All podcasts / Lenny's Podcast / Summary

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

2025-11-20 - 90 min - source - Read full transcript

Lenny Rachitsky (host)Stewart Butterfield

Key insights

Product value follows an S-curve (utility curve), so the question for any feature is where on that curve you sit, not whether you have the feature.

Butterfield describes value against invested effort/quality as flat at first, then a steep rise past a threshold, then flattening again. Many teams treat features as binary (have it or don't) and keep adding complexity below the threshold where it starts mattering, which adds cost without adding value.

product-craft-and-taste

User quality expectations keep rising even for parts of the product you never touch, a dynamic Butterfield calls divine discontent.

Because users experience better search, login, and checkout elsewhere, their bar for your search, login, and checkout rises continuously. Features that get implemented once and rarely revisited (account creation, forgot-password flows) fall behind even without changing.

product-craft-and-taste

Positioning a product around the outcome it enables, not its features, is what Butterfield's 'We don't sell saddles here' memo argues for - and it means you are creating a market, not just a product.

Drawing on the book Positioning, Butterfield argues it's nearly impossible to install a genuinely new idea in someone's head; it's far easier to combine two familiar ones ("Uber for pets"). If your product differs meaningfully from alternatives, you must sell the transformation (horseback riding), not the component (the saddle).

product-craft-and-taste

Reducing friction is the wrong default goal - the actual obstacle for most products, most of the time, is comprehension, not the number of steps.

Friction reduction only matters when a user's intent is already high and specific, like buying Taylor Swift tickets. For most new signups, intent is barely above threshold and vague, so the real job is helping people understand what the thing is and what to do next - Butterfield estimates comprehension is 70-80% of product design.

friction-vs-comprehension

Minimizing the number of clicks or taps is often exactly the wrong optimization; grouping choices into a few obvious options plus an 'Other' submenu beats flattening everything onto one screen.

Early Uber's two-option home screen (destination or Other) is the example: eight easy, low-stakes taps beat two taps that each require comparing many options, because comparing many choices to each other is cognitively expensive in a way that adds up geometrically as the option count grows.

friction-vs-comprehension

Users suffer from the 'owner's delusion' - builders forget that users show up distracted, impatient, and with minimal context, not sitting attentively like the owner who lives inside the product every day.

Butterfield frames this via a crowdsourced Twitter reply ("the owner's delusion") to his own observation about unusable restaurant websites: makers assume visitors want to engage deeply, when in reality most people are one confusing screen away from bouncing.

friction-vs-comprehension

Organizations grow headcount independent of real work demand because career incentives reward managing more people, not because managers are lazy or dishonest.

Citing Parkinson's 1955 essay and Royal Navy administrator data, Butterfield notes that promotion, pay, and internal authority all track team size, so a 27-year-old PM will reflexively want a direct report even with no clear task for them - a dynamic that shows up in every budget cycle where headcount rarely shrinks voluntarily.

organizational-bloat

As known-valuable work runs out, organizations backfill with 'hyper-realistic work-like activities' - meetings, dashboards, and analyses that are structurally identical to real work but produce near-zero value.

Butterfield's example: Slack spent enormous engineering, analytics, and meeting time re-litigating whether to pre-populate an @-reply in threads, based on a statistically shaky 0.03-message difference in thread length - a guaranteed loser once the cost of running the analysis is weighed against the tiny possible upside.

organizational-bloat

If a problem involving a large organization looks simple to solve, you probably don't understand it - Butterfield's own coinage, 'Butterfield's Law.'

He states it as "everything is simple if you have no idea what you're talking about," and uses it to explain why no CFO or engineering head ever proposes shrinking headcount in a budget cycle: the complexity is real, not a failure of will.

organizational-bloat

Real pivots require exhausting every realistic option and making the call with deliberate emotional distance, because staying in a failing venture out of pride is the more common failure mode.

