Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
[Music]
0:15
Everything can be improved, iterated and refined and if you don't think that's true,
0:19
maybe you haven't analyzed it enough. Calculator choices carefully considered. Absolutely analytical.
0:25
Analytical is part of the Engineered Network. To support our shows, including this one,
0:28
head over to our Patreon page and for other great shows visit engineer.network today.
0:32
Activity. Are you working today? Are you being productive today? Are you demonstrating activity
0:40
today? You might think activity is a thing that others need to see, but I wonder if that's really
0:48
completely true. When we combine produce with activity we get productivity. On balance though,
0:57
It's more about expectation over a certain time horizon to get a result.
1:03
So, if we were to look at an activity that a major corporation, let's take for example Apple
1:10
because they're a good example, why not? And let's say that they release a new phone every year or
1:15
maybe an incremental update every year and then a major update every two years. So, when people look
1:22
at what Apple is doing in the meantime between those two sorts of release dates, every 12 months,
1:27
to 24 months, there's a lot of speculation, a lot of guesswork, a few leaks of information
1:33
sometimes from the inside. But ultimately, though, there's very, very rarely any actual
1:39
evidence that there's any activity happening. There's no proof that anything is going on.
1:47
We all assume, we all know, but there's no evidence to suggest it. There's no external
1:52
visibility that anything is happening. So, so far as we know, Apple's not doing anything.
1:57
Well, we know that that's not true though, because after 12 months or 24 months, we see
2:00
the results. So we know, and we can infer in the interim that activity was happening.
2:06
When the next device is released, it's probably amazing. People love it. Well, probably, maybe
2:11
they do, maybe they don't, but faith is restored. There was activity all that time. Of course
2:17
there was and we didn't need evidence of that to get a good result at the end.
2:22
So our expectations are set because we understand implicitly and we're trained by observing
2:29
a cycle of productivity that these things, like making a new phone, takes a certain amount
2:35
of time, effort and energy and that's normal and that's okay.
2:40
And a lack of visible activity isn't indicative of all of a lack of productive progress.
2:48
So as we sort of break that down, it comes to the underlying trust that actions are occurring.
2:55
So let's consider the opposite extreme.
2:58
We're going to look at now micromanagement and metric driven reporting.
3:02
Let's say that you're a new manager coming into a business and you're trying to understand
3:06
what a team is actually doing.
3:09
You have no long-term view, no real time horizon, no set expectations, no evidence of a productive
3:16
output and hence no evidence of an end result.
3:20
On the surface, the team isn't visibly doing much, it's not producing much visible activity
3:26
and the pressure is on to demonstrate that activity is happening.
3:31
To do that, the new manager sets some activity-based KPIs and starts reporting those KPIs to the
3:37
next layer of management above them. The team shifts from a trust-based system of work handover
3:43
and progress internally to pushing to close bug reports faster, open change control requests,
3:50
formalize interactions with other team members, and everything starts to become very transactional
3:58
and takes on a transactional overhead. The problem is that overhead isn't free and that
4:03
overhead costs, let's say 25% of their overall efficiency, where that previous energy spent
4:09
doing the doing is spent proving the doing and their final delivery is proportionally
4:14
late or hours worked by the team increases to compensate for that. So that overhead costs
4:20
about 25% of your overall efficiency where previous energy that was spent doing the doing
4:26
is instead spent proving the doing and then the final delivery will be proportionally
4:32
late or the hours that they're working increases to compensate for all that extra overhead.
4:37
The manager though is demonstrating that activity in the team is happening and their management
4:42
above have great visibility that activity is happening but it's come at a huge cost.
4:49
So trust, trust is eroded, interactions in the team and between teams become transactional
4:57
and impersonal, actual productive work is slowed down, the result is actually a net
5:03
loss overall. Morale starts to drop, people become more distant, more hostile, and then
5:10
there's an unhealthy tension that builds between teams and silos begin to be built for some
5:17
perceived self-preservation or superiority versus inferiority or this team is doing better
5:23
than that team, so we're not going to help them, they're not going to help us.
5:27
That works directly against teamwork in the context of the whole organization.
5:34
The infection of KPI-driven or metric-driven activity measurement seems to come from ignorance,
5:41
I think, and a belief that humans can be evaluated as discrete productive entities with different
5:48
levels of performance for different activities.
5:52
That kind of mindset completely ignores the fact that humans are complicated.
5:56
Invisible work like tidy up activities, a desire to help others in need is a real thing,
6:03
along with learning curves. And humans, we change focus based on our outside lives and pressures and influences.
6:10
That changes any equation that people might have and could possibly ever document to say,
6:15
"This person can produce this amount of work in this amount of time," because all of it
6:19
is extremely variable. But ultimately, if you're going down the path of metric-driven
6:25
activity measurement, and there's a push down the chain to demonstrate that movement is
6:30
occurring in a team, and activity is happening, and the longer that time horizon is for those
6:37
final deliverables, the more the pressure will build. You get comments like, "Oh, are
6:43
Are they still working on that? It's been two weeks, it's been two months, it's been a year.
