never had any sort

30:51

of success like that, it really freaked me

30:53

out because I pride myself in being super

30:55

reliable. I'm always there. I'm always on time.

30:57

And the thought of, would I ever have

30:59

a product that meant that I would like,

31:02

cancel on people or be flaky or whatever?

31:04

I hope not. But we did

31:06

all of our exploding in 2020. Like,

31:09

some of the biggest part, lots of it

31:11

was in 2020. There was nobody

31:13

to cancel on. No like being

31:15

out and pay-to-do duties going off. I didn't

31:18

have to deal with that, which was nice.

31:20

I didn't have to let anybody down

31:23

or anything like that. So what happened

31:25

was mid 2020, there were some viral

31:27

tweets. And it

31:29

was funny because the people who

31:31

tweeted were just, they weren't anybody

31:34

like famous or well known. They

31:36

literally had about 200 followers,

31:38

but the tweet just like hit,

31:40

the tweets just hit. They just resonated

31:43

with a bunch of people because

31:45

essentially they were calling out a bunch of

31:47

factors. The main tweet that kicked

31:49

everything off, it essentially like hit

31:51

on a key point that was very salient at

31:53

that time. So one was it,

31:55

you know, appealed to the book Twitter, book Twitter community.

31:58

So it was calling out the book Twitter community. And

32:01

it basically spoke about, we

32:03

all hate Goodreads, so annoying. It

32:06

called out the fact that, so it said

32:08

there's a great independent alternative. It

32:10

referenced the fact that it was run by

32:13

a black woman, which this was like a week

32:15

after George Floyd's murder. So at this time, everyone,

32:17

there was this big push to

32:19

support black creators. So

32:22

that was mentioned and that was

32:25

jumped on. And then also it

32:27

was mentioned the fact that Goodreads owned Amazon.

32:29

And at the time in the pandemic, there

32:31

was a lot of anti-Amazon sentiment because of,

32:33

I think Amazon's profits had been

32:36

released or something and everyone was struggling in a

32:38

pandemic and it was like, and look, Jeff

32:40

Bezos is still rich and getting

32:42

richer. And in general in publishing,

32:44

there was a lot of anti-Amazon sentiment because they

32:46

have a monopoly on huge swathes of

32:48

the industry. And so this one

32:50

tweet, which had all of this packed in, just

32:53

exploded. And we had thousands of likes

32:55

and retweets. And I'm minding

32:57

my business, cracking

32:59

on with some code. And I

33:01

see these emails coming in because back then

33:04

I got an email every

33:06

time somebody signed up and kicked

33:08

off a Goodreads import. So at

33:10

the time, the maximum we'd ever

33:12

had was eight. And I'm

33:14

seeing like eight emails, dozens of emails coming

33:17

in. What's going on? Going to Twitter, seeing

33:19

what's going on. This system was

33:21

not made to handle like

33:23

more than about eight at a time.

33:26

And it was actually with over

33:28

three days, we got tens

33:30

of thousands. And the

33:32

other thing was that the email said,

33:35

your Goodreads import is underway. But

33:38

to 20,000 people, this was a lie. 20,000

33:40

people got an email saying your Goodreads import was

33:43

underway. It was not underway. It was an excellent

33:45

sidekick queue. And this was the

33:47

first time I had to deal with a sidekick queue that

33:49

had... This queue never had anything in it.

33:51

And now it had... Lots of stuff in it. 18,

33:53

20,000 things in it. Oh,

33:56

so we also have to stop the

33:58

imports while... point my co-founder Rob

34:01

had joined me and we had to

34:03

stop the import because we needed

34:05

to re-architect everything. He was

34:07

managing some servers on his side, so he

34:09

needed to re-architect all of that. I just

34:11

needed to sort out my code because it

34:14

was not ready to handle lots of stuff.

34:16

It just wasn't performing. I can't even remember

34:18

some of the stuff I was doing, but

34:20

some of the loops, some of the... It

34:22

just was not ready for the scale. That

34:25

was that. It took two weeks to get

34:27

the app back up and running again. From

34:30

that time, we just kept on growing

34:32

by at least a thousand users each

34:34

day sign up. Since

34:36

that time, so that was 2020, from

34:39

middle of 2020 to now,

34:42

essentially, it's been steady growth with

34:44

the odd spike when some social

34:46

media, especially a book influencer, makes

34:49

a YouTube video about us or

34:51

a TikTok pops off. Then

34:54

we've had the odd media. We had

34:56

an article in the New Statesman, which

34:58

is a newspaper over here about the

35:00

headline was, Why Good Reasons Bad for

35:02

Books? We were

35:04

mentioned as an up-and-coming alternative that brought

35:06

in thousands of users. We've

35:09

just been steadily growing since then. Behind the

35:11

scenes, I'm still sticking with my customer

35:13

research-focused approach. Love it.

35:15

It's intense. Very intense.

35:17

The queue would have just... It

35:20

had crushed me. You just said something that reminded

35:22

me of a point I wanted to hit on. When

35:24

you asked me the last question, you said

35:26

I was in an endiable position. Actually, there's

35:29

a key part of this story. I

35:32

always felt was entrepreneurial and

35:34

I always wanted to have something successful. I

35:37

remember following all the indie

35:39

developer forums and things, I

35:41

was like, I need to start a B2B

35:43

company because that's how I'm going to make

35:45

money fast. I can start charging from day

35:47

one. It suits not getting

35:49

venture capital. There was a time in my

35:51

past where I was all about... When I

35:53

was younger, I was all about VC,

35:56

venture capital. Then over the years,

35:58

reading more about it. and

36:00

seeing certain stories and you know, just understanding

36:02

more about the world, realizing that's not the

36:04

type of company I wanted to build. So

36:07

I was like B2D is the way. And

36:10

so I always thought

36:12

like I sell into a B2C company

36:14

because a freemium B2C company used by

36:16

millions of users is not what you

36:19

want as a solo dev who

36:21

doesn't want funding. It's just not what you want.

36:24

And I remember when we had that first

36:26

initial strike, when we

36:29

had those tweets, we had all these tens of thousands of

36:31

users trying to use it. I felt

36:33

stuck for a while because I didn't know

36:36

what to do. I just didn't

36:38

know what to do with my code. And

36:40

I was worried. And I remember, you

36:42

know, I had to take a step away

36:45

from my computer. I remember going into my

36:47

dark bathroom and like sitting in there and

36:49

literally saying, I thought I wanted

36:51

this. I don't want this. And

36:54

also saying to myself, is this

36:56

it? Are you going to die

36:58

now? Because you actually got the

37:00

users you wanted and you can't handle it. So

37:03

now the product dies. And also

37:05

I always say that I never actually said this,

37:07

but I feel like it was on the tip

37:09

of my tongue to just be like, I can't

37:11

do this. But I wouldn't allow myself to actually

37:13

say it. So I remember saying, nope,

37:16

that's not you. You're not going to say that. You're going to

37:18

get out of this bathroom and you're going to go back to

37:20

your desk and you're going to fix it. But for

37:22

a while, there was a period of time where I

37:25

thought, I don't want this. This is like awful.

37:30

How did I end up in this

37:32

situation where I'm building a company

37:34

that's not making me any money so

37:36

I can't survive like this. I

37:38

don't want to get VC funding. And I felt

37:40

like I was trapped in it. And we are

37:42

comfortable now, which is great. But for years,

37:45

especially as it was growing, I was like, I feel like

37:48

I'm stuck. What am I going to do?

37:50

Because my money is going down. And

37:54

the costs are going up. Right. What's

38:17

up friends? I'm here with

38:19

one of my new friends, Zane Hamilton from

38:21

CIQ. So Zane, we're coming up on a

38:23

hard deadline with the CentOS end of life

38:26

later this year in July. There

38:28

are still folks out there considering what

38:30

their next move should be. Then last

38:32

year we had a bunch of change

38:34

around Red Hat Enterprise Linux that makes

38:36

it quote less open source in the

38:38

eyes of the community with many saying,

38:40

Rello is open source but where is

38:42

the source and why can't I download

38:44

and install it? Now Rocket Linux is

38:46

fully open source and CIQ is

38:48

a founding support partner that

38:51

offers paid support for migration,

38:53

installation, configuration, training, etc. But

38:56

what exactly does an enterprise or a

38:58

Linux to 7 get when they choose

39:00

the free and open source Rocket Linux

39:02

and then ultimately the support from CIQ

39:04

if they need it? There's

39:06

a lot going on in the enterprise Linux space today.

39:09

There's a lot of end of life of CentOS. People

39:11

are making decisions on where to go next. The standard

39:13

of what enterprise Linux looks like tomorrow is kind of

39:15

up in the air. What CIQ is

39:17

doing is we're trying to help those people

39:19

that are going through these different decisions that they're

39:21

having to make and how they go about making

39:24

those decisions. And that's where our expertise really comes

39:26

into play. A lot of people who have been

39:28

through very complex Linux migrations, be it from the

39:30

old days of migrating from AI X or Solaris

39:33

onto Linux and even going from version to version

39:35

because to be honest enterprise Linux version to version

39:37

is not always been an easy conversion. It hasn't

39:39

been and you will hear that from us. Typically

39:41

the best idea is to do an in place

39:44

upgrade, not always a real easy thing to do.

39:47

But what we've done is we have started looking at

39:49

and securing a path of how can we actually go

39:51

through that? How can we help a customer who's moving

39:53

from CentOS 7 because of the end of life in

39:55

July of this year? What does that migration path look

39:57

like and how can we help? And that's where we're

39:59

looking at it. looking in ways to help automate

40:01

from an admin perspective. If you're working with us,

40:03

we've been through this, we can actually go through

40:05

and build out that new machine and do a

40:07

lot of the backend manual work for you. So

40:09

that all you really have to do at the

40:11

end of the day is validate your applications up

40:14

and running in the new space, and then we

40:16

automate the switch over. So we've worked through a

40:18

lot of that. There's also the decisions you're making

40:20

around, I'm paying a very large bill for something

40:22

I'm not necessarily getting the most value out of.

