0:00

The fourth quarter's here, and it's time

0:02

to hit revenue targets and finish strong. But

0:04

before you can win Q4, you need

0:06

to close deals. Before you can close

0:08

deals, you need easy-to-use, all-in-one

0:11

sales software like HubSpot Sales Hub.

0:14

The new and improved Sales Hub is where sales

0:16

teams can prospect more efficiently,

0:19

grow revenue faster, and mine

0:21

deeper customer insights. It's smart

0:24

software for smart sales teams who

0:26

want to close the year strong. With

0:28

Sales Hub, Q4 will never be the

0:30

same again.

0:31

Learn more at HubSpot.com slash

0:34

sales.

0:39

Hey folks, today we have Andrew Ettinger

0:41

on the show, a gentleman that's been a friend for

0:43

a number of years. He's kind of my go-to guy

0:46

on DevOps and infrastructure.

0:48

Pretty unique sales motion. He's

0:51

been a sales leader across

0:54

six or seven companies.

0:56

He was the one that took Pivotal

0:58

from $0 to $500 million in ARR, and

1:01

it was eventually sold to VMware. Another

1:04

one he did was Astronomer, where

1:06

we're going to unpack that a little bit and sort of the

1:08

early signs of community-led growth. And

1:11

more recently, he's joined Appen as

1:14

CRO, a $300 million

1:17

revenue company out of Australia that's publicly

1:19

traded, as he tries to help

1:21

the team pivot to more of a software

1:24

product-led company. All right,

1:26

here's my conversation with Andrew. We

1:31

got a lot to cover today. I do want to

1:33

start with Astronomer. Give us the 30 seconds

1:35

on what Astronomer does, and then tell

1:38

us a little bit of what

1:39

the context was like there when you joined.

1:41

Yeah, happy to, Mark, and thanks for that. I

1:43

think like most folks on your show, I like to build, I

1:45

like to fix, I like to grow.

1:47

And it was one of those opportunities in your career

1:50

where there was an open-source community that was

1:52

vibrant but also fractured at the same

1:54

time and needed to be repaired

1:56

and restored for growth.

1:58

At the end of the day, it was an an open

2:00

source company around Apache

2:02

Airflow, which very simply put was data orchestration.

2:05

Basically connecting all your disparate parts inside

2:07

of your modern data stacks, your data gets from point

2:10

A, right to point B in a safe, secure

2:12

and reliable way. So when I joined,

2:14

it was like an opportunity to say, hey, look, how many times

2:16

in your life do you get to go join a top level Apache

2:19

project, all the vanity metrics

2:21

of 20,000 plus stars and a million

2:23

downloads, et cetera, and really actually

2:25

give that some new energy. There were about 20,

2:27

25 people in the company and

2:30

sub a million dollars in revenue when I started.

2:32

So we went from that sub million to the mid

2:35

twenties, which included launching a cloud

2:37

product and taking some other revenue that

2:39

were, you know, in other channels

2:41

down. So it was actually more growth than that, but it was a

2:43

really exciting time and companies continue

2:45

to do absolutely fantastic under Andy Byron

2:48

and team and they're crushing it right now. Awesome.

2:50

You said coming in, it

2:53

was a little fractured, the community.

