Here’s something I don’t understand: Software engineering is an art snd a science. But so is sales. So is marketing. Except sales can break its art down into numbers. It’s made itself *legible* to the business. Marketing is the same. - ThreadSky

hazelweakly.me • 57 days ago

Here’s something I don’t understand:

Software engineering is an art snd a science.

But so is sales. So is marketing.

Except sales can break its art down into numbers. It’s made itself *legible* to the business. Marketing is the same.

Reposted from Marco Rogers

This is one of my favorite spicy topics. I am one of those people who says that engineers don't actually want to be held to any evaluation criteria. They definitely want *other* people to be held accountable. But not themselves. And if given a choice, they'll choose nobody being held accountable.

Comments

hazelweakly.me•57 days ago

If you ask a Chief Sales Officer if they have a set of evaluation criteria for everyone in the org, along with a fairly unambiguous line of “here’s where we fire you” they’d say of course they do.

They can break those numbers down and show you where the money goes and how and why.

hazelweakly.me•57 days ago

Marketing can do the same! If you’ve ever asked a Chief Marketing Officer if they know what the most effective strategies they have are and if they understand how much their budget contributes to revenue, they’d say of course they do.

It’s right there. Tabulated and broken down in 74 different ways

hazelweakly.me•57 days ago

Then there’s engineering.

Can a CTO tell you the top sources of productivity in the org, or the top sources of waste? Have they down headcount into revenue effectively. Do they have evidence behind any of their metrics?

Is this knowledge transferable to any other company?

hazelweakly.me•57 days ago

Does anyone in tech have a well understood and articulated strategy for understanding when to fire people, when to promote them, when to invest in a strategic initiative, and when to divest?

Is that strategy intelligible to the rest of the industry? Using common layers of understanding?

No?

hazelweakly.me•57 days ago

Sales is an art. Marketing is an art. Every sale or campaign or effort is fully unique. You can’t cookie cutter any of it and you can’t cut corners.

Same is true for software.

But only one of those three seems to believe a fantasy that “art” is incompatible with “being understood by the business”

hazelweakly.me•57 days ago

By the way, here’s a shout out to @polotek.bsky.social: he’s been a delightful person to learn from and chat online with over the years.

He’s got a great knack for asking questions people *should* be asking, but don’t want to.

Thanks for the food for thought :)

auska.esq•57 days ago

There’s an important caveat in that many of the established business metrics are… reductive to the point of being harmful but it’s just accepted, and it’s not necessarily the kind of numbers we want to present.

But very much agree on need to solidify practice around getting the numbers we *do*.

adamhjk.me•57 days ago

I think your frame is a little off, Hazel. Yes, every marketing campaign is unique in its *content*, but not unique in its *form*. Software is usually unique in both, once you get beyond "how many engineers are typing".

kateviolette.com•57 days ago

"We owe it to our buddies in other departments to be understandable and accountable to the business" is one of my most common coaching angles for teams I manage.

You can get that message through!, it just takes thoughtfulness and repetition (and not hiring jerks to begin with)

phillipcarter.dev•57 days ago

Microsoft did a study a while back that a lot of people panned but it basically boiled down to:

All the criteria you see in the discourse around good engineers matter, but the most important thing is that for roles where you code a lot, the more code produced, the better the engineer

phillipcarter.dev•57 days ago

Key thing is that was not a metric in isolation as many in the discourse have wrongly accused the study about, and it’s my belief that this was itself a bit of collective cope, no matter how well intentioned.

hazelweakly.me•57 days ago

Yeah my suspicion with that study is that LoC ended up being the most "legible" metric in the organization for a whole slew of other factors and had almost nothing to do with directly correlating to the engineer's skill.

But, y'know, I'm biased (and can back it up with not-microsoft research)

phillipcarter.dev•57 days ago

The other spicy take to consider is how the “ghost engineering” thing is actually not entirely bullshit because it claims 1/10 engineers at Big Tech barely do work, and BOY OH BOY that is definitely not bullshit.

Now the fact that monied interests seek to exploit this is another matter entirely

cail.bsky.social•56 days ago

Ah, but are their numbers predictive? Causative? Or just happen to move the right way at the right time?

Given you mention ‘legibility’ have you read Seeing Like A State? I think the challenge here is not how to build better mechanistic models of work, but changing the management mindset

skamille.themanagerswrath.com•57 days ago

Marketing? The same business with the tagline “Half the money I spend on advertising is wasted; the trouble is I don't know which half.”? Marketing may produce metrics but the idea that they are clearly metrics-driven is.... not at all true in any of my experience

skamille.themanagerswrath.com•57 days ago

To be clear I agree with the point engineering should measure themselves more, I just think that part of the reason people resist it is that the pressure turns to measuring repeatable stuff that can be compared over time which is mostly just productivity divorced from value

hazelweakly.me•57 days ago

Oh I totally agree (and your point about marketing is also super valid!)