40:24

I don't want to continue down that path. We

40:26

can help you make that shift over to an

40:28

open source operating system, Rocky Linux, and help drive

40:31

what's next, help you be involved in a

40:33

community and help make sure that that environment

40:35

you have is stable, it's going to be

40:37

validated by the actual vendors that you're using

40:39

today. And that's really where we want to

40:41

be as a partner from not just an

40:43

end user perspective, but as an industry perspective,

40:45

we are working with a lot of those

40:47

top tier vendors out there of certifying Rocky,

40:50

making sure that it gets pushed back to

40:52

the RESF, making sure that we can validate

40:54

that everything is there and secure that needs

40:56

to be there and helping you on that

40:58

journey of moving. And that's where we, CIQ,

41:00

really show our value on top of

41:02

an open source operating system is we

41:04

have the expertise. We've done this before.

41:06

We're in the trenches with you and

41:08

we're defining that path of how to

41:10

move forward. Okay, ops and sysadmin folks

41:12

out there, what are you choosing? CentOS

41:14

is end of life soon. You may

41:16

be using it, but if you want

41:18

a support partner in the trenches with

41:20

you, in the open source trenches with

41:22

you, check out CIQ. They're

41:24

the founding support partner of Rocky Linux.

41:27

They've stood up the RESF, which is

41:29

the home for open source enterprise

41:31

software, the Rocky Enterprise Software

41:33

Foundation that is. They've helped

41:35

to orchestrate the OpenELA, a

41:38

collaboration created by and upheld

41:40

by CIQ, Oracle, and

41:42

SUSE. Check out Rocky Linux

41:44

at rockylinux.org, the RESF at

41:47

resf.org. And of course, if

41:49

you need support, check out

41:51

our friends at ciq.com. Did

42:19

you have a business plan though? I think

42:21

you have alluded to quite well that you've

42:23

had a product plan. Yes. And

42:25

a product plan is not a business plan. Mm-hmm. Did

42:28

you think about the business while building

42:30

the product? I did. I

42:32

did think about the business while building the product. And

42:35

pre-building the product, my plan was

42:37

always, oh, because I was

42:39

still like, how can I make this B2B somehow?

42:43

So I said, oh, what I'm going to do is, publishers,

42:47

I'm going to make like, give

42:49

publishers a platform to connect to their

42:51

readers so they would have a space on the app.

42:54

And also, maybe there's a way that

42:56

I can do these reports, like these

42:59

industry reports or reader reports where it's

43:01

like all anonymous data, but publishers can

43:03

see trends of everyone is

43:05

looking for like a dark

43:07

fantasy romance and it doesn't like there are

43:09

only, there's not like commissioned dark, more dark

43:12

fantasy romance books, things like that, where it

43:14

would be like insights. And because we had

43:16

the moods, which no one else had, we

43:19

would have different insights, unique insights to specific

43:21

things that people were looking for. Yeah. So

43:24

there are two things. One is I wanted

43:26

to kind of avoid anything that was data,

43:29

even though giving user data, even though it

43:31

would have been anonymized and would have been

43:33

more trends. But two, I had

43:35

a chat with a friend, colleague,

43:37

kind of sometimes mentor, kind of

43:40

fella. And he said to me, the

43:42

best businesses always make

43:45

money or get value directly from their customers.

43:47

Like they don't have a middle person, like

43:49

a middle entity. And if you're

43:52

going down this publisher plan, you're

43:54

going to have a middle entity. You're like,

43:56

it's not direct. Oh, sorry.

43:58

The other thing is I didn't want to do ads. was nothing.

44:00

If I could avoid ads, that would be great.

44:02

One of the things that people hated on Goodreads

44:04

were the ads. So I was

44:07

thinking, okay, maybe it is freemium then. And

44:10

so really, the thing that

44:12

made us start to build out

44:14

the paid plans was just our costs.

44:17

We were spending thousands of dollars a month

44:20

to keep everything running on the highest

44:22

Heroku plans and the database.

44:24

Every time we had a spike, we needed to upgrade to

44:26

the next database. We were getting to the highest tier. What

44:28

happens when the highest tier runs out? What do we do?

44:30

And it felt like every time we

44:32

go on influx of users, our costs went up. So we

44:35

needed to do something about it. And

44:37

so that's when I just

44:40

made up a paid plan. I literally

44:42

made a page. It didn't exist. I

44:45

said, the plus plan coming at

44:47

this point, it would have been 2022. So it was in 2021. Oh,

44:51

wow. I'm like, is it 2021 or 2022? It must have been 2021 because we

44:53

became profitable in 2022.

44:56

So in 2020, at the end of 2020, October, I made

44:58

a page that was apps.storygraph.com.

45:06

And I said, hmm. And I went through and I

45:09

looked at the features we had.

45:11

And I thought, how can we

45:13

enhance these? Okay, similar books, we're

45:15

going to give you personalized similar books. So you

45:18

have a unique page just for you that

45:20

takes into account your unique preferences, your

45:22

unique themes, topics, and tropes

45:24

that you've told us you've liked and shows

45:27

you the books similar to this one, but

45:29

also highlighting the ones that

45:31

have the things that you like. Okay,

45:33

cool. And then it was like stats.

45:37

Everyone loves the stats, but what if you

45:39

could compare two of the stat

45:41

pages side by side? So just taking the features

45:44

we had and just made up a

45:46

list of features. And then I

45:48

said, we're also going to have a public roadmap

45:50

on which you can vote and post. You're

45:52

going to get priority support. So we're

45:54

going to put your support. And I

45:56

said, pre-order now, $30 after that. $50

46:00

paid it up and we

46:03

got over like three months we

46:06

got 1400 pre-orders, 1400 people

46:08

paid $30 for

46:11

something that didn't exist. So

46:13

then it got to the point where I said, okay, well we

46:15

have to build this thing now. If I just go and give

46:17

all this money back, we have to build

46:19

it. And so I spent a

46:21

month building it. I remember

46:23

because we had January, which was a big, January

46:26

of 2021, we hit 100,000 users. And

46:30

then I remember saying, I told these

46:32

people early 2021, I

46:36

don't want people starting to get mad and say like, when is

46:39

this coming out? So I just changed it

46:41

to say February 2021, just

46:43

back myself. And

46:45

it came out, it's funny, I was

46:48

up, that February was wild, but it

46:50

came out, it was March 1st in

46:53

the UK, but for most of America, it

46:55

was February 28th or 9th,

46:57

I can't remember. There

46:59

you go, it's February somewhere. It was

47:02

February somewhere. That was a title of that

47:04

week's user. So like the next

47:06

week's user said, it was February somewhere. That

47:08

was what it felt. Love

47:10

it, love it. That's awesome. What

47:13

percentage of your user base is on plus?

47:17

Like a small percentage. I'm saying because

47:19

we've just had a big spike and

47:22

I haven't done the math again, but

47:24

normally it's less than 1%. It's

47:26

like sometimes I think it's 0.01% or something. We

47:30

have right now 10.5K plus users and

47:34

2.3 million, almost

47:37

2.4 million registered users. So

47:39

it's like, what's that? Like 0.0. And

47:42

small. Yeah. You

47:44

mentioned businesses though. And the

47:46

next best thing to a B2C would be a

47:48

B2G, business to

47:51

group, right? Like you've got book

47:53

clubs are super popular and

47:55

a team of people paying X dollars a month

47:57

is totally plausible because they're buying. and

48:00

wine, they're hanging out, it's groups. How

48:04

does Plus factor into book

48:07

clubs? Those are super popular. I

48:10

had to message my wife because

48:12

my wife is a group club leader. She's

48:14

very good at it and she's

48:16

got some serious friend groups from book

48:18

clubs. I said, hey,

48:21

have you heard about this babe? She's like, no. I

48:23

asked her for some feedback and I could share that

48:25

with you in this call, but I'm curious what you

48:28

think about B2G, book the groups, or business a

48:30

group kind of thing. I know

48:32

you have a book club feature, but how fleshed out

48:34

is it? Oh my gosh.

