Lenny's PodcastStewart Butterfield: Why thinking is what kills your product
Through Slack details like the shouty rooster and magic links; reducing thinking, not clicks, drives love, while utility curves tell teams when to stop.
CHAPTERS
- 0:00 – 5:05
Relentless customer value and the “giant piece of shit” standard
Stewart opens with a blunt Slack-era quote about being embarrassed by the product’s current state, using it to frame a mindset of perpetual improvement. The discussion quickly ties product success to the real value delivered to customers, not to internal narratives or optics.
- •Using harsh honesty as a tool to maintain high standards
- •Belief that designers/leaders must see “limitless” improvement opportunities
- •Customer value as the ultimate measure of success
- •Distinguishing pride in a component vs. complacency about the whole product
- 5:05 – 6:45
Stewart’s post-Slack life: family, philanthropy, and concern about phones
Lenny asks what Stewart has been doing since leaving Salesforce. Stewart describes focusing on family (raising a young child), exploring philanthropy and art, and wrestling with how software can reduce—not increase—phone dependence.
- •Leaving Salesforce and shifting to family life
- •Ongoing conversations about a possible next project with longtime collaborator Cal Henderson
- •Motivation to build software that helps people use phones less
- •Philanthropic and creative work as current focus
- 6:45 – 10:13
Utility curves: where product effort actually creates value
Stewart explains utility curves as an S-curve mapping effort/cost to value/convenience, emphasizing thresholds and diminishing returns. He applies the model to features to decide whether to double down, stop, or recognize that a feature is still on the ‘flat’ part of the curve.
- •The S-curve: slow start, steep value creation, then diminishing returns
- •Effort/value framing applies to features and product quality
- •Common failure: shipping a feature that never crosses the ‘magic threshold’
- •Using the curve to decide if you’re too early, at payoff, or past the point of returns
- 10:13 – 15:11
Divine discontent and the rising bar: iterating the “boring essentials”
Building on utility curves, Stewart introduces Bezos’ “divine discontent” to explain why standards keep rising as users acclimate and competitors improve. He argues teams underinvest in crucial flows like time zones, signup, and password reset—places where delight and trust are built or lost.
- •As products improve, user expectations rise and the target moves
- •Many orgs rarely revisit core workflows after initial implementation
- •Google Calendar time zone picker as a case study in neglected essentials
- •Delight creates emotional connection that drives advocacy and retention
- 15:11 – 16:49
Taste as a competitive advantage (and a learnable skill)
Stewart defines product taste as discernment that can be trained, not just innate talent. He argues most companies underinvest in taste, making it a durable advantage for teams willing to practice craft and empathy.
- •Taste can be developed through practice (like becoming a better chef)
- •Natural advantages exist, but training matters
- •Leaning into taste can create a durable edge because others don’t invest
- •Beware “over-optimizing” tiny details past meaningful returns
- 16:49 – 20:20
“Tilt your umbrella”: empathy, courtesy, and the mechanics of Slack’s delight
Stewart shares the origin of “tilt your umbrella,” a Slack cultural shorthand for being considerate of others’ experience. He connects small, thoughtful conveniences to Slack’s viral, cross-company growth—users advocating for a tool because it feels humane and respectful.
- •Umbrella story: most people don’t adjust for others; Slack should
- •Three explanations for inconsiderate behavior: malice, inability to imagine alternatives, or obliviousness
- •Product craft as empathy translated into interface decisions
- •Emotional connection drives word-of-mouth and cross-pollination adoption
- 20:20 – 28:31
Designing defaults and shaping behavior: magic links, notifications, @everyone, and DND rollout
Stewart walks through concrete product decisions where craft and behavior-shaping mattered more than raw speed. Examples include passwordless login links, onboarding notification defaults, discouraging misuse of @everyone with a rooster prompt, and a careful, admin-aware rollout of Do Not Disturb.
- •Magic link login to avoid painful mobile password entry
- •Onboarding tradeoff: default ‘notify on every message’ then nudge to recommended settings
- •The “shouty rooster” prompt to discourage noisy @everyone behavior
- •Complex DND rollout balancing admin control, user overrides, and safe defaults
- 28:31 – 38:34
Friction vs. comprehension: stop reducing clicks—stop making users think
Stewart argues the real enemy isn’t friction but comprehension—especially when user intent is low and the product is unfamiliar. He explains why minimizing taps can worsen usability, and why design should reduce cognitive load and avoid making people feel stupid.
- •High-intent flows (tickets/checkout) benefit from friction reduction; most product usage is low-intent
- •Comprehension includes: what this is, what to do next, and what will happen
- •Decision fatigue is real (metabolic + emotional cost)
- •Bad UI makes users feel dumb and permanently damages affinity
- 38:34 – 45:52
Simplicity through hierarchy: menus, “Other,” and learning from Snapchat’s speed
Stewart expands on interaction design: fewer taps isn’t the goal; easy, obvious steps are. He critiques cluttered action menus (Gmail) and praises designs that foreground the primary action with secondary actions behind “Other,” using Snapchat as an example where many taps can still feel effortless.