2:55

Can you tell us what that means? Like, how

2:58

did you know it was fractured? How do you measure

3:00

that? And what do you think was the cause of it? Yeah,

3:02

look, again, you know, I think the developer

3:04

is very powerful and they have a very powerful

3:07

voice. And there was a number of kind

3:09

of core and foundational features that

3:12

had been asked to be included in the

3:14

project for a number of years. It just never materialized

3:17

into a packaged upgrade

3:19

and piece of software that folks could use. Namely,

3:22

the scheduler, which made all of these things

3:24

work was not highly available. So

3:26

as you saw to put this into your

3:28

mission critical workloads, you couldn't have a single

3:30

point of failure. Something just very basic,

3:33

such as that. And there were a lot of promises made

3:35

that weren't fulfilled. And so ultimately

3:37

to astronomers credit, the engineering team really

3:39

delivered on that promise. And that really kind

3:42

of gave a rebirth to

3:44

the growth of the project. And a lot

3:46

of hard work went into that and further, but that

3:48

was really it. There were just a lot of skeptics that maybe in

3:50

fact, there was a new way of doing things because

3:52

there was not enough attention paid to some of

3:54

the legacy features that were required. So

3:57

this is a pretty unique context. We haven't really

3:59

unpacked. on the show and that

4:02

is this very developer centric open

4:04

source context. As a newcomer

4:07

to that, as a revenue person, as

4:09

a salesperson, I would be a little freaked

4:12

out. Developers, it's a culture

4:14

of its own and I don't, generally

4:16

speaking, I might say that developers

4:19

don't have a ton of respect for salespeople

4:22

per se. Was that the context

4:24

there? How did you navigate that? How would you

4:26

recommend a sales leader coming into

4:28

a context like take that on? Yeah,

4:31

well, I think, Mark, you hit the nail on the

4:33

head. Look, I've had a lot of history before this

4:35

with open source developers at Pivotal and Spring

4:38

Boot and that whole movement. So I had a little

4:40

bit of experience here and as a

4:42

result, I built technical chops and capabilities

4:44

that I didn't have in my career. And

4:47

so the first thing to your point, you

4:49

cannot do is sell and you have to add

4:51

value. And so obviously

4:54

I show up looking like a salesperson, probably

4:56

talk like a salesperson and probably

4:58

I'm not that incognito there. But once

5:00

you start to add value and actually

5:02

show that you can help and understand

5:05

and have empathy to solve the

5:07

problems that they have and show them the way you

5:09

start to earn the credibility and the respect.

5:12

And I think it all comes down to understanding

5:14

the user journey that they're on, because

5:16

once you understand that user journey, you can

5:19

understand the pain that they're in, in

5:21

their environment, their situation, ultimately the impact

5:23

you could have. So it still comes down to the basics of

5:25

fundamental selling and qualification. You're

5:27

just doing it a much different way. That's much more

5:29

technical led than sort of sales

5:32

and qualification led, if that makes sense. It

5:34

does, but can you unpack value a little

5:36

bit? Because I'm going to channel

5:38

a new seller or leader coming into this context.

5:41

It almost feels like I

5:42

have to go

5:43

spend three years in some sort of computer

5:45

science department to go

5:48

toe to toe with these folks. How can I possibly

5:50

add value to a developer

5:53

who's been coding for 15 years when I've never

5:55

touched code? Or did you have to go and

5:57

literally learn to code a little bit? I

5:59

would say. that I had to learn to code, but

6:01

I did learn how to use the product.

6:04

And I did learn how at the

6:06

end of the day, if you're a developer

6:08

and you're deploying something to production and

6:10

being able to view and look at logs and

6:13

metrics in a dashboard that could tell you

6:15

the health, the state, and how things

6:17

are running in production, you could very easily

6:19

show them how that might be something that would add

6:21

value to them that was previously

6:24

impossible in the open source world, or

6:26

that they would have to spend their precious time

6:28

coding, configuring, managing, upgrading,

6:31

supporting, and scaling out to the rest of the enterprise.

6:34

And that at the end of the day was not something

6:36

that a data engineer and that core persona

6:39

was really interested in doing. They were interested in

6:41

doing their daily lives of getting data ready

6:43

for the organization to consume and for business

6:45

leaders to make informed decisions and actions

6:47

to run the business off of. Okay. So if

6:50

I know like medic or I know like challenger

6:52

sale, something off the shelf, it

6:54

feels like I have to customize

6:56

that to this buyer context.

7:00

Like what falls flat if I'm just like

7:02

a straight out medic, implementer,

7:05

and I did that in whatever healthcare, finance,

7:07

whatever, and now I'm going into the developer environment.

7:10

How does that have to be tweaked? Yeah. Well,

7:13

for starters, most of the developers in

7:15

the open source world don't have any money. They don't

7:17

actually care about your metrics. They don't actually care

7:19

about being qualified and worse.

7:22

They actually have no idea how software gets procured

7:24

inside of their organization. At the end

7:26

of the day, you have to implement a process

7:29

such that they understand that you can

7:31

make their life better. And once you

7:33

do that, you've earned the credibility to then ask

7:36

for the next step on how you

7:38

would get to their boss, what this would

7:40

mean and how you would articulate the value. And

7:42

we've learned this the hard way, right? There were plenty

7:45

of times I called a Casper the ghost. You

7:47

thought you did a great job. You let them go

7:49

sell on your behalf to their boss and

7:51

guess what you see in Salesforce close

7:54

unresponsive. And so we had to really

7:56

radically change our engagement model around

7:59

situation. main impact that would

8:01

then frame up sort of where they

8:03

were trying to go and be very prescriptive about

8:06

how the engagement model worked and

8:08

earning the right and the credibility to

8:10

then take that next step together with

8:13

them and enabling them with the tools

8:15

and us and our technical personas

8:18

to be able to communicate with their VPs and heads

8:20

of engineering on the value that we could provide

8:23

and so it was a delicate dance mark there's no doubt about it.