48:36

We feel like the book club, yeah,

48:38

I'm triggered right now. It's a fake

48:40

feature. Is it a fake feature? No,

48:42

I'm triggered because book

48:45

clubs were meant to be out by December 31st of

48:47

2023, for

48:50

years we were met. So

48:53

they didn't come out. So I said, they'll come out

48:55

in the first week of January. We

48:58

had a wild January, two weeks of

49:01

downtime. We couldn't handle the demand. There's

49:03

a whole thing there. We

49:05

had to re-architect a bunch of things and

49:07

we still have some plans for re-architecting how

49:09

our database works. I don't

49:11

know if we'll have time to get into that in this

49:14

conversation. So I have

49:16

only recently been able to pick up book clubs

49:19

again. I think there was, in one of

49:21

the ideas that I was throwing around, there was a plan

49:23

that essentially I thought that

49:25

plus users would be able to host

49:27

book clubs. My thinking was that

49:30

I didn't think you could have a model

49:32

where every member would pay for a book

49:34

club because people can run book clubs for

49:36

free on several other platforms. But I thought,

49:38

oh, if we have it such that plus

49:40

members could host book clubs, that would

49:42

be great. However, my customer research revealed

49:44

to me a few things. And

49:47

this is customer research of people who use

49:49

plus, customer research of people who

49:51

are book clubs, all these things. And

49:53

what I learned was plus is not working. We

49:56

have product markets fit with the free product. We do

49:58

not have product markets fit. it fit at

50:00

the moment with Plus. Why? Because today,

50:03

two years later, the Plus plan is

50:06

still that grab bag of features that

50:08

I made up. And so what happens

50:10

is most people who originally, who take

50:13

out the Plus plan, it's because they

50:15

want to just support an independent alternative

50:17

to Goodreads. And

50:19

they probably have one of the eight

50:22

features that they use. And most of

50:24

them gravitate towards

50:26

the SaaS features. And

50:29

so I did this whole customer research and

50:31

I said, we need to completely change how

50:33

the Plus plan looks. So I'm currently, and

50:35

I say currently, it's been on pause for

50:37

a while, but I should be picking it

50:39

up in February. I'm currently

50:42

in the process of completely changing

50:44

the Plus plan to be an advanced

50:46

SaaS plan. So all it's going to

50:48

have is going to be just

50:50

all the extra charts. You're going to be able to

50:52

create your own charts because I discovered

50:54

through my research that the power users are the

50:56

ones that care about the SaaS. They have their

50:58

own Excel spreadsheet. Separate of

51:01

that, I've been doing customer research around

51:03

with Book Club people, but also just seeing,

51:06

thinking about the vision for the product and what works

51:08

well. And it's far better because

51:10

there are a bunch of features within the Plus

51:12

plan right now around buddy read book suggestions, where

51:15

you can put in the usernames of people that

51:17

you want to read a book with. And we

51:19

have a machine learning thing that will tell you

51:21

the best books to read for this group,

51:23

which will obviously be part of

51:25

that will be included in our Book Club feature. Those

51:28

features in terms of like network effects, it's

51:31

better that those are free and

51:33

that they actually more people can use them

51:35

and come on board. And then maybe the

51:37

power users in the group or the ones

51:39

that care about SaaS might upgrade their accounts

51:41

to Plus, but now the group

51:43

and no one stops from joining. Because

51:46

the other thing I realized is that with

51:48

that Plus plan, what someone might be into

51:50

staff, someone else is into running a book

51:52

club, someone else is into, you know, running

51:55

a buddy read or whatever, whatever it is.

51:57

And so I definitely need to see. separate

52:00

out the identity. The identity of the

52:02

Plus user is going to be the

52:04

advanced SaaS power user. Then

52:06

book club, buddy reads, read alongs, all

52:08

our social features, you can opt in,

52:10

it's all free. If you're a

52:13

Plus user, you will get advanced extra SaaS

52:15

insights into any of the things you do.

52:17

That's going to be the core of Plus,

52:20

but it will be free to run your group. I

52:23

had to change the estimate to end of January,

52:26

but it's 31st of January. There's

52:28

a reason I chose to do this interview

52:30

on the 31st of January in

52:32

my evening. I thought by this point, I

52:34

have everything done that I want to get

52:36

done. It's still not done. You need to

52:38

tell your wife to check it out again

52:40

when book clubs are out. I'm

52:45

sure she will. The one thing she did say though was

52:47

that, in quotes, Goodreads, and

52:49

this my wife is a designer, an

52:52

interface designer even. She says, in

52:54

quotes, Goodreads has a terrible UI,

52:56

it's clunky and not very intuitive.

53:00

She says, we'd like to make a group in

53:02

Goodreads and then suggest books to each other. I

53:04

like AI, but still like people too. I

53:07

know that a lot of things that she does is around

53:09

this group of folks she's with, and they're

53:11

always talking about what book's coming

53:14

next, which is core to

53:16

the help you give to an individual, but

53:18

then you move that to a group. I'm

53:22

not designing your business, but I just think there's maybe

53:24

some fruit there with, because

53:27

groups tend to be more invested. An

53:29

individual will just be not

53:31

so much not invested, but groups tend to

53:33

invest further emotionally into

53:36

a product and be daily active

53:38

users or at least monthly active

53:40

if they're a monthly book

53:42

club. She's been

53:44

doing this for multiple years. These

53:47

friendships are deep and they

53:49

are super meaningful to her. Being a

53:51

leader of this group is super important to her because

53:53

it's her friends. It's not just people,

53:56

it's her girlfriends. how

54:00

much deeper you can go there if there's like that's where the

54:02

crux of the business is I don't

54:04

know in terms of oh, sorry

54:07

something I didn't mention is I Asked

54:10

what I did my custom interviews with the book

54:12

club host. I asked like money questions to say

54:14

like, you know What do you

54:16

spend money on? Are there any membership fees? Do you know

54:18

do you pay for the tools you currently use and it's

54:20

basically there's nothing there, right? so

54:23

that's not to say that there won't be something in the

54:25

future, but I think that I'm

54:28

trying to focus on kind of what you

54:30

Said which is giving people a digital space

54:32

to have to host their book club within

54:34

an all-in-one solution So for example, I'm so

54:37

excited for the I spoke to

54:39

when I did my research round Normally,

54:41

I mentioned how I I speak to

54:44

very specific groups of people But

54:46

this time I on Instagram and Twitter I

54:48

asked for anyone who runs a book club

54:52

And I asked is to ask them to

54:54

tell me is it remote or in person?

54:56

How regularly do you meet is there a

54:59

theme or not? And then I

55:01

said to Abby Can you go through and choose

55:03

ten people who are as

55:06

different as possible? So I want small

55:09

Remote Large in person.

55:11

I wanted to start there because I said to

55:13

myself it might not be possible

55:16

But if I can based on these

55:18

interviews of interviewing wildly disparate styles of

55:20

the club groups If

55:22

I can build a feature that helps all of these

55:24

then that would be awesome And

55:26

so I have developed a feature set

55:29

that will cater to whether

55:31

it's an in-person Whether it's a

55:33

hybrid whether it's virtual in that you're

55:36

gonna basically be able to you invite

55:38

people You're gonna be able to

55:40

have story graph help suggest books for

55:42

the poll to have the community vote of what

55:44

shall we read next? But also the host can

55:47

just choose if that's how they currently do it

55:50

You're gonna be able to have an

55:52

agenda where you can add photos from the

55:54

meeting You're gonna be able to

55:56

like add links and you know, everyone's rating will show

55:58

up and all that kind of of stuff, then

56:00

that will be feed into having this journal,

56:03

this digital journal that have the history of

56:05

all your meetings, what was said, the photos,

56:07

whatever, so it's like a digital journal. You'll

56:09

also have a leaderboard for your club where

56:12

it's like, oh, these are the top 10

56:15

books that we've ever read. So even if someone

56:17

joins, they can go back and say, oh, this

56:19

is what this club likes. Let me

56:21

go and read these in the times when I'm not

56:24

reading the book club picks or just so that you

56:26

have that history as well. And we

56:28

want to look into actually being able to

56:31

host the meetings on the app for the

56:33

digital ones as well. So basically just having

56:35

this one space. So we are really looking

56:37

at trying to create a space where people

56:40

feel like, ah, finally my book club has

56:42

a home on Storygraph. You

56:45

said you have 2 million users? Registered

56:47

users. What's the activity of those 2.4?

56:51

Is like half of them pretty active? A

56:53

quarter of them pretty active? I need to

56:55

review it because around this time it's really hard

56:57

because we have this spike and it kind of

56:59

needs to settle down again to see where we

57:01

even out. But historically, we tend to fluctuate

57:04

between 25% to a third. Typically,

57:07

our active. And

57:09

that's like we look at most active

57:11

users because most people read a book on average a month.

57:14

And we look at checking in

57:16

the dashboard, adding or doing something to a

57:19

book in a library. I

57:21

have a list of key actions which

57:23

count that classify you as active. And

57:26

so we typically have about a quarter to a

57:28

third active user. Well,

57:31

as I was sharing with you about this whole

57:33

business to group idea, I was like, bite my

57:35

tongue because I'm like, gosh, there's like 2.4 million

57:38

people that have a problem that have already

57:40

given you their opt-in. Why just focus on

57:42

the groups in terms of

57:45

discovering your value to that many? Because

57:47

you obviously have a value to individuals.

57:50

I mean, with that many users, what

57:52

problem do they all have versus

57:54

just the groups who decide to meet together and

57:56

share their fascination with their books? A lot of

57:58

people are just like. introverts and

58:01

don't share aside from

58:03

the folks that are blogging about

58:05

it or books, the grammars, etc.,

58:07

they're the ones that are being extroverts

58:10

with their process and what they read.

58:12

You probably have a lot of people just like,

58:14

I just want to organize how I read better

58:16

and have my own personal journey, not my collective

58:18

journey. So I was like kind of bite

58:21

my tongue on suggesting that you

58:23

should dig into groups when it could be more just

58:25

like, you've got 2.4 million, what

58:28

do they all have in common in terms of a problem that they

58:30

would pay to solve? And this is the two

58:32

in this. One point is to

58:34

say, we haven't yet done this

58:37

transition that I believe will accelerate

58:39

the adoption of plus. So I feel the reason

58:41

why the adoption of plus right now, what you

58:43

see people saying is, I don't

58:45

even need to get plus because the free product does everything

58:47

I want. And it's

58:49

this grab bag of features that adds little enhancements

58:51

to a whole range of features of

58:54

which most people are not interested in all of them. And

58:57

so what I'm curious to see is, when

58:59

we transition it to, you

59:01

can literally create your own chart, you can create your

59:04

own, let's see what adoption looks like

59:06

then. But we also have another

59:08

revenue stream, which is

59:10

giveaways. So publishers

59:12

and authors pay to list

59:15

competitions for their books on

59:18

the app. And we've got two tiers there. We've

59:20

got a standard tier and a premium tier. And

59:23

with the premium tier, your book is on the homepage.