I've seen so many organizations run into the ground because they developed a data driven strategy that turned into "{palatable} quantitative metrics at all cost" strategy

Which, of course, gets you nowhere useful

elchefe.me•57 days ago

This is ESPECIALLY evident when you see software engineering organizations that use “velocity” as their guiding star, rather than quality or efficiency.

elchefe.me•57 days ago

ime, most tech marketing jobs are metrics driven except the CMO. This hasn’t always been the case, but has become very much the case in the past few years. I’ve seen this working as and with marketing teams as a DevRel, especially when you have to justify your budget.

danciruli.cloud•57 days ago

When I (a product manager) was a GM of a $100M business and got deep insight into the sales team, I was *blown away* at the dashboards that they had and the fidelity of the information. They were SO GOOD at tracking progress week-to-week, with hard numbers.

I've never seen the like in product/eng.

hazelweakly.me•57 days ago

It’s one of the reasons why I’m so stubborn in insisting that engineering teams focusing on observability need to bring the entire organization along for the ride.

Engineers need to be seeing that sales and marketing tracking and need to be making their observability legible to the business as well

hazelweakly.me•57 days ago

“Oh we can track error rates per endpoint now”

That’s sweet, can you track error rates per customer bracket and prioritize fixing the issues of the largest clients first? No?

Hmm, that makes it real hard to justify working on long tail work when the larger clients have surface issues, doesn’t it?

danciruli.cloud•57 days ago

Yes!

danciruli.cloud•57 days ago

By requiring every single sales person to update things like deal size, closing date, and likelihood of closing (with a hard formula for calculating the latter), our CRO had an amazing view into how the quarter was going.

I've never seen dashboards like that from an eng team.

danciruli.cloud•57 days ago

Sure, I've seen burndown charts. And ideally you're really story pointing well and can predict your velocity.

But wow -- the fidelity of the data that our CRO got from the sales tools was amazing.

k17.dev•57 days ago

Because most directors do some sales and marketing. We inherently understand a lot of what they entail. We learn about them in an MBA.

Most directors have no idea what any kind of technical specialist does, and none know all.

k17.dev•57 days ago

This is a generalisation, but sales love a leaderboard - it motivates them to try harder and that ranking is usually closely related to profit.

These sorts of leaderboards can be utterly toxic for technical specialists.

k17.dev•57 days ago

Folks don't understand what software engineers are actually doing, so let's pick another speciality: doctors.

You can use mortality rate as a KPI, right? The best doctors save more people...

k17.dev•57 days ago

Only then you get lots of specialists in non-life-threatening conditions, and nobody working on acute deadly conditions.

I need the neurosurgeons and acute oncologists on my team. I also need the knee specialists and cosmetic surgeons.

I don't want them to compete, it's counterproductive.

k17.dev•57 days ago

Sometimes, I need to put my top developers in the toughest trenches, where they'll do better than anyone else could but have worse metrics. I need to keep them motivated while more junior devs get the easy work.

k17.dev•57 days ago

Yeah, they're accountable, they just don't break down to numbers someone outside can easily understand.

juliaferraioli.com•57 days ago

I find that software engineers, in aggregate, are bad at identifying metrics without gaming the metrics becoming the goal.

I'm not saying we're bad at stats; I'm saying we're very good at lying to ourselves with data.

juliaferraioli.com•57 days ago

Inevitably, making the metric the goal and having it be gamable leads to the predictable conclusion of a bad outcome, which results in cognitive dissonance...

Because if we're good at stats and good at software engineering, why does this happen???

juliaferraioli.com•57 days ago

Totally ignoring the psychological aspects.

It's "safer" to handwave.

I don't agree with this conclusion, but I do get it. It's fear based.

hazelweakly.me•57 days ago

I really love what you’re saying here. I wonder if the handwaving comes from how unpredictable thought work is and how difficult it can be to make that legible to the rest of the business?