6:49
When in fact it's always taken a year between phones, just like every other year for the
6:55
past decade.
6:58
Your ignorance of the past is not a measure of the team's performance, it's just a measure
7:04
of your own ignorance.
7:07
All it takes is a lack of trust or even a so-called respectful challenge of why is this
7:12
taking so long? Can we go faster? So that drives evidence-based management versus trust-based
7:19
management. Now don't misunderstand, I just think it's very important to acknowledge that
7:26
cost tracking is important, sure. Milestone reporting is also important and I think reporting
7:34
up the management chain about roadblocks and disruptions is very, very important.
7:40
But I do think that reporting on activity alone is entirely and completely pointless on every
7:47
single level. And yet it seems to persist, probably because it's easier for some people
7:54
to quantify and understand, despite the fact that it doesn't actually add any value.
7:58
Now, you may be wondering, I'm talking in very general terms, why don't we talk about something
8:05
specific. So let's talk about an example in programming terms. One of the activity-based
8:10
metrics is bugs closed over a time period. I mean it could be that we could all just as easily be
8:15
talking about service tickets and so on. It doesn't really matter but let's go with bugs closed over
8:20
time. So assume there are three teams and all of them are working on different sections of the code
8:25
maybe different code libraries, different controllers physically, it really doesn't
8:28
matter. The business though, it's trying to determine performance between those teams to
8:34
grade them and by measuring activity metrics. Team A closes 5 bugs, Team B closes 10 and Team C closes
8:42
50 bugs and that's all over the same time period. So Team C, 50 bugs? Well they're the highest
8:48
performing team aren't they? Well maybe they are, maybe they're not. What if I actually told you
8:54
that Team A objectively was the highest performing team? Because when we deep dive into it, well it
9:02
It turns out that Team A worked for 25% more hours
9:06
than the other two teams did, and resolved issues that had caused system crashes
9:10
and catastrophic failures.
9:13
And both teams B and C had been trying
9:16
to fix them for months, and they'd failed many times to resolve those issues.
9:20
Turns out that Team C's bug fix count list
9:24
consisted of 30 of them, which is typographical errors.
9:27
And the balance of those bugs they closed very straightforward, low priority bugs that really didn't add any specifically huge value
9:35
and yet they counted against their metric and it looked amazing but it just wasn't. So the problem
9:43
with the metric you might say is actually the lack of depth of detail since not all bugs are
9:47
created equal in terms of effort to resolve them and their impact to the system which is true
9:52
but then we reach that point how do we honestly objectively and fairly assess
9:58
just how difficult a bug is to resolve. Sometimes we measure priority of bugs like a P1, priority
10:04
one, you know, can't ship without it, P2 can wait for the next service pack and so on, but estimating
10:08
effort to fix a bug is extremely difficult even for experienced developers. There's always an
10:14
element of interpretation, professional judgment and balance to determine which team was actually
10:19
the highest performing and ultimately only the manager that technically understands what their
10:24
team is doing is able to interpret this and if they can't understand it then their only option
10:31
is to metricize their team's work. Of course there's other options, maybe they don't want to
10:38
spend the time to understand, or as in like they're too lazy, or maybe their management layer above
10:44
demands that they must have a metric. So when I come across the situation of activity measurement
10:50
as a measure of output. The things that strike me are, first of all, A) the manager either doesn't
10:57
understand what their team does or B) they're too lazy to do their job or C) they've allowed
11:02
their management to push irrelevant activity metrics upon them. For all these reasons or any
11:09
of these reasons, they shouldn't be in management. Try something else like cross stitch or something.
11:15
Thinking of it like a mathematical equation just for a second.
11:18
Productive plus activity equals productivity, which is what we care about. Whereas,
11:25
activity minus productive equals, well, just activity, which is a waste of time,
11:33
a waste of money, just a waste of everything.
11:37
If you're enjoying Analytical and want to support the show, you can via Patreon at
11:43
at patreon.com/johncheejee or one word.
11:46
With a thank you to all of our patrons and a special thank you to our silver producers
11:49
Mitch Bjorgar, John Whitlow, Joseph Antonio, Kevin Koch, Oliver Steele and Shane O'Neill.
11:55
And an extra special thank you to our gold producer known only as R.
12:00
Patrons have access to early release high quality ad free episodes as well as bonus
12:03
episodes and this is done by Patreon.
12:06
Visit engineer.network/analytical to learn how you can help this show to continue to
12:10
be made. Of course, there's lots of other ways to help like favoriting the episode in your podcast
12:15
player app or sharing the episode or the show with your friends or via social.
12:19
Some podcast players let you share audio clips of episodes, so if you have a favorite segment,
12:23
feel free to share that too. All of these things help others discover the show and can make a big difference too.
12:29
You can follow me on the Fediverse at [email protected], on Twitter @JohnChigi or the network at engineered_net.
12:38
Accept nothing. Question everything.
12:41
always a good time to analyze something. I'm John Cheerjee. Thanks so much for listening.
12:46
[Music]
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More