59:26

And it's like there's a link everywhere your

59:28

book shows up. There's like a clear gold

59:30

link of like enter giveaway and things like

59:32

that, those extra features, you get a custom

59:34

notification. So you can, because you would have

59:36

like tens of thousands of entries. So when

59:38

the giveaway ends, your personalized message gets sent

59:40

to all of those entries. And

59:42

essentially, this is a win-win-win,

59:45

because it's a win

59:47

for the publishers and authors,

59:49

because they essentially have a way to advertise

59:51

their books to our community, millions

59:53

of users. It's a win for the users,

59:55

because to them, it's not really ads, it's a chance to

59:57

win three books. And then it's a win

59:59

for us. because it's another revenue stream. And right

1:00:01

now, we're getting more revenue from Plus, but I

1:00:03

do see a world where depending on how it

1:00:06

goes, the giveaways overtakes that we'll have to see

1:00:08

it's hard to tell right now. But

1:00:10

we do have that other revenue stream, which is our kind of B2B

1:00:13

kind of... Do you know how many users

1:00:16

Goodreads has? It's rumoured around

1:00:18

80 to 90 million, I

1:00:20

believe. So I was doing the math, 10.5 on

1:00:23

2.4 million is 0.4 of a percent. What

1:00:25

do you know about typical

1:00:30

is on Framium? I would expect

1:00:32

1% in terms of upgraders. It

1:00:35

is about 2% or something.

1:00:37

Oh, in terms of converting? Yeah,

1:00:40

like a typical converter. I feel like 0.4% is low. Yeah,

1:00:43

it is low. I think it's like you're

1:00:45

looking at 2% to 5% or something like

1:00:47

that. I think that's the standard

1:00:49

rate. Yeah. I'm just

1:00:51

over here thinking what's the

1:00:53

best user of your time cranking

1:00:55

that percent up or just trying to get

1:00:58

to 80, 90 million? Because a small

1:01:00

conversion rate on 80, 90 million is still a lot

1:01:02

of money or converting that

1:01:04

rate. Obviously, we can walk and

1:01:06

shoot them at the same time. But if you're a solo

1:01:08

dev, you kind of can't, right? You're working on one feature

1:01:10

or the other. Yeah. And our

1:01:12

intention is to keep the team small because it

1:01:15

has a lot of benefits. Because the

1:01:17

other thing is what 80 to 90

1:01:19

million users look like in terms of customer

1:01:22

support in terms of... You know what I

1:01:24

mean? Right. Cost. Operational.

1:01:27

Mm-hmm. So there's that to keep in

1:01:29

mind. I would hope that part of

1:01:31

the reason why we can... If we could grow

1:01:33

to 80 to 90 million is because a lot

1:01:35

of the UX pain points have been sorted out,

1:01:38

everything's super intuitive. We've got

1:01:40

a very helpful help base

1:01:42

or whatever for the... Things

1:01:44

like that. So it's hard to tell. Yeah.

1:01:47

So how do you pick on what to work on then?

1:01:49

How do you choose? I'm starting... I'm

1:01:51

currently working on is getting that presented up. That's

1:01:54

what I'm currently working on, right? How

1:01:56

did you decide that then? Versus the

1:01:59

other stuff? Honestly... before we

1:02:01

became profitable, this is actually what happened.

1:02:03

It was more about me

1:02:05

and being like, I don't have a salary and

1:02:07

I'm on my last year of runway, right? So

1:02:10

it got to the point where I'm like, for Storygraph,

1:02:13

for us to not get funding and for it to

1:02:15

stay independent, I need to start

1:02:17

getting a decent salary from Storygraph. Rob needs to

1:02:19

start getting a decent salary from Storygraph. We've

1:02:22

always been paying Abby a decent, I've always been paying Abby

1:02:24

a decent salary from day one and we

1:02:27

were like, what can we do? So this is

1:02:29

pre giveaways, plus was

1:02:31

there but it was kind of just, you know, it was

1:02:33

a 0.01% at one point. And

1:02:36

so I was like,

1:02:38

could we do ads in a smart way? Because of

1:02:41

all the page views we have. And the thing is,

1:02:43

even when we thought we were being smart about it,

1:02:45

we always just came out at,

1:02:48

even if we weren't sending any user

1:02:50

data to publishers, like if it was

1:02:52

just a behind the scenes matching process

1:02:54

that the publishers had their input and

1:02:57

then we matched behind the scenes to

1:03:00

make sure that users were seeing relevant

1:03:02

ads. We just thought that even

1:03:04

just the thought of user thinking we'd given their

1:03:06

data to publishers just didn't sit right with us.

1:03:09

So there was this one like series of

1:03:12

formative conversations with Rob and

1:03:14

with Saron, my friend Saron, his wife,

1:03:17

where I was like, okay, that's

1:03:19

where I kind of thought, oh, giveaways, that's how

1:03:21

we get another revenue stream. And then I said,

1:03:23

we don't have product market fit with plus. How

1:03:25

do we get product market fit with plus so

1:03:27

that that percentage goes up? That was kind of

1:03:30

the journey to just choosing to do that.

1:03:32

Yeah, makes sense. It came

1:03:34

from a personal like, well,

1:03:36

my runway runs out in two years. So

1:03:40

this is something needs to change. Yeah,

1:03:43

right. I like the giveaways idea.

1:03:45

I think it is it can be more of a

1:03:48

win-win-win because you have, you know,

1:03:50

there's something there for the user, not just hey, look

1:03:53

at this book that we even if we've better the

1:03:55

book. Yes. And it's like, well,

1:03:57

we've entered the book, the person's paying to

1:03:59

promote it. it but you know, he has a chance

1:04:01

at a free copy. It's like, everybody likes

1:04:03

free stuff. So yeah, that's a

1:04:06

good idea. It's a good way of doing

1:04:08

it so that it's not a win win

1:04:10

lose for the end user, which ultimately, you

1:04:12

know, drives down enjoyment and

1:04:14

satisfaction, right? And drives

1:04:16

people away if it becomes egregious.

1:04:19

Less so I think if it's a really good

1:04:22

fit. I mean, we are advertising based company here

1:04:24

at changelog. We are sponsored. We

1:04:26

also have a membership so you can opt out

1:04:28

of that. I don't know, Adam, 1%. What

1:04:30

do you think is our percentage? 0.001? Pretty

1:04:33

small. We don't know what it is. It's pretty small.

1:04:35

Yeah, but that's all we're optimizing for

1:04:37

either, at least now. No, it's not. But our,

1:04:40

the point I was getting to was that we

1:04:44

are very intentional with the sponsors that we work

1:04:46

with, and the way we reduce the ads

1:04:48

and all this kind of stuff. So that

1:04:50

we think it's a decent trade off for

1:04:52

keeping the podcast free for everybody. And our

1:04:54

listeners even sometimes ones that sign up for

1:04:57

plus plus or like, can I sign up

1:04:59

for your sorry, it's better. It's better

1:05:01

than yours now. Yours is plus ours is plus plus.

1:05:03

So it's slightly better. Oh, it was actually plus plus.

1:05:06

Changelog plus plus, you know, because

1:05:08

we're nerds. We like to increment

1:05:11

things. I love it. Plus

1:05:13

plus. Love it. Yeah. Some people sign up

1:05:15

and say, Hey, can I keep the ads in there?

1:05:17

Because they're valuable to me. I enjoy it because we're

1:05:19

exposing them to new things. And so if done right,

1:05:21

I think you probably could crack that nut.

1:05:24

But if you can figure out

1:05:26

even more interesting ways like giveaways of

1:05:28

doing the same thing, then more power to

1:05:30

you. When we were sketching out what

1:05:32

introducing ads to the platform would look

1:05:35

like, I said to Rob, what

1:05:37

we should do is have it such that

1:05:40

if you pay for plus, you can turn off

1:05:42

the ads. But I want it such that most

1:05:44

plus people do not turn off the ads because

1:05:46

they're that good. That was like the goal. That's

1:05:48

what you what I said. That's

1:05:51

kind of how we think. That's great. Yes.

1:05:54

But then, of course, we do have people to say, I signed

1:05:56

up because I don't want the ads anymore. So there you go.

1:05:58

I mean, it's not like everybody. But they're

1:06:00

good enough that some people find value in them. And

1:06:02

so we treat them like content as much as we

1:06:04

can, which is a lot of

1:06:06

hard work. So I'm sure giveaways probably some hard

1:06:08

work as well at this point. Yeah. It's

1:06:10

a whole other product that I'm building basically. Yeah,

1:06:12

exactly. It's a whole other platform because there's a

1:06:15

whole backend dashboard to that. Um,

1:06:17

and it's still in beta because there's still some

1:06:19

manual parts of the process. And I really hope

1:06:21

that we can make it live today. Just

1:06:24

on the trusting, a couple of points that I

1:06:26

want to mention that I think

1:06:28

are cool, but they could also factor

1:06:31

into lower numbers in that

1:06:33

actually not really. Anyway, the

1:06:35

point that I want to make is we don't take

1:06:37

any payment details when

1:06:39

you sign up for the trial and

1:06:41

also the trial, the trial is also 30 days. And

1:06:45

so on one hand, I

1:06:47

think the trial is quite long and that

1:06:49

it's very easy in 30 days for someone to like

1:06:51

not feel the pressure to use it. Then

1:06:54

they kind of don't use it. And then when it comes

1:06:56

to try and sign up, they're like, I didn't really use

1:06:58

it. I guess I don't need it. And so they don't.

1:07:01

Whereas maybe if it was shorter, they might immediately

1:07:03

try and explore all the features and

1:07:05

then sign up. But the

1:07:07

cool part of all of this is knowing that when we

1:07:09

look at our subscriber number, when I look at about 10,570,

1:07:11

whatever it is, knowing that every single one

1:07:16

of those people actively, their

1:07:19

free trial expires and they actively came back

1:07:21

on and said, yes, no, I

1:07:23

want this. And they put in their car details.