Glitch had $9 million left and a team that still enjoyed the work when Butterfield shut it down - the decision came only after exhausting every non-ridiculous path to commercial success. He distinguishes this from early-stage repositioning ("three of us pivoted after six months" doesn't count as a real pivot) and calls the decision humiliating precisely because founders have convinced investors, employees, press, and users to commit to the original vision.

pivoting-and-quitting

Acts of generosity function as strategic, iterated-game moves, not just personal values - they signal cooperation that invites reciprocal loyalty from employees and customers.

Butterfield frames Slack's paid-in-full employee health insurance, no-lockup direct listing, COVID credits, and proactive downtime refunds through the lens of the prisoner's dilemma: consistently visible cooperation encourages others to cooperate back, producing better outcomes than defecting first. He is candid that this generosity had a real cost: the 100x-money-back SLA policy triggered an automatic ~$8 million credit after a major outage post-IPO, forcing a change to the terms of service.

generosity-as-strategy

Books referenced

Don't Make Me Think - Steve Krug - Source of Butterfield's core mantra - reduce the thinking a product demands, not just the number of clicks or amount of friction.
Positioning - Al Ries and Jack Trout - Cited as the classic on why it is easier to combine two existing ideas in someone's head ("Uber for pets") than to create a genuinely new one.
Why Nothing Works - Recent book Butterfield calls repetitive but substantively important, on the accumulating veto points that make anything hard to build or ship.
Principles - Ray Dalio - Source of the Michael Jordan ski-lesson story Butterfield uses to argue criticism should be treated as a gift, not an insult.
Quit: The Power of Knowing When to Walk Away - Annie Duke - Duke uses Glitch's pivot to Slack as her example of a 'smart fold' - abandoning a venture once its expected value has dropped below the next best alternative.

Companies

Slack - Butterfield's second company, sold to Salesforce; source of most of the craft and product examples in the episode.
Flickr - Butterfield's first company and first pivot, also from a video game (Game Neverending).
Salesforce - Acquired Slack in one of the largest tech acquisitions at the time.
Google - Repeatedly criticized for poor comprehension design in Gmail (recipient field, action menus) and Google Calendar's alphabetical time-zone picker.
Stripe - John Collison and Patrick Collison referenced; Patrick's fixation on a small Gmail UX flaw used as an example of obsessive product taste.
Uber - Early Uber app's two-choice menu (destination or 'Other') held up as an example of comprehension-first design.
Snapchat - A teenager's fluid, rapid-tap usage cited as proof that removing friction is the wrong goal when users already have full comprehension and intent.
Canva - Melanie Perkins' 100-plus investor rejections raised by the host as a persistence counterexample to Butterfield's pivot advice.

Techniques and frameworks

Utility curves - An S-curve of value versus invested effort/quality; early investment yields little value until a threshold, after which value rises steeply then flattens - used to decide whether a feature is worth further investment.
Divine discontent - As users get used to better software elsewhere, their quality bar for every part of your product keeps rising even if you never change it.
Owner's delusion - Product builders assume users bring the same attention, patience, and context to the product that the builder does, when most users arrive with minimal intent and are ready to bounce.
Parkinson's Law (organizational form) - People hire more reports not from evil or stupidity but because headcount correlates with career growth; supply of hires outpaces genuine demand for work.
Hyper-realistic work-like activities - Butterfield's term for meetings, dashboards, and analyses that look exactly like real work but exist only because staff have run out of known-valuable work to do.
Butterfield's Law - "Everything is simple if you have no idea what you're talking about" - if a large-organization problem looks simple, you don't understand it.
We don't sell saddles here - Butterfield's early internal memo: sell the outcome (horseback riding) rather than the feature (the saddle) - you are creating a market, not just a product.
Tilt your umbrella - Slack's internal riff on Bezos's 'your margin is my opportunity' - most competitors have bad manners/craft by default, so small acts of consideration compound into differentiation.
Shouty rooster - A Slack UI pattern that interrupted @everyone messages with a warning showing how many people/time zones would be notified, which sharply reduced overuse of the feature.