8:25

How be unpacked that is like a sales

8:28

coach or manager so it's like okay

8:30

you're listening to a first

8:33

meeting between one of your reps and these

8:36

this engineer and the engineer

8:38

is loving it like oh my gosh this is great I

8:41

had no idea you've showed me exactly the

8:43

value I'm gonna take this to my VP

8:45

engineer and I need this product and

8:47

the engineer and this salesperson is like okay

8:49

that's great can I join the meeting

8:52

I'd like to be part of that and

8:54

the engine is no no no my VP

8:56

of engineering is extremely busy I'll just handle

8:59

this on my own and your salesperson

9:01

is like okay that's not acceptable

9:04

from what you're saying so how do you handle

9:06

that objection of like getting

9:08

the power in this context? Yeah it's

9:11

a very good question various student

9:13

mark and look at the end of the day

9:15

we did a lot of that early on and we learned our lessons

9:18

right and they were it was variable

9:20

obviously hard to predict hard to forecast and hard to

9:22

understand and so what we did was we learned

9:25

that if we drove them into the product which

9:27

was not product-led growth it was sales

9:29

led growth into the product and we got

9:31

them to deploy literally in a few hours

9:34

in a workshop a few of their pipeline

9:36

so that we we could show their

9:39

management their code working in this cloud

9:42

we drove an engagement process that we were

9:44

upfront and unapologetic about is like look

9:46

here is our process here's how we engage

9:49

here's your investment and here is what

9:51

we would like to do in return if we prove these things

9:53

like does this work for you and we were

9:55

able to attract them because we were willing to

9:57

invest the time in the product with their working

10:00

code and our engineers to solve

10:02

the problems that they had. And

10:04

ultimately because you're making their life 10 X

10:07

better and because you're giving them their time

10:09

away from focusing on the undifferentiated

10:12

heavy lifting required on a lot of developers

10:14

these days, they ultimately were ready to

10:16

take you there. Hey folks,

10:18

just Mark here. I stopped the recording. Yeah,

10:21

this is an extremely abstract,

10:23

common and very challenging

10:26

objection in sales that

10:28

I suppose we'll categorize as getting to power.

10:32

It's coming up more often because historically

10:34

in sales, we always call power

10:37

and due to some of these new motions

10:40

like product led growth or even community led

10:42

growth, like we're talking about here, it's

10:44

actually beneficial to engage the

10:46

non decision maker,

10:48

the people closer to the front line first,

10:51

and then use their excitement and their

10:53

support

10:54

to get to the decision maker. Now,

10:57

in most situations like this one with

10:59

Andrew, the engineer, even

11:02

though they're so excited for this product

11:04

would prefer to go to their VP

11:06

of engineering boss themselves

11:09

without the salesperson, they know

11:11

the boss, they know she is very busy and

11:14

they want to be very protective of her time.

11:17

And most of the time that's going to be a close

11:20

lost. And here's the reason why. Engineers

11:24

are not salespeople, and this is true in almost

11:26

any other product you're trying to sell. They're

11:29

going to position the product in how it benefits

11:32

them, not how it optimally benefits the

11:34

VP of engineering. And that's

11:37

what we have to get through first and foremost by getting access ourselves.

11:41

So Andrew walked through a couple techniques that he

11:43

uses with his team to get that access. I'll

11:46

give you two more. One that I like to

11:49

do is I like to say, okay, you know, first off, just ask, like, do you mind if I join

11:52

you in the meeting? Quarter of the time that

11:54

might actually work. And if they

11:56

say no, I'd prefer just to take this might be a good

11:58

time.