I’d imagine that would result in missed expectations to the point that the theater comes in, perhaps?

hazelweakly.me•57 days ago

Being good at lying to ourselves with data would totally exacerbate this, I think

juliaferraioli.com•57 days ago

I think there's a really complex discussion to be had about this and I hope we can catch up to have it!

jeffreyhulten.com•57 days ago

I wonder how sales and marketing would react if their organizational goalposts were moved like tech goalposts are. I wonder how you retain salespeople when you tell them "just make 40% of the sale and we will come back to the rest later."

elchefe.me•57 days ago

tbh, I blame this failure on leadership/culture. If you are using metrics as a bludgeon, you're going to get gamifications.
Metrics should be lagging indicators of behaviors, and it's the function of leadership to encourage and promote those behaviors without using metrics as the end goal.

hazelweakly.me•57 days ago

Hmm, I wonder if a deeper understanding of how to use metrics and do this fuzzy data qualitative analysis and reasoning exploration is a big thing holding us back from being more effective here?

( @grimalkina.bsky.social I’d be curious of your thoughts on this too 👀 )

grimalkina.bsky.social•57 days ago

Me looking at the "healthy measurement" paper I haven't yet finished posting to an arxiv 😐😐😐

Lots of interesting stuff here (more than I can do justice to in a post!!). I would venture this is essentially the same struggle we're having about representing human needs plus achievements in all realms

juliaferraioli.com•57 days ago

Yes yes yes

elchefe.me•57 days ago

I think the main problem is the fundamental lack of understanding people have of human behavior, motivation, and fear + absolute lack of empathy because none of them can be put on a fucking powerpoint slide, so we use metrics instead because someone needs to justify their MBA debt.

I have opinions.

juliaferraioli.com•57 days ago

Agreed. But most people I've chatted with identifying metrics don't know what a lagging/leading signal is.

juliaferraioli.com•57 days ago

See also goal culture based on buzzwords.

"How much delight did we deliver?"

"What % increase did we have on time to ship over last quarter?"

"Do our customers AI the AI things???"

elchefe.me•57 days ago

sounds like they shouldn't be leading/managing anyone

ringtheory.bsky.social•57 days ago

It's because leadership doesn't agree on the prioritization and importance of the work before started. So if project A is more important than project B, then it shouldn't matter that B comes in with more lines of code or project A takes longer to deliver.

The bare bones metrics are not comparable.

aptakhin.name•57 days ago

This is interesting. I followed mostly @davefarley77.bsky.social definition “Engineering is the application of an empirical, scientific approach to finding efficient, economic solutions to practical problems”. But now it applies to marketing engineering too, if we go empirical way:)

productnerd.bsky.social•57 days ago

On helping the business understand the value of our work in progress or what's coming up next, is this not the point of product management?

I feel like when there's a push for metrics when it comes to product delivery, it's always metrics on productivity, as if engineering is an assembly line.

hazelweakly.me•57 days ago

I think it depends on how the company views engineering and what their worldview is around thought work in general. But typically, at some point in the process, engineering *is* considered an assembly line.

It’s easy to do, given how new and poorly understood software is, but it doesn’t go well :)

productnerd.bsky.social•57 days ago

Yeah, one of the first things I do once I start trying to make things better is to help people understand what the different parts of the organization do. It's easy to make fun of sales/marketing/product/ops/dev, but building that empathy opens access to a much better conversation all around.

productnerd.bsky.social•57 days ago

That concept of "Together, we are the team" goes a long way to being able to talk about more than just raw number metrics. ^_^

jjminter.bsky.social•57 days ago

This is honestly one of the most difficult things for me as developer who now has to make a resume and market myself. I keep getting told that I need to break my projects down into concrete numbers. How many labor hours did I save, how does that translate into dollars for the organization?

jjminter.bsky.social•57 days ago

Tbh, idk because I'm doing this to save Mary from having to copy and paste spreadsheets. I want to make Asher's job easier by automatically sending this notice. How am I supposed to quantifying the quality of work that happens when you give your team breathing room?

joemonti.bsky.social•57 days ago

Performance metrics are an art across each of those disciplines. Sales and marketing also have the advantage of being around since the dawn of commerce, while Software Engineering is a baby by comparison. Software Engineering has also been burned by ridiculous metrics like lines of code written.

brittonbroderick.bsky.social•57 days ago

@krishnablue.com has been writing about this one a lot, and this is one I think a lot about.

One key is sales and marketing have activity metrics that can be tied back to probabilistic revenue outcomes, which (afaik) software doesn’t have.

I think you could get closer with product in the mix.

krishnablue.com•57 days ago

But agree that it has very little to do with whether these are art/science/intellectual work etc.

krishnablue.com•57 days ago

It’s not generically true of software. project based delivery, outsourced development, has clear unit economics.

So it depends on the business context.

Products are generally the harder case.

Internal IT development is similar.

Comments

Posting Rules

Reply