1:07:25

That's really cool. No one's a mistake there because

1:07:27

it's just not possible. Right. If

1:07:30

I have a lower turn that way, because the

1:07:32

people are very intentional about signing up. Yeah. How

1:07:34

many conversations have you had with those 10,500 folks to

1:07:36

make plus better,

1:07:40

to give them more value? Dozens, I would

1:07:42

guess. So, uh, no, dozens, it would

1:07:44

be more than, yeah, probably dozens up

1:07:46

to a hundred because I've done

1:07:48

a bunch of like over the years,

1:07:51

a bunch of summer conversations. I've had

1:07:53

reset rounds where, you know, there was

1:07:55

one where I figured out, oh,

1:07:58

no one really cares about plus they're just paying. to

1:08:00

support, you know, oh, what do

1:08:02

the plus, what do people have

1:08:04

in common? The power users, oh, they love

1:08:06

stats, like things like that, which helped me

1:08:08

get to the advanced stat transition. But there's

1:08:11

also been surveys and stuff over the

1:08:13

years too, not always direct conversations. Sometimes

1:08:16

I'll send out a survey of

1:08:18

checking in, maybe testing the waters

1:08:20

for a particular feature I'm thinking

1:08:22

about, things like that. So I've

1:08:24

spoken every single plus user,

1:08:26

well, not the 2000 who joined in the

1:08:29

last month, that was great. But no, the

1:08:32

8000, whatever they've heard from me in some shape

1:08:34

or form, whether it's just a form or a

1:08:36

survey saying, Hey, we're thinking of doing

1:08:38

this, like how do you, this currently? What

1:08:42

does success look like for this for you? Like,

1:08:44

do you have, uh, some sort of

1:08:46

thing jot down where like, if this gets to

1:08:48

some point, cause you don't, you want to stay

1:08:50

small. So it doesn't

1:08:52

sound like, you know, total world domination

1:08:54

is what you're after necessarily, or

1:08:57

maybe you are, but just with a small team. Maybe.

1:09:00

No, it is definitely not

1:09:02

a world domination, but I think I

1:09:04

would love for us to be almost

1:09:06

the default place that people, when they're

1:09:09

thinking about book recommendations or finding a

1:09:11

book where like the default place to

1:09:13

go, like people are

1:09:15

just like, Oh, check it on Storygraph or go to Storygraph.

1:09:18

And another, I guess one other measure

1:09:20

is whenever we have a spike, like

1:09:23

a big spike, we end up getting

1:09:25

to like the highest number two. So

1:09:27

we've been number one in the app store once,

1:09:30

but we've hovered between

1:09:32

two and five, right? In

1:09:34

the, in the, in the books category. And

1:09:36

I think good reason is basically entrenched at

1:09:38

number one, pretty much popping up one time

1:09:41

where we were briefly above them.

1:09:43

That was when we basically launched the app and

1:09:45

then the TikTok went viral. So

1:09:49

I would love basically, I'm imagining us

1:09:51

being up there, like entrench, like one

1:09:53

end to being there and being known

1:09:55

as the only kind of default main

1:09:57

place, personally feeling.

1:10:00

I do see Strogoth as my

1:10:02

life's work, assuming things continue.

1:10:05

I would just keep doing this as long

1:10:07

as it stays fun. I feel like that is

1:10:10

a possibility that I keep doing this. In

1:10:12

which case, it would need to get

1:10:14

to the point where I feel like I

1:10:16

can get a salary that would support whatever

1:10:18

life choices I want to make, whether that's

1:10:20

relocating, whether that's having a wedding,

1:10:23

whatever it might be. Because right

1:10:25

now, I don't feel like that. I've got a decent salary

1:10:28

now and Rob has two, but for

1:10:30

me, it's not enabling my savings. The savings

1:10:33

that I've threw down over the

1:10:35

last five years, they're not growing back up

1:10:37

right now. It's

1:10:39

things like that where there's a personal

1:10:41

lifestyle where it's like, okay, Strogoth is

1:10:44

supporting me and my lifestyle and

1:10:46

any future changes that might come to that

1:10:48

lifestyle. And also, we are

1:10:50

just known as the de facto default. You

1:10:53

want books? Strogoth. I

1:10:55

have an idea for you. Oh, yes, please. I

1:10:58

don't even know if it's good. I'm thinking

1:11:00

out loud here. I

1:11:03

love problems. I love problems. I

1:11:05

could tell you've been steering away. Even with

1:11:07

the group things, I could just feel your

1:11:09

face. You're like, hmm. Well,

1:11:11

because there's no right way to

1:11:14

success. Even as Jared asked that

1:11:16

question, you really have a

1:11:18

not struggle to describe it, but it just

1:11:20

seems like just don't die as your success.

1:11:23

Just don't die. And if you can really

1:11:25

succeed well, that's obviously a plus. No

1:11:28

pun intended. But I wonder

1:11:31

if you want to be the default for

1:11:33

a good book recommendation, then don't put it

1:11:35

behind a sign up. Give

1:11:38

somebody someplace to go. Ask

1:11:41

them three questions. Give them a result, and

1:11:43

they're going to love you immediately. I

1:11:46

don't know how your recommendation works, so take this with

1:11:48

a grain of salt. But what if that recommendation was

1:11:50

so compelling that they were like, I have to sign

1:11:52

up afterwards? Or maybe you really don't struggle

1:11:54

with sign ups, but if you want to be

1:11:56

the default, find a way to give

1:11:58

them a result that's worth it. worth their

1:12:01

attention without any friction in

1:12:03

the process. The signup process

1:12:05

is always pushy because

1:12:08

you're like, what am I signing up for? And then you get

1:12:10

a form to fill

1:12:12

out to give somebody some information. And maybe

1:12:15

the thing that came from the TikTok video or

1:12:17

the YouTube video, or maybe that was the compelling

1:12:19

artifact that doesn't make that part of struggle for

1:12:21

you, and I don't know. But

1:12:23

if you want to be the default, you have to

1:12:25

kind of give that thing away, ad nauseam for free.

1:12:27

Just give it away for free. And everybody's like, man,

1:12:29

this is the place to go. And

1:12:31

they have no problem with giving you more keys to

1:12:34

their kingdom. So here's the thing.

1:12:36

We have the previous current site.

1:12:38

We had two million registered

1:12:40

users, but we were getting 4 million

1:12:43

unique visitors every month. So there's

1:12:46

like double are just someone's

1:12:48

link them. It could be a giveaway page, which

1:12:50

you can view all the giveaway pages without signing

1:12:53

up. We also have, you can view

1:12:55

all the book pages too, sure. We

1:12:58

also have a browseable

1:13:00

books page which have a filter menu. It's not

1:13:03

like the recommendation engine where you actually put

1:13:05

in your preferences. But you can say,

1:13:07

oh, I want something funny, fast paced, whatever. And

1:13:09

you can see, you can order by last updated

1:13:12

or page size

1:13:14

or whatever it is. Sorry, yeah, book

1:13:16

size. But there is a feature

1:13:18

that we released recently that we have been talking

1:13:22

about and saying, can we make

1:13:24

this free as in such

1:13:26

that you don't need to sign up, which is

1:13:28

that we currently have something called a Storygraph preview,

1:13:31

personalized preview. And essentially,

1:13:33

it's an AI generated summary

1:13:35

of readers the

1:13:37

book would be good for. And also, if

1:13:41

you have the personalized version, which is currently

1:13:43

plus because originally, we thought

1:13:45

it was going to be too costly to offer it to everybody.

1:13:48

The personalized version tells you if we think

1:13:50

the book is going to be good for you. And it's

1:13:52

a text that's completely in our solution. And

1:13:55

there was a world where we're excited for

1:13:57

it to be free to everyone because there's

1:13:59

Then it's like, wow, even if your main

1:14:01

tool is Goodreads, if you're in

1:14:03

a bookstore, you want to be looking at

1:14:06

it on Storygraph because we're

1:14:08

the only place that are going to tell you, oh,

1:14:10

given your interest

1:14:13

in so-and-so themes, blah, blah, blah, blah, blah, blah, you

1:14:15

might like this book. But what we also do is

1:14:17

if you've specified any type of thing you don't like

1:14:19

reading about, like if you don't like reading sad books

1:14:21

or if there are certain triggers and things you don't

1:14:23

want to read about, we'll alert you

1:14:25

to say, be warned, it is still

1:14:28

the themes of blah, blah, so be

1:14:30

careful, proceed with caution. This is a

1:14:33

new feature which we're hoping

1:14:35

to make free for everyone, but we have discussed,

1:14:37

is it possible to just have this be on the

1:14:39

pages? We need

1:14:41

to talk about it some more because there's something

1:14:44

about whether it's to do with AI opt-out or

1:14:46

whatever. We need to just look

1:14:48

into the implications cost-wise, resource-wise, because

1:14:50

we're getting full million users a

1:14:52

month and just check. Yeah,

1:14:54

just review if that is actually a possibility,

1:14:57

but that would get at what you're talking

1:14:59

about, which is giving people a result without

1:15:01

needing them to sign up. So

1:15:04

I don't have the answers, right, obviously, because I'm not

1:15:06

deep in your team. I don't know what's required to

1:15:08

give this thing away, but I just wonder if that

1:15:11

could be compelling, because if anyone

1:15:13

could go to a non-friction

1:15:15

page and get some value, then

1:15:18

they're willing to give you more of their attention

1:15:20

in time. It's just my rough thought on that.