Summary

Stewart Butterfield, founder of Flickr and Slack, gives Lenny Rachitsky a rare, wide-ranging interview built around mental models he used to run both companies. The through-line is a rejection of two common product instincts: that features are binary (you either have them or don't) and that removing friction is always good. Instead, Butterfield frames value on a utility curve - an S-shaped relationship between invested effort and payoff - and argues that most teams either underinvest below the curve's inflection point or keep polishing well past the point of diminishing returns, without ever explicitly diagnosing which is happening.

A large portion of the conversation is spent on comprehension versus friction, illustrated with a string of specific, sometimes profane examples: Google Calendar's alphabetically sorted time zones, Gmail's inconsistent action menus, an iPhone alarm app's inscrutable "Sleep" toggle, and Slack's own "shouty rooster" warning for @everyone messages. Butterfield's consistent claim is that reducing clicks or steps is the wrong lever almost everywhere except high-intent flows like checkout or ticket purchases; the real cost is forcing users to make decisions they don't understand, which he ties to both a literal metabolic cost (glucose burned making decisions) and an emotional one (users blame themselves, not the software, when they're confused). He credits Steve Krug's "Don't Make Me Think" as formative and describes Slack's "we don't sell saddles here" internal memo - never actually published externally under that framing beyond the well-known blog post - as an early attempt to get a tiny founding team aligned on selling outcomes rather than features, citing the marketing classic "Positioning" as backing for why net-new ideas are hard to plant and combining familiar ones is easier.

The episode's second major thread is organizational dysfunction, framed through Parkinson's Law and Butterfield's own coinage, "hyper-realistic work-like activities." He argues headcount growth is driven by career incentives rather than malice or stupidity - more reports means more status and pay - and once an organization runs out of known-valuable work, people fill the gap with meetings and dashboards that look identical to real work but return almost nothing. His central example, a months-long analytics effort at Slack to justify pre-populating an @-mention in message threads based on a barely-significant 0.03-message length difference, is presented as a cautionary tale about how much organizational machinery can be spent proving something that was never going to matter. He extends this into "Butterfield's Law": problems involving large organizations that look simple to an outsider almost never are, which is why budgets rarely shrink even when everyone agrees they should.

On pivoting, drawn from his own history at Glitch (formerly the game Game Neverending, which became Flickr) and again at the game Glitch (which became Slack), Butterfield argues for treating the decision with deliberate emotional distance rather than either premature abandonment or stubborn perseverance. A real pivot, in his framing, only happens after exhausting every realistic path and requires founders to accept a humiliating admission after having convinced investors, employees, and users to commit to the original vision - he notes Annie Duke used the Glitch-to-Slack transition in her book "Quit" as an example of a rational "smart fold."

The conversation closes on generosity, which Rachitsky frames as the most consistent theme other people raised about Butterfield: paying 100% of employee health insurance, a no-lockup direct listing, COVID-era customer credits, and an automatic 100x-money-back SLA for downtime. Butterfield connects this to game theory - visible cooperation in an iterated prisoner's dilemma invites reciprocal cooperation - while being candid that the generosity had real costs, including an approximately $8 million automatic credit triggered by a major post-IPO outage that forced Slack to revise its terms of service.

Notable Quotes

"If your software kind of stops me and asks me to make a decision, and I don't really understand it, you make me feel stupid." - Stewart Butterfield

"Everything is simple if you have no idea what you're talking about." - Stewart Butterfield

"The reason I say you have to be coldly rational about it is because it's fucking humiliating." - Stewart Butterfield

"In the long run, the measure of our success will be the amount of value that we create for customers." - Stewart Butterfield

"The acts of generosity to me are a way of demonstrating that I am going to cooperate as we iterate in this game." - Stewart Butterfield