11:59

myself, then I'll

12:01

ask them, okay, great, just walk me through

12:03

how you're going to explain the product to them. And

12:08

there's a lot of different variations that you can

12:10

take. But essentially, after

12:13

they do so, you express some concern. You

12:16

express some concern like, hey, I

12:19

know you really want this product, but I

12:22

do this every day. And if I were

12:24

your VP of engineering, I'm not sure I would be that

12:26

excited based on how

12:28

you just walk me through the product. Are you sure

12:30

I can't just be sitting there in the room with you

12:32

or sitting on the Zoom with you to walk them through

12:34

it? I'm going to take another stab at it. And

12:37

if they absolutely deny me, now

12:40

I am sales coach. I'm

12:43

going to build the slide with

12:45

them that they're going to walk into

12:48

the office or walk the VP of

12:50

engineering through on Zoom. I'm

12:52

going to explain to them and build out

12:54

the content that positions my

12:57

product in the way that resonates best with

12:59

the VP of engineering to make sure that this

13:01

doesn't stall, that I can get through the

13:03

stage of the opportunity. Really

13:05

common objection. And thank you for Andrew

13:07

for burning this up. All right, let's get back to Andrew. Now,

13:11

yes, were there times in which there was friction

13:13

and it just felt like, of course, just like

13:15

any perfectly run medic process, it doesn't work 100%

13:17

of the time, right? But

13:20

what that gave us was a far greater

13:22

level of predictability. And most importantly,

13:24

when you look at the consumption metrics, right,

13:27

those were the ones that ultimately wound up consuming

13:29

more, which meant their net retention rates went up,

13:32

right. And as you very well know, and preach better than

13:34

anyone in the world, that is the best leading indicator, right,

13:36

of a healthy software company. And so sometimes

13:38

there's a little bit of pain involved in getting

13:40

there, but it's necessary if you want to have right,

13:43

a sustainable business. Do you think that

13:45

the shift in the last decade or two

13:47

of like, if we went back 20 years ago,

13:50

the VP of engineering was doing more top down,

13:53

they were like, this is going to be my tech stack,

13:55

this is going to be our developer environment. And

13:58

people can sort of just conform.

13:59

And

14:00

have we shifted to a more

14:03

engineering first culture where

14:05

it's like almost like a servant leadership,

14:08

you know, like as the VP of engineering, my job

14:11

is to make my engineers as productive

14:13

as possible. And if they're servicing cool

14:15

tools that will help that they like, that shows

14:17

that productivity, that's my job

14:19

is to clear that space and make that happen. Has that

14:22

flipped a little bit? I absolutely believe

14:24

so, Mark. And, you know, I speak with many

14:26

C-level executives and they do reach

14:28

down into their organization to find out the trends

14:31

and what's happening. And I think there's this

14:33

old saying that goes, don't ever bet against open source.

14:35

And it's because it is so easy. But

14:37

at the end of the day, these developers are powerful.

14:40

They're smart. They spend their spare time

14:42

evaluating these things and participating

14:44

in these communities. And it's a really nice leading indicator

14:47

of the future of, you know, what products,

14:49

particularly in cloud infrastructure and

14:51

data are going to rise to the top. Community

14:54

led growth has become a bit

14:56

of an acronym, you know, like product

14:58

led growth. I think it's about maybe

15:01

a five year delay in a way. I'm quite intrigued

15:03

by it. Would you consider this motion

15:06

that you all used as a astronomer as community

15:09

led growth? Yeah. I mean, look, anytime

15:11

you're an open source, like your first job

15:13

is the community. That's it. Period.

15:16

The end and without a vibrant community, you don't earn the

15:19

right to actually monetize anything. And

15:21

so the first strategy was called the three C's, which

15:23

was community content conversion. And

15:25

if we didn't do all of the right things to engage

15:27

the community, which is a combination of

15:30

obviously solid technical leadership and

15:32

future velocity and development

15:35

into the core open source, but

15:37

also rich documentation, rich

15:39

libraries, rich certifications,

15:42

right? Where we hired the top person from Udemy to

15:44

run these classes and gave away 25,000 free certifications

15:48

to go do that. All of the right content

15:50

that HubSpot and others made very famous

15:52

where it was like, look, you're looking to integrate this with snowflake

15:55

or Databricks and provide all of that with

15:57

no paywall, no emails. I don't

15:59

want any. from you just use all that information.

16:02

You have to combine those two things before

16:04

you have the right to even think about converting

16:07

them into some paid product or into

16:09

a conversation around where your value

16:12

can materialize itself. Hey

16:15

folks, just Mark here. Yeah, I'm just

16:18

loving this journey with

16:20

Andrew right now. I'm loving learning from

16:22

him. You know, there's some

16:24

big movements at play

16:26

here. This is getting back to empowered

16:29

buyers. I

16:31

think that's partially what drove the

16:34

interest and success of product-led

16:36

growth, empowering buyers to go

16:39

try product. And I feel like community-led

16:41

growth is coming fast. This

16:44

is a new marketing mechanism.

