Slack founder: Mental models for building products people love ft. Stewart Butterfield

Back in 2014, right as Slack launched, I told an interviewer that what we had was embarrassing and nowhere near good enough. If you can’t see endless room to improve, you probably shouldn’t be making the product.

Slack helped define consumer‑grade craft in business software, and that relentless standard shows. Before we dive in, quick table‑setting: today’s guest is Stuart Butterfield, co‑founder of Flickr and Slack, and someone I’ve wanted on the show for years; so first question, what are you up to post‑Slack?

Mostly family time and creative projects, with no new company to announce. I talk with Cal Henderson about possible ideas, do some philanthropy, and keep wondering how software might help people use phones less, but I haven’t cracked it.

People you’ve worked with keep mentioning your “utility curves” lens; how does that guide product decisions?

Think of an S‑curve where early effort yields little, then value spikes, then flattens. If a feature never reaches the steep part, you’ve only added complexity; so we asked where a feature really sits on that curve before investing more.

So it’s a way to prioritize work until users hit the “I can’t live without this” moment, at both feature and product level?

Exactly, and the bar keeps moving because everyone else improves too. Even small choices shape loyalty; take the painful time zone picker in a popular calendar app—it ignores context and makes a common task feel dumb, which erodes goodwill.

Let’s talk taste and Slack’s famous craft. How do you think about taste in product?

Taste can be trained, and leaning into it is a competitive edge because most teams underinvest. Our rule of thumb was “tilt your umbrella,” meaning practice tiny acts of empathy that remove irritation and make people feel cared for.

Give us moments where you chose craft over speed and it really paid off.

We used magic‑link style sign‑in on mobile to avoid password pain, nudged new users off noisy notifications once they saw a few messages, and added a playful warning for @everyone to curb spam and teach the cost of blasting a whole channel.

I remember that warning. You also rolled out Do Not Disturb thoughtfully.

We staged defaults, let admins and users override each other in a sensible order, and announced it early so nobody felt blindsided. The goal was to set healthy norms without breaking critical workflows.

You often argue that friction can be useful; what do you mean?

“Remove friction” is the wrong reflex when the real blocker is understanding. Slack’s site visitors often had thin intent and low comprehension, so clarity about what it is and what to do next mattered more than shaving a click.

Cognitive effort is expensive and emotionally punishing, so avoid forcing decisions. Optimize for trivial choices, not fewer taps; surface the common actions and tuck the rest under “other,” like an earlier Uber design did, and aim for the fluid, thought‑free feel you see in fast social apps.

That reframing—reduce thinking, not clicks—really lands. You seem wired for constant dissatisfaction with the status quo.

We embraced perpetual improvement, even posting my “this is a piece of junk” quote on the wall as a reminder. Clear, candid critique can be motivating when the culture trusts that it’s about making the product better.

You’ve also ranted that the default state of most systems is “not working.” What’s the lesson for builders?

Reality is multivariate and slow to change, like the years it can take to get a stop sign installed. Simple villain stories are seductive; inside companies that same urge blinds us to complexity and the hard grind required to make anything actually work.

You cite Parkinson’s law a lot; how does it show up in companies?

Work expands to fill time, headcount gravitates upward, and people invent work when known valuable work runs short. We once spent huge effort A/B‑testing a tiny threads tweak whose potential upside could never justify the analysis—classic hyper‑realistic work‑like activity.

So the fix is leadership creating clarity and saying no early, not scolding after the fact?

Right. It’s on leaders to define priorities, align people on what truly matters, and prune everything else up front.

What’s the backstory behind your “we don’t sell saddles” memo?

It was a pre‑launch note to align the early team that we sell outcomes and change how teams work, not features. You often have to create the market story, which means positioning with familiar ideas so people grasp what’s new.

On pivots, when should a founder stick versus switch?

Be brutally rational about expected value and whether you’ve exhausted realistic paths. Pivoting is humiliating and painful, but sometimes the best move; Slack benefited from years of internal use before we admitted the game wouldn’t work.

People also describe you as unusually generous with employees and customers; what drives that?

Partly how I was raised, and partly game theory—signal cooperation and others reciprocate. We ran the company on a simple chant: over time we win in proportion to the value we deliver, which leads to choices like fair billing and proactive outage credits, even when one version later proved too costly for a public company.

Those credits were automatic, right?

Yes, we applied them proactively and let people know, even aggregating downtime that might not have hit a given customer directly.

Quick curveball: your given name was Dharma, which makes your last concept feel full circle—what’s the “owner’s delusion”?

It’s when builders design for their own pride rather than user jobs to be done. Think of old restaurant sites with music and photo slideshows when all a diner needs is hours, address, menu, phone, and a way to book; always pause, pretend you’re a regular person, and sanity‑check that it makes sense.

Name the bias, bake that check‑in into your process, and avoid the trap. Stuart, this was special—thank you for doing it.

Thanks for having me. I really enjoyed it.

Thanks for listening. If you found this valuable, follow the show on your favorite app and consider leaving a rating or review; it really helps. Past episodes and more are at Lenny’s Podcast dot com.