1:15:22

And that could be plausible. It could not be

1:15:24

plausible for you to accomplish. But

1:15:26

if you want to be the default, you kind of have to give

1:15:29

somebody a taste of what the goodness

1:15:31

is, right, as some sort of tantalizer to

1:15:33

get in. It could be a landing

1:15:35

page. It could be an E.B. test. Doesn't even have to

1:15:37

be to everybody. It could be, let's test this out on

1:15:40

an ad campaign that we do, or some way

1:15:42

we drive and funnel folks to this one page

1:15:44

that test out this theory in a

1:15:47

20,000 person spectrum versus a

1:15:49

200,000 spectrum because you're

1:15:51

getting four million visits a month or whatever

1:15:53

the number is, where you really can't spend

1:15:55

the money on that and say, OK, wow,

1:15:58

what was the conversion on $25? 5,000

1:16:00

folks, does it make sense to

1:16:02

scale that? Does that really bring us in high quality

1:16:05

people who really love to read and

1:16:07

really can find value not only in

1:16:09

our platform and the recommendations but then

1:16:11

are a high candidate for whatever Plus

1:16:14

evolves into? Yeah. That's how I think of it. If

1:16:37

you're listening, you may remember the

1:16:39

early days of the internet where

1:16:41

open networks like HTTP and SMTP

1:16:44

led to explosion of websites and

1:16:46

online communities. Building a fan site

1:16:48

and connecting over shared passions led

1:16:50

so many of us to careers

1:16:53

in software. Back then it seemed

1:16:55

like anything was possible because the

1:16:57

internet of the 90s was built

1:17:00

to democratize information not consolidate it

1:17:02

with a handful of big tech

1:17:04

companies. Read, write, own,

1:17:07

building the next era of the

1:17:09

internet is a new book from

1:17:11

startup investor Chris Dixon that explores

1:17:13

how network architecture plays out in

1:17:15

our online lives and the decisions

1:17:17

that took us from open networks

1:17:20

governed by communities of developers to

1:17:22

massive social networks run by internet

1:17:24

giants. Read, write, own

1:17:26

is a playbook for reclaiming

1:17:28

control and for reimagining applications

1:17:30

so users can own, co-create,

1:17:32

and even profit from the platforms they

1:17:34

use every day. From AI

1:17:37

that compensates creators to protocols that

1:17:39

reward open source contributions, this is

1:17:41

our chance to build the internet

1:17:43

we want not the one we

1:17:45

inherited. Order your copy of Read,

1:17:48

Write, Own today or go to

1:17:50

readwriteown.com to learn more.

1:18:12

Here's something to think about, Halia. Yeah,

1:18:14

cool. Let's go. Have you

1:18:16

entertained or would you entertain an acquisition?

1:18:20

No, as in the answer

1:18:23

right now is no. Like no, you

1:18:25

haven't and no, you wouldn't. The thing is,

1:18:27

I'm just in general with life. I'm wary

1:18:29

of saying never say never because

1:18:31

I don't know what's going to happen in five

1:18:33

years, 10 years. Am I still going to be

1:18:35

having fun? Am I, I feel like I'm pretty

1:18:37

good at looking after myself. Like everyone's always

1:18:40

worried about my wellbeing and I'm like, I'm

1:18:42

so far away from burnout. Like I

1:18:45

really look after myself. I, you know, so

1:18:47

the answer right now is like, no, definitely not.

1:18:50

We couldn't, but I'm not going to say never because life

1:18:53

is long and hopefully that's good. And

1:18:56

so yeah, but the answer right now

1:18:58

is no. And that is not in my plan and it's

1:19:00

not in my intention. My plan is to continue running

1:19:03

and owning Storygraph forever. Let's

1:19:07

assume success then. Let's go down the road five

1:19:09

years, maybe 10, but let's at least go down

1:19:11

the road five and say that

1:19:13

Storygraph has displaced Goodreads and the

1:19:15

Nezeitgeist of human knowledge and you're

1:19:18

number one on app downloads and

1:19:21

you have more users than you know what to do with your

1:19:24

own. Everybody loves the product and here comes Amazon

1:19:27

and knocking. No, Amazon's not making

1:19:29

Amazon. That's too easy. Yeah,

1:19:31

Amazon is no. Like that's like, that

1:19:33

would almost be like portraying our users.

1:19:37

Adam has an opinion though. Like

1:19:39

my dad,

1:19:42

my dad has an opinion on that. Okay. Yeah,

1:19:45

I bet. All right. Amazon doesn't

1:19:47

come knocking, but somebody else does

1:19:49

who has goodwill in heart. Then

1:19:52

you're thinking about it, then you may be like, well, sell.

1:19:55

If it's going well and we're running it, then no reason

1:19:57

to sell. All right. I

1:20:00

like it. Yeah, I think. Right now I

1:20:02

would say that the only. V C. Know

1:20:04

I'd it because it's again if what

1:20:06

were possible now so we sits live

1:20:09

on that Vc train. I still like

1:20:11

a wolf only bring bad thing to

1:20:13

focus because once you've been vc on

1:20:15

it's almost a with the learn math

1:20:17

of race to. Acquisition.

1:20:19

Life, yo! I would hate to feel pressured

1:20:21

to do either of those things. View

1:20:23

to get paid of. Exit Say now I'm

1:20:26

unless that's a of venture capitalists that comes

1:20:28

in at a profit share versus a. As.

1:20:30

A chair that's a different venture

1:20:32

capitalists there's lots be more than

1:20:34

best based upon their return over

1:20:36

multiples based upon friends. But.

1:20:39

That's plausible. Yeah, But then

1:20:41

sitting you're talking is the coming along. what

1:20:43

would be. A lines yeah that

1:20:45

kind of investment the been only

1:20:47

money but smart money like connected

1:20:49

them words they would have to

1:20:51

advised to the. The. Giveaways

1:20:53

in a can you help me content

1:20:56

with better publishers? have deeper relationships and

1:20:58

publishers have a longer contracts, more reliable

1:21:00

revenue, etc. Yeah, that's how I would

1:21:02

view that kind of investment. Or like

1:21:04

that, who's the most famous I Susie?

1:21:07

Oprah Avast a gram in a booth?

1:21:09

Bang Yeah! I know that the Layla going

1:21:11

and say what if Oprah comes along through

1:21:13

know how she's gonna go there for hundred?

1:21:15

hey this is a lion to our like

1:21:17

one that around a disaster and are you

1:21:19

okay so hit on something Hit Welcome adds

1:21:21

a single over his book. Clubs are causing

1:21:23

a like natty as were club like with

1:21:25

you did that because like you could grow

1:21:28

that. Audience. And maybe have

1:21:30

your thing be in of the new

1:21:32

over as book club eventually. Citizen but

1:21:34

have operated Combine was like. Interested.

1:21:37

But. That seems like I are about. Spiritually.

1:21:40

Aligned about least in terms of product and

1:21:42

losing season. good at show as a as

1:21:44

that money like maybe that would be a

1:21:46

now she good value as is my only

1:21:49

point out on us along with us but

1:21:51

he kinda cincy somebody or minutes in it

1:21:53

and that's great literary as in any success

1:21:55

of semester over if you're out there. And.

1:21:58

Somebody listen to the senior. network, get

1:22:01

in touch. If you're Oprah's software

1:22:03

developer IT person, and you listen to

1:22:06

this, that's right. Good job, yeah. If

1:22:08

you're handling Oprah's DNS, get

1:22:11

in touch, right? Yeah,

1:22:14

exactly. Next time she asks for

1:22:16

a book link, send the story graph one.

1:22:18

That's right. I like

1:22:20

that idea. I mean, I think I would, if I were

1:22:22

you, that's how I would align it, because I would not

1:22:24

sell out to the other

1:22:27

Goliath, like an Amazon, even an

1:22:29

Amazon competitor. I would, because

1:22:31

like you said, you're profitable, and for now you're

1:22:33

enjoying it, right? Like you said, you're not sure,

1:22:35

five years or 10 years how you feel, because

1:22:38

none of us do, but you assume you will. And

1:22:41

so, and it's to stay on the, what's the

1:22:43

good, the gymnast who are on

1:22:45

the walking, like a plank kind of thing, whatever

1:22:48

that is. To stay on the

1:22:50

bar, don't fall off, don't look down. The balancing

1:22:52

beam, are you talking about the balancing beam? You

1:22:55

make it sound like it was

1:22:57

really stressful. And I don't know.

1:22:59

Like it's tenuous. Like you could fall off

1:23:01

and die. Exactly. Well, it's just

1:23:03

a few feet. You know, it's

1:23:06

not like, there's no sharks down there.

1:23:09

Oh, a walk in the park, no? It doesn't feel like that. It feels

1:23:11

like a walk in the park for you? I want to get, well,

1:23:13

not right now, but I want to see

1:23:15

the future of it feel like a walk in the park.

1:23:18

Right. Yeah. Hang on

1:23:20

a second. There's this idea that Jerry just mentioned, I

1:23:22

think has legs. What if there's a book club for

1:23:24

everybody? What do you mean? That's just there

1:23:26

to join, right? Groups to find

1:23:29

in a way. Oh, like pre-made book clubs.

1:23:32

Well, the Nadia book

1:23:34

club is there that she's like, hey, everyone's

1:23:36

welcome here. And then it's

1:23:39

like the bees and honey kind

1:23:41

of thing. Like, you know, it's sweet. So you

1:23:43

come and then wow, there's a recommendation engine here.

1:23:45

And you want to be the default of that.

1:23:47

So rather than say, come get recommendations, it's more

1:23:49

like, come find your group. Maybe

1:23:52

this is the group for you, but

1:23:54

something where everybody's invited. Yeah. Pretty

1:23:56

funny because we launched a read along feature at the

1:23:59

end of last year. year where you can read a

1:24:01

book with up to a thousand people and it's

1:24:03

basically for people who don't want to be in a regular

1:24:05

book club but they can opt in and you

1:24:07

can choose a book. And I was running it

1:24:09

from the Storygolf account but it was clear that

1:24:11

Nadia was running it. And we

1:24:13

had a thousand people join and we all read

1:24:15

Erin Morgenstern's The Night Circus and it was awesome.