16:47

This is a new go-to-market motion that's

16:49

going to be just as disruptive and

16:51

just as powerful as product-led growth. It's

16:53

obviously crawling out of the open

16:56

source, arena, and

16:58

I believe it's applicability beyond

17:00

that domain, beyond DevOps, beyond infrastructure,

17:03

to almost any type of B2B software is

17:05

going to be powerful.

17:07

So I love digging in here with

17:09

Andrew to learn these concepts like

17:12

community,

17:12

content, and conversion. To

17:15

really dive into the uniqueness

17:17

of dark social, a concept

17:19

that's coming up here, but it's becoming ever

17:22

increasingly important

17:23

to marketers,

17:25

to be at the front of

17:27

when the demand and awareness cycle happens

17:29

and get visibility into that and nurture it

17:32

and create a moat at the marketing side

17:35

for your business, through your community. All

17:37

right, this is important. Let's hear more from Andrew on this

17:39

concept. And

17:41

so we really worked very hard on those three Cs,

17:44

but really the first two to earn the right for the third.

17:47

So define that for us, because I think as PLG

17:49

took off, I mean, it was kind of called freemium

17:52

before, and then people called

17:54

it PLG and that really took off. And it

17:58

was a little difficult to... to

18:00

find the boundaries of where PLG

18:03

sat. And after staring at this for a long

18:05

time, for me, it was basically

18:09

if I could use your product

18:11

and extract value from your product without

18:14

talking to a human, whether

18:16

it was for seven day trial, whether it was forever

18:19

without having a payroll, that to me

18:21

was a product led growth motion. How

18:24

do you define community led growth? Like where

18:26

is the, is there a clear boundary there? Yeah,

18:30

I mean, look, PLG is, as you suggested,

18:32

and many folks have done a world-class job

18:35

against that and instrumenting for the aha moment

18:37

and when users convert into teams,

18:40

and then you can go after the enterprise and all of that. And

18:42

I think that's well defined and well pioneered

18:44

by many of the grades that are out there. I

18:46

think on the community side, there's a lot

18:48

of adjacencies there. So you have the core community of

18:50

like the personas that you serve, where

18:52

you want them speaking, evangelizing

18:55

and talking to their peers about using

18:57

your technology or your set of technologies.

19:00

But then you also have the spheres of influence

19:02

and sort of the adjacencies like the VP of engineering,

19:04

right? In this case, or the CIOs or

19:07

the heads of business or the folks that

19:09

run, you know, your cloud data warehouses

19:11

in this example. And so you have a multitude of these

19:13

communities that ultimately then can intersect, right?

19:16

And that's much different than PLG. PLG

19:18

is like just getting the product at all costs. And then our

19:20

system takes over. The communities are

19:22

a lot more intricate and a lot more

19:25

complicated and intersect in many

19:27

different ways. And in fact, actually

19:29

running both strategies is really powerful.

19:32

It's tricky. But if you think about it, actually

19:34

executing against that community process

19:37

into a PLG motion gives you perfect

19:39

nirvana. Because in theory, then they get into your funnel

19:41

a lot more educated, a lot more willing

19:44

and sort of already down that path versus just,

19:46

oh, I landed here. Let me check this out and see what it's like.

19:48

Let's stretch the boundaries on that and

19:51

go way outside this context where it's typically

19:53

applied. You know, there's just a lot

19:55

of companies, for example, a lot of entrepreneurs right

19:57

now, creating SDR copal.

20:00

technology, something that our audience can relate

20:02

to, right? So basically these are, you

20:05

know, AI, AI co-pilots where it's like, okay,

20:07

I as an SDRI go out, I find 50 accounts

20:10

a month, I create sequences,

20:13

I execute them, I manage connect

20:15

calls, I set appointments, whatever. And

20:18

you know, a lot of that can be streamlined,

20:21

automated using AI. As

20:24

an entrepreneur in that use case,

20:27

can I use community-led growth? Absolutely.

20:30

At the end of the day, you just still

20:32

have to balance, right, the boundaries

20:34

of these technologies like anything else, whether

20:36

you're using community-led growth in that AI context,

20:39

or you're using, right, those AI tools just

20:41

to automate outreach or other sequences.

20:44

You need to still ensure that you have that

20:46

personalized touch, and you don't come

20:48

off as a robot and a machine because

20:50

it's very disingenuous. And I've used

20:52

some of these things and seen some of these messages

20:54

from Mike. I'm not really sure if the

20:56

person on the other end is going to look at this and

20:59

actually believe that I'm a human. Now, some of them

21:01

are good enough, and so I guess my long-winded

21:03

answer is to say yes, but you absolutely

21:05

need to be careful like you would with any other applied

21:08

use of that technology.

21:11

If

21:13

you're starting a new job, we're getting a fresh

21:15

start at work. There's a new podcast

21:17

for you. It's called New

21:20

Here from our friends at the Harvard Business Review.

21:23

Every Wednesday, host Eleni

21:25

Mata talks with the career experts

21:28

and friends who share stories,

21:30

tips, and advice. They will help

21:32

you build a meaningful career on your own

21:35

terms. Episodes cover

21:37

networking authentically, recovering

21:39

from mistakes at work, handling

21:41

a bad boss, getting

21:43

your first race, and more. New

21:46

Here is the young professionals guide

21:48

to work and how to make it work for

21:51

you.

21:52

Listen for free at New Here from

21:54

HBR, wherever you get your podcasts.

22:03

What new roles come

22:05

up in taking a community-led

22:07

growth first mentality? Because

22:10

if I draw an analogy to PLG, you

22:13

know, in B2B software,

22:16

the PLG movement created this growth

22:19

role.

22:19

It kind of like blurred the

22:22

boundaries between product and marketing

22:24

and brought them together. This

22:26

sort of growth role would own

22:28

this human-less funnel. They would run

22:30

experiments like crazy. They would

22:33

be data-driven.

22:34

These data scientists became popular.

22:37

They often had data analysts on the team.

22:40

They were product managers working with engineers,

22:42

working side-by-side with paid digital

22:44

designers and marketers.

22:47

What's the analog to that in

22:49

the community side? Do we have community managers

22:52

now? How does this

22:54

evolve the roles in the team?

22:57

In the open source world to developers

23:00

is developer advocates. You

23:02

can argue whether they belong in product,

23:04

whether they belong in marketing, they shouldn't

23:06

be in sales. They certainly can be under a CRO

23:09

org if you have enough of the wall set up. Their

23:12

job and their team's job is simply to go

23:14

out there and help the community. It

23:16

does drive sales professionals crazy

23:18

when they find out that Mark was at a conference

23:21

and spoke with Jane. Then Jane's in charge of

23:23

data engineering at JP Morgan or whatever

23:25

company. You didn't sell sales? What

23:27

are you doing, Mark? It's like, no,

23:30

no, no, that's actually the worst thing you could do. There's

23:33

definitely friction there, but it is necessary.

23:37

Revenue leaders in that space and incumbent

23:39

on us to give those teams the space

23:43

to breathe and to foster those relationships

23:45

because it is something that compounds

23:48

over time into material value for your

23:50

organization. It's not a one-hit

23:52

wonder. You've got to be comfortable with that. It's hard

23:54

and it takes time. That's the flywheel

23:57

that ultimately gets going there. You can choose

23:59

to invest in one. or 10 or 20 or 30

24:01

of those and how you scale that out, but it's

24:04

a developer advocacy. This is

24:06

intriguing, Andrew. I want to make

24:08

sure we have time to get to your latest role,

24:10

because there's a whole new set of go-to-market

24:13

challenges. So Appen,

24:15

Rockstar Company, your

24:17

CRL, very different

24:19

context that you're walking into. Pay

24:22

that for us and what's ahead for

24:24

you in this role? Yeah. So look,

24:26

we publicly trade a company in the Australian Stock

24:29

Exchange, our revenues, $330 million a year.

24:31

Appen is so exciting. And on one hand, it's the forefront

24:37

of where AI is being developed and

24:39

consumed. And on the other hand, it's a very mature

24:41

company that's been around for a while, helping

24:44

companies create the necessary

24:46

data to train their AI model. We've had

24:49

those tools forever that Appen

24:51

in a more services-led approach. And now we're combining

24:53

those to be much more product-led with services

24:56

that are enabled around that. That's really exciting, given

24:58

the rise of LLMs and all the conversation we

25:00

find ourselves in. So how

25:03

does a $300 million plus

25:05

services

25:07

consulting business run? And in

25:11

parallel, you are launching and

25:13

growing a product business within that. That

25:15

must create some operational

25:17

challenges and tension. That's

25:20

a very good question, Mark. Look, at the end

25:22

of the day, this is a set of tooling that

25:25

we've run 24,000 AI

25:27

projects over the last 15 years on

25:30

our own tooling. Some of our customers

25:32

use that tooling. And it's really just

25:34

to say, hey, look, we've done that

25:36

many projects with 50 plus million

25:38

man hours on these tools that

25:40

now we're just going to expose out

25:43

to folks to be able to use. Because what we're finding...