1:24:17

And at the end I made a forum that

1:24:20

was like, share your feedback on

1:24:22

this feature, given that it's new. And

1:24:24

you know, one person wrote something like, I

1:24:27

feel like running one of these is

1:24:29

not for me, it's more for influencers

1:24:31

like Nadia. Like, and I just remember

1:24:34

laughing, you know, I didn't see myself

1:24:36

as an influencer. But you are. I

1:24:38

think you are. Here's a different angle.

1:24:40

I would be friends with somebody based

1:24:43

upon particular books I know I've read that

1:24:45

I've enjoyed. And if they've read and enjoyed

1:24:48

these books, then I could be

1:24:50

better friends with them. As a matter of fact,

1:24:52

that happened this week. Like I like Nick Moore

1:24:54

because I know specific books he and I have

1:24:56

read together and we shared like similar likes and

1:24:59

dislikes with the book. And so I

1:25:01

feel like, man, if I was like Nick's neighbor, I'm

1:25:03

talking about Nick Niek's from J.S. Partly, by the way.

1:25:05

Okay. If I was Nick's literal neighbor, I would like,

1:25:07

not that I'm not his friend because we're far

1:25:09

away, but I would be more willing to like

1:25:12

hang out with Nick because

1:25:14

I know he's read the Babaverse, you

1:25:16

know, Quintology book or whatever

1:25:18

it is, like that series, or because

1:25:20

he would take a recommendation, you know,

1:25:23

so like the finder group

1:25:25

people or the finder friends in the world,

1:25:27

the internet is similar likes and

1:25:29

that similar likes kind of comes from a book.

1:25:32

So this is a whole new direction for you now,

1:25:34

it's like Tinder for book lovers, you know. Oh,

1:25:36

they get restarted on the Tinder for book

1:25:39

lovers as a single person who has been

1:25:41

dating for years trying to find somebody. The

1:25:44

number of times people have said like, Oh,

1:25:46

have you thought about dating at spinoff? And

1:25:49

I was like, I wish I had the bandwidth for

1:25:51

a dating at spinoff because it would

1:25:53

be incredible to find a man

1:25:55

who loves books and that we can bond

1:25:57

over that. So let's come on. a

1:26:00

few times over the years, the dating up

1:26:02

spinoff. This is why you take money in. If

1:26:04

you hire a dev, they're working on that. You

1:26:06

keep working on Plus and then it's true. Good

1:26:09

things happen. It's called lack of focus is also what you call

1:26:11

it. And then it's that dev that ends

1:26:13

up being my husband. Oh my goodness. Oh

1:26:15

my goodness. Now we're on a homework

1:26:17

movie. Oh my gosh. We're like matchmakers.

1:26:19

Look what we're doing for her. Nadi,

1:26:21

are you a homework fan by the

1:26:24

money chance? Are you a homework movie

1:26:26

junkie during the holidays? No,

1:26:28

but I know the reference. Okay. You

1:26:30

laughed really hard. So I was thinking maybe you were. I was like,

1:26:32

wow, not only. Well, I kind of was part of

1:26:34

one last year. I was part

1:26:36

of one that was basically like a homework movie,

1:26:38

but it didn't end the way a homework movie

1:26:41

ends. Oh, I

1:26:43

get sucked into it. There's

1:26:47

a framework to every one of those movies. I get sucked into it.

1:26:49

My wife watches them. Sometimes she has friends over and I'm watching them

1:26:51

with her just because I'm in the room. Then

1:26:53

I'm the guy who's like, don't do that. Oh

1:26:55

my gosh. I'm the, you know, I'm like in

1:26:57

it. I get invested in this. You're the husband of

1:26:59

the back. He's like, I don't know what

1:27:01

to do. Oh, but you're going to

1:27:03

do this. But I am watching it.

1:27:06

That's half the reason I'm like, don't turn it on

1:27:08

because if you turn it on, I will watch it and

1:27:10

then I'll get invested and then I'll be one of the

1:27:12

girls. You'll have a scooter your way in

1:27:14

the front. You're crying crocodile tears. Can you believe how he

1:27:16

treated Sally? She

1:27:21

did not deserve that. The point is

1:27:23

though, is we all bond over stories, right? And I

1:27:25

was just thinking like going back

1:27:27

to the default, it's maybe the

1:27:29

way to being the default isn't to put the default

1:27:31

out front. So maybe my recommendation was incorrect. Maybe

1:27:34

it's just helping people find the platform

1:27:36

and then wow, there's this recommendation engine

1:27:38

that should be the default for everybody.

1:27:40

And I just want to tell everybody I

1:27:42

know or find more

1:27:45

people that have read the same books I've read. And

1:27:47

this is why, by the way, I think it's very

1:27:49

important that all the social reading features like body reads,

1:27:52

like read-alongs and book clubs are

1:27:54

free because you get people who

1:27:56

don't even intend to move or use a

1:27:58

book tracking app. their book club

1:28:00

hosters. We're trying this for this meeting or

1:28:02

we're doing this one thing and then they

1:28:04

try it and then they say, oh, what

1:28:07

is this thing? That was awesome. I want to do

1:28:09

this again. Oh, it's just easier if I have all

1:28:11

my stuff on here. And we have seen that journey

1:28:13

a little bit, but there's definitely more that we can

1:28:15

do on that space. There's

1:28:18

any more technical stories you can share with us? Like your

1:28:21

cues all jacked up? How's your servers

1:28:23

doing? What are you doing

1:28:25

for observability? Do you do observability? Like what's

1:28:27

your SRE stack? How do you deploy? Well,

1:28:30

for Oke now, that's a big story.

1:28:32

We migrated. Rob manages all of our

1:28:35

infrastructure. So we were on Heroku. It

1:28:37

was getting, we were on the database

1:28:39

farm that was costing us like eight

1:28:42

to $9,000 a month on top of all the other

1:28:44

costs. And at this point, plus is

1:28:46

like, you know, we have a few thousand

1:28:48

users, not enough. And so we,

1:28:52

you know, Rob, who had 15 years

1:28:54

in infrastructure experience said, Hey, Nadia, I

1:28:58

think we should move off Heroku. Me,

1:29:00

who is a boot camp graduate. All

1:29:02

I know is Heroku. What do you mean

1:29:04

we're moving off Heroku? No, no, no, no, no. And

1:29:07

he did all the research. He laid everything out for

1:29:09

me. He asked all of my questions. It was a

1:29:11

month long process of this is the way to go.

1:29:13

And so eventually we

1:29:15

did this big migration, the beginning

1:29:18

of 2022, January 2022. It

1:29:21

was like months and months of research,

1:29:23

lead up planning. You know, we

1:29:25

told everyone we're gonna have this one of those, we're

1:29:27

gonna have scheduled downtime from this time to this time,

1:29:29

we were like, now we're gonna be

1:29:31

down for five hours, moving everything over.

1:29:34

So we did everything, we

1:29:36

followed our steps. We

1:29:38

turned the app back on, everything's

1:29:40

working. And then the emails start

1:29:42

coming in to support account deleted,

1:29:45

my account we were like, oh,

1:29:47

my like my heart sunk, like

1:29:49

it was awful. Eventually,

1:29:52

we realized that what happened was in the process of

1:29:54

the upgrade, everybody had been signed out and some people

1:29:56

had not had to sign in for like three years.

1:29:58

So A lot

1:30:00

of people had two accounts or something, some

1:30:02

people were two accounts, or they just couldn't

1:30:04

remember their login details, or they didn't realize

1:30:07

they weren't planned in. Basically, we were able

1:30:09

to match every single person back to an

1:30:11

account eventually, but it was that initial minute

1:30:13

of when we turned it on, and

1:30:15

then the emails that are coming in, they were like, well, my account's

1:30:17

deleted now, my account was that, it was scary. We

1:30:20

had backups, but you don't want to be dealing with that.

1:30:22

You're like, I trusted you, Rob, no.

1:30:25

Yeah, so that was that. That's one story.

1:30:27

We had so many issues with

1:30:29

search, and

1:30:32

I started the act with PG search,

1:30:35

and it was just searching for a book

1:30:37

took like 10 seconds. It

1:30:39

was awful. I would tweet out saying,

1:30:41

what can I do? A

1:30:44

lot of people eventually got to the point where I

1:30:46

was like, I thought we need something like Elastic Search.

1:30:48

Even working out with Elastic Search,

1:30:51

that eventually had its own performance issues, and I

1:30:53

realized, I need to shard now. I just learned

1:30:55

things as and when we have problems, and it's

1:30:58

like, oh, now I need to shard. I need

1:31:00

to index it this way. I

1:31:02

was just learning on the job. One

1:31:04

day, the server

1:31:08

running our Elastic Search just died

1:31:11

or something? It was run, it was a managed, this one

1:31:13

was still in her roast at this point. We

1:31:17

were using the Elastic plugin

1:31:19

or plan, and suddenly

1:31:21

the search wasn't working, and it's very

1:31:24

key for search to work. Search nowhere

1:31:26

was working. You couldn't search for books.

1:31:30

We were like, what's going on? We couldn't

1:31:32

see anything. We emailed the support, and we

1:31:34

were paying. We were on there like, what's

1:31:36

the word for the opposite

1:31:39

of penultimate, i.e. the second from

1:31:41

the highest? I think it's

1:31:43

just called second. I love

1:31:47

penultimate.