25:45

I don't know why

25:48

more entrepreneurs don't start

25:50

as a services business and then

25:52

turn into a product business. Yeah,

25:55

there are some issues. I think I do know

25:57

why. And

25:58

that is because...

26:00

Back in the day, we used to think that services

26:02

businesses were VC fundable

26:05

and they became a really Tarnished

26:08

business model, especially for venture

26:10

capitalists. They weren't as scalable And

26:12

so I think a lot of people have walked away from

26:15

pursuing a services business and

26:18

yet I think it's an amazing foundation

26:21

upon which to build a product business

26:24

You are essentially getting paid

26:27

by the market to do your market research

26:30

You are not sitting in an

26:32

insular room Hypothesizing

26:35

on where the demand is you are out there Solving

26:38

real problems for real customers

26:41

driving value and understand how

26:43

it works You are seeing the patterns from

26:45

client to client to client You are

26:48

in such a better position to

26:50

architect and get right from the beginning

26:52

your minimal viable product as you transition

26:55

Now

26:56

there are certainly challenges in that

26:58

as you make that transition But

27:01

I wish I saw more entrepreneurs start

27:04

in their pursuit of their problem They're solving with

27:06

a services orientation and

27:08

then use that to transition into

27:10

product. All right, let's get back to Andrew

27:14

It's like in the case of a large financial

27:16

institution If you're gonna create an LLM

27:19

in sort of wealth management and you're looking

27:22

at 401k rollovers The

27:24

experts on that data and their call

27:26

center transcripts and your internal acronyms

27:28

and how your software works are best held

27:31

internally and so you're going to need these

27:33

tools and these platforms for your internal

27:35

people to go create all of

27:37

those workflows and Collaborate with each

27:39

other and the tools today and the option that they

27:42

have our spreadsheets. They have wikis You

27:44

have Jira tickets you have SharePoint

27:46

portals Then there is no streamlined

27:48

way to help them mitigate their

27:50

risk of that data going to the wrong

27:52

people's hands or the wrong things Happening. There's

27:55

no tools in place to help them actually accelerate

27:57

their time to value in into production

28:00

with this massive backlog and

28:02

at the same time increase the effectiveness

28:04

and the efficiency in the collaboration across

28:06

a dozen plus personas that go involved

28:09

into creating a simple LLM product

28:11

into production. Because if you look at the lifecycle

28:13

of personas involved, it's a lot and

28:16

they need a platform to go work on and collaborate.

28:18

And we've been doing that for a long time. We just

28:20

have to reposition it so people understand that it's

28:22

a really exciting spot to be in. Yeah,

28:24

it's a path of entrepreneurship I wish was

28:27

attempted more frequently to

28:29

really understand the market through the services

28:32

and then productize it once you understand

28:34

where the needs are. But I

28:36

don't want to undermine the

28:38

go to market transitions that need to happen.

28:40

So let's just talk for a second. You're

28:43

mentioning the bank in this wealth management group.

28:46

What did you do? Are you just giving the existing

28:49

account executive this product

28:51

now and they're bringing it or is this

28:53

a different type of seller and motion

28:56

to make this happen? Very good

28:58

question, Mark. And again, to your point of entrepreneurship,

29:01

maybe it's just actually called a glutton for punishment.

29:04

I'm not sure. But these are fun and interesting

29:06

problems that I find myself working on. And

29:08

this AI journey is actually no different.

29:11

And the common thread here is a really technical

29:13

sell, a real technical engagement,

29:15

having to hire, you know,

29:17

the professionals that interface with these

29:20

personas as peers. So instead of

29:22

a traditional sales engineer, you know,

29:24

I'm hiring data scientists and machine learning

29:26

engineers that go and partner very deeply

29:28

with these customers, run the onsite

29:30

workshops, whiteboard their problems

29:32

with them, help them understand the data that's

29:35

necessary here, and help to build those

29:37

prototypes with them so they can understand

29:39

in working code in the platform, how

29:41

this would work, which allows us to create

29:43

the business case, and the opportunity

29:45

for them to see the light and how this can have

29:48

a dramatic impact across your organization.

29:51

Yeah, we're swinging back

29:53

to these very technical

29:56

roles.

29:57

Yeah, a lot of us live in B2B SAS and

29:59

And there's

30:01

kind of been a progression from

30:03

like 25 years ago of like, hey,

30:06

can the technology even do this? And

30:08

we did have technical people sit inside by side

30:11

to go head to head with the CTO to suddenly

30:14

like, yeah, we pretty

30:16

much know how this tech is going to work.

30:19

It's well understood. It's relatively commoditized.

30:21

And it's more about the go to market.

30:24

It's more about the brand. And it's more about the category

30:26

creation that's

30:27

going to drive the success of the business.

30:30

Like I haven't sat around and be like, oh,

30:33

is that SAS business even technically

30:36

feasible? That hasn't come up for me in like a decade.

30:40

Here comes AI. And here comes those

30:42

challenges again.

30:43

Is this even technically feasible?

30:46

How does this work? What kind

30:48

of security implications does this have?

30:51

I don't know how to adopt this stuff. So

30:53

here comes the technical sales support

30:55

staff back. And a lot of us

30:57

haven't had experience

30:59

in that in a while. Just a heads

31:01

up.

31:02

This is not unique to Appen. This

31:04

is not unique to DevOps. If 90%

31:08

of B2B software sales is going to be

31:10

AI

31:11

in the next five or 10 years, this

31:13

is coming, especially in the early phases.

31:16

So pay attention to Andrew's guidance here

31:19

and think about it if you're going in that

31:21

direction with your product. All right,

31:23

let's get back to Andrew. And

31:26

so it's really just pattern matching from my last 12 years

31:29

in different areas of this space. And

31:31

arguably, this is the largest one of them all inside

31:33

of AI and really excited about it. Okay,

31:36

that's fair. Now be a coach

31:38

that as someone who has experience in there. So

31:41

I'm sitting here as CRO of

31:43

a pretty big team. And I want

31:46

to benefit from SDR co-pilot

31:48

technology. But my legal team

31:50

is saying, no, we're not ready. We need

31:52

another six months to get our act together.

31:55

And on the other side, there's this startup

31:57

seller who's like,

31:58

hey, we just raised money if I can.

31:59

huge valuation. We need some pilots

32:02

going on here.

32:03

And the CRO is like, I love this. I

32:05

want this.

32:06

But my legal team has given me

32:08

the Heisman for

32:10

six months.

32:11

What is your recommendation as a sales leader,

32:14

as a sales coach to get through that? The

32:17

thing that we did successfully in some of my

32:19

earlier journeys we spoke about, right, was

32:21

we just ran these small pilots and then engaged these

32:23

teams earlier. Again, it's counterintuitive,

32:26

but the sooner you get security involved or risk or

32:28

legal into these engagements, the

32:30

better. And at least you understand where your objections

32:32

are a lot sooner. It's not a fun saying.

32:35

If I'm going to lose a deal, I'd certainly rather lose it in 30 days

32:37

and 90 days or the last day of the quarter when I had it

32:39

forecasted. But again, blockers

32:42

still exist. This is new. There

32:44

are no standards and your mileage will vary

32:46

depending on person to person and company to

32:48

company. But that's sales 101, right? I

32:50

was so excited, Andrew, when I

32:52

heard sales leader, Caliber,

32:55

like yourself was jumping

32:57

into Appen to help them with this exciting

33:00

transformation and to help

33:03

lay the groundwork for this AI

33:05

revolution that's coming. So I'm going

33:08

to watch your story closely. I'm

33:10

sure it's going to be a fun one. And I

33:12

really appreciate you joining us today to drop

33:14

knowledge. Thanks, Mark. It was a pleasure.

33:17

I appreciate it. Thank you.

33:23

Today's

33:25

episode is written and produced by Matthew

33:27

Brown. Our show is edited by Pizza

33:30

Shark Productions. Big thanks to

33:32

HubSpot for startups and to the HubSpot

33:34

Podcast Network for keeping the audio on.

33:36

Hey, also we're a new show. So

33:38

if you like what you hear or if you hate what you hear,

33:41

leave us a rating and review over on your favorite

33:43

podcast site. I love the feedback. Also

33:46

check out stage two capital. We're the first

33:48

VC firm running back by

33:50

over 500 CROs, CMOs,

33:52

CCOs.

33:53

So if you're an entrepreneur looking to scale your

33:55

business, check out stage two.capital. All

33:58

right, that's it for today. I'm Mark Rabea. to

34:00

see you next week.