1:31:50

It's one of my favorite words. I do

1:31:52

like that word. Yeah. Whenever

1:31:54

I say it in conversation, everybody

1:31:56

always says, my friends

1:31:59

still say, hmm, a note to me and I'm

1:32:01

like but you knew exactly what I meant and

1:32:03

it's like yeah. I feel like everybody knows it

1:32:05

yeah. Yeah exactly everyone calls it out

1:32:07

it's like a fancy word but you know

1:32:09

what I mean. So I

1:32:11

was hoping that there was another one but

1:32:13

actually there might be. Second high in

1:32:15

San. Yeah exactly second. And

1:32:18

the support was just like oh yeah it looks

1:32:20

like that survey just failed like we can spin

1:32:22

it out for you again but it's gonna take

1:32:24

like X hours and we were like right we're

1:32:27

getting off of Elastic we're going on to Bonsai

1:32:29

so that was a wild day because it was

1:32:31

eight hours of like well not quite eight hours

1:32:33

because we were slowly indexing the millions of records

1:32:36

again but it was like that was one

1:32:38

thing. I guess one of our biggest kind of

1:32:40

tech learning was when

1:32:43

we had the first initial spike you know

1:32:45

the response time graph on Heroku we'd

1:32:47

be like okay okay and then BAM it would

1:32:50

like spike up and we

1:32:52

tried so many things we architected

1:32:54

bits of code we even

1:32:56

at one point you know my friend Andy

1:32:59

Kroll he texted and said you

1:33:01

know you're very right heavy like which

1:33:03

database turn are you on and

1:33:05

at this time I was on the $50

1:33:08

database plan and Andy said well

1:33:10

that's never gonna work like you need to

1:33:12

upgrade like this is actually a case of

1:33:14

throw more money at it you're graded to

1:33:16

the $200 database you know

1:33:18

that fell at the time oh my gosh $200

1:33:21

a month on the database that felt very

1:33:23

wild I mean especially at this time it's

1:33:25

money from my pocket and eventually Rob's as

1:33:27

well. It got to the

1:33:29

point where even when we upgraded database it worked

1:33:31

fine for a time and then bang we had

1:33:33

the response time issues we were

1:33:35

looking at the logs we were trying to

1:33:37

go through everything and that's when we learned

1:33:40

about IELTS and

1:33:42

burstability like we didn't understand

1:33:45

we didn't know about any of this stuff so we

1:33:47

learned oh now we see

1:33:49

what's going on like our right IELTS

1:33:51

are like over what the database

1:33:54

can handle so we now

1:33:56

know we actually need to upgrade to this

1:33:58

plan whatever that plan was. Also,

1:34:01

the reason why it works for

1:34:03

a time and then fails is burstability.

1:34:05

It's basically giving you, you can go

1:34:07

this amount of time going over your

1:34:09

IOPS rate and then we're

1:34:11

going to say no. So that

1:34:13

was another, you know, so many

1:34:15

like learnings over the years of

1:34:17

like technical, like I said, sharding

1:34:19

with Elasticsearch, IOPS on the database,

1:34:21

like even just setting up Elasticsearch

1:34:23

and optimizing that. And

1:34:26

yeah, it's been a real amazing

1:34:28

technical journey for me. When I was at

1:34:30

Pivotal, you know, I always had my pair

1:34:33

and so I could always, there was always someone

1:34:35

and given I came straight from a boot camp, there

1:34:37

was always somebody who knew more than me sitting right

1:34:39

next to me. Or if they didn't

1:34:41

know more than me, they would, you know, kind of have a sense

1:34:43

of what to look for if we were stuck. And

1:34:46

a lot of this is just

1:34:48

like figuring out while

1:34:50

you're putting out the fire. So

1:34:53

yeah, there could be a whole other episode

1:34:55

on like technical fires

1:34:58

and learnings. Well, we might have to have you

1:35:00

come back at some point and tell us all

1:35:02

these stories because we didn't dive into the technical

1:35:04

very much at all. Go ahead. I

1:35:07

was going to say we're on the verge of something

1:35:09

that I'm scared of, but I'm less scared of now.

1:35:12

We're going to go for a distributed database because

1:35:14

we had all these issues at the beginning of this

1:35:16

year and we thought having a replica would be the

1:35:19

solution. But that's given

1:35:21

like the right heavy nature of the database

1:35:23

because essentially when we have a spike, it's

1:35:25

also a spike in Goodreads import. So

1:35:28

when we have 10,000 users on the app, there

1:35:30

are about 8,000 people that want to import their

1:35:32

Goodreads and we need to do that relatively quickly

1:35:34

while they're excited, while they're interested. How

1:35:36

long does the standard Goodreads import take? Are we

1:35:39

talking seconds, minutes? Minutes. But

1:35:42

we had people waiting for weeks this

1:35:44

January. Sure. Well, because you had a backup.

1:35:47

Yeah. And then all the 20,000,

1:35:49

I said to the world, wow, haven't seen these Q

1:35:51

numbers since June of 2020. And

1:35:54

are there API limits slowing you down there as

1:35:56

well? Like, can you hit Goodreads just as fast

1:35:58

as you can or they limit you? We

1:36:00

don't we don't touch good reads at all. We

1:36:02

don't touch good reads at all in-house stuff How

1:36:04

do you do your imports then? Okay,

1:36:20

I Want to know

1:36:22

how you're doing it? Will

1:36:24

you reveal it for our change like plus plus listeners

1:36:26

ears only? No Not

1:36:30

telling anyone this is yeah, tell anybody got a

1:36:32

lot of secret stuff. Okay. All

1:36:34

right It's mechanical Turk behind

1:36:37

the scenes making a phone call to good read

1:36:39

support acting as the user Hey,

1:36:47

I need my import list and you hung with that

1:36:49

please and they get emailed and I upload

1:36:51

it my name is Sally Yeah,

1:36:54

and I would like to export

1:36:56

I can talk about a decision I made early on

1:36:58

actually which was that I remember when I

1:37:00

started I never Used

1:37:03

good reads API because I always

1:37:05

said to myself if this thing gets big they

1:37:07

could just take away a pack So

1:37:10

from day one I said export your data

1:37:12

from good reads his set Give

1:37:15

me the CSV will sort it out and

1:37:17

I remember the beginning people were to complain

1:37:19

like say this is very funky All

1:37:22

the apps do it this way And

1:37:25

what happened two years ago you go good

1:37:27

we said we shine down our API now Well

1:37:30

smart that is how you do it to this day is CSV

1:37:33

import. Oh, yeah, we do is a CSV

1:37:35

that we're given. Yeah. Yeah, okay That

1:37:39

makes sense and then you

1:37:41

hold on to it for a few weeks and then eventually you do

1:37:43

something with it I Just

1:37:48

love it It

1:37:51

was just funny. I'm just messing with you.

1:37:53

I would love to have your problem I

1:37:55

mean you got serious problems like technical problems.

1:37:58

Let's not discount them. Those are You're the

1:38:00

fun ones. Like you are, it's baptism by

1:38:02

fire because you're having success and you're

1:38:04

having to learn and adjust and like

1:38:06

scramble. And a lot of us, you know,

1:38:08

we just put our stuff out in the

1:38:10

world and nobody ever uses it. And so we're just like,

1:38:13

we would love, we're just age topping, looking at our CPU

1:38:15

doing nothing. We're like, come on now. That's why I said

1:38:17

an enviable scenario, you know, even though at first you didn't

1:38:19

really necessarily want it because you were not

1:38:22

aligned, you thought, with your original

1:38:24

goals, but I mean,

1:38:26

you're having fun, right? Seems like you're

1:38:28

having fun. Overall I'm having the

1:38:31

best time. Like I love

1:38:33

my life. How amazing is this that I

1:38:35

had this idea, I had this little dream.

1:38:38

Do you remember when I said it wouldn't be, it

1:38:40

would be quite a stretch to say that

1:38:42

I never thought about a Guru's alternative in

1:38:45

those early like months. I

1:38:48

remember thinking, okay, I'm not building

1:38:50

Guru's alternative, but imagine if he did,

1:38:54

but we're not, we're not building it.

1:38:56

But we're not. We're not. I

1:38:58

have a little bit to say. So the

1:39:00

fact that we are now like the main

1:39:03

alternatives to goodreads, the team is so small,

1:39:05

especially me being like a bootcamp

1:39:07

grad, right? That's something else that

1:39:09

I always like to stress actually

1:39:11

because I'm a bootcamp grad. I

1:39:13

started coding when I was what, just graduated of

1:39:15

UD. So I was 21, 22. I

1:39:18

said wow, it's because I'm 32 now. So that's

1:39:20

like a decade. Yeah, you're

1:39:22

a senior dev. Yeah, that's wild. So, oh my

1:39:24

gosh, yeah, 24, 20, 13. I've

1:39:27

passed my 10 years of actually learning to code.

1:39:29

That's wild. Completely escaped

1:39:32

me. You've been

1:39:34

too busy. I've been too busy. I didn't

1:39:36

even realize I passed my 10 years of being

1:39:39

a dev anyway. You're a

1:39:41

senior dev for sure. Wow. So,

1:39:44

but I always like to stress that I'm a

1:39:46

bootcamp grad. My degree was

1:39:48

in philosophy, politics and economics. I

1:39:50

was going into banking. Like I'm not, I haven't been

1:39:53

coding since I was young. So

1:39:55

my point is like anybody can do

1:39:57

it. Like, you know what I mean? Like you

1:39:59

can, if. you want to do it, you

1:40:01

can do it. It

1:40:03

might take you a long time, but

1:40:06

you can do it. It is

1:40:08

possible. And you don't have to, you

1:40:10

could be listening to this and not even know how to code yet

1:40:13

and still end up building something in some years that. I

1:40:16

think you represent more than just you can do it as

1:40:18

a bootcamp graduate. I

1:40:21

think you represent that you can