- •Expose the 2–3 most common actions; hide the rest behind “Other”
- •Gmail’s fragmented menus increase cognitive burden via search-by-elimination
- •Uber’s early “Where to?” simplicity as a model for clear intent
- •Snapchat: many taps are fine when each action is obvious and low-cost
- 45:52 – 50:13
Constant dissatisfaction and continuous improvement as an organizational norm
Returning to the theme of dissatisfaction, Stewart describes cultivating a culture where blunt critique is motivational and improvement is the default. He ties this to trust, Kaizen-like iteration, and the ability to treat mistakes as ‘gems’ that accelerate mastery.
- •The 2014 quote and internal reaction as a cultural artifact
- •Separating critique of work from critique of people requires trust
- •Continuous improvement as a shared goal across disciplines
- •Using feedback as fuel rather than as a threat to identity
- 50:13 – 54:31
Why nothing works: complexity, veto points, and the hidden work of coordination
Stewart shifts to a broader lens on institutions and organizations, arguing that the default state of the world is dysfunction due to complexity and layered veto points. He connects this to why simplistic blame narratives are appealing but often wrong, and why execution requires persistent coordination.
- •Anecdote about the years-long effort to install a stop sign
- •Political narratives over-attribute outcomes to villains vs. systemic complexity
- •Anthropomorphizing systems is natural but misleading
- •Organizational progress requires persuasion, alignment, and navigation of vetoes
- 54:31 – 58:43
Parkinson’s law and org growth: why headcount expands and harms focus
Stewart explains Parkinson’s law beyond time management, focusing on the tendency for administrative layers to grow regardless of core output. He argues most roles have incentives to expand their teams, which creates organizational bloat unless leaders actively constrain it.
- •Work expands to fill time; similarly, orgs expand independent of real need
- •Incentives: more direct reports correlate with status, pay, and authority
- •Mistaking growth for progress leads to complexity and inefficiency
- •Leaders must resist automatic headcount expansion and enforce clarity
- 58:43 – 1:08:32
Hyper-realistic work-like activities: fake work that looks like work
Stewart introduces “hyper-realistic work-like activities” as the outcome when organizations run out of “known valuable work to do.” He illustrates how A/B tests, dashboards, decks, and meetings can consume massive effort for tiny gains, and argues leaders must ensure a clear pipeline of high-value priorities.
- •Threads @-prefill A/B test as an example of huge cost for negligible upside
- •“Known valuable work”: you know what it is and you know it matters
- •As orgs grow, demand for ‘work’ can exceed real valuable work supply
- •Leadership responsibility: create priority clarity, say no early, prevent fake-work spirals
- 1:08:32 – 1:13:20
We Don’t Sell Saddles Here: creating the market by communicating outcomes
Stewart explains the memo’s origin as an early internal alignment tool before Slack launched. He emphasizes that for novel products, building the product and building the customer’s understanding of it are inseparable—positioning and narrative are part of product work.
- •Memo was written for an ~8–10 person team during private beta
- •Marketing isn’t optional when comprehension is the bottleneck
- •Positioning: combine existing concepts rather than invent totally new ones
- •Sell the outcome/experience (the ‘horse ride’), not the feature (the ‘saddle’)
- 1:13:20 – 1:18:33
Pivoting with cold rationality: exhausting real options and surviving humiliation
Stewart shares how he thinks about pivots: you persevere through good ideas, but fold when expected value collapses. He stresses emotional distance because pivoting is humiliating and disruptive—impacting employees, users, and communities—so it must be a deliberate, rational choice.
- •Slack’s advantage: years of proto-work while building Glitch
- •Pivot criteria: have you exhausted non-ridiculous long-shot paths?
- •The emotional difficulty: public commitments, layoffs, community loss
- •A pivot isn’t “changing ideas early”; it’s a high-stakes strategic fold
- 1:18:33 – 1:25:49
Generosity as strategy and ethics: cooperation, customer value, and where it breaks
Lenny surfaces stories of Stewart’s generosity toward employees and customers, and Stewart frames generosity as both innate and game-theoretically rational. He returns to his core mantra—creating customer value—and shares an extreme SLA policy that later became untenable at public-company scale.
- •Generosity as signaling cooperation (Prisoner’s Dilemma framing)
- •Chanted mantra: success equals value created for customers
- •Ethical practices attract better employees and reduce internal dysfunction
- •SLA ‘100x money back’ story: generosity can backfire at scale, requiring adaptation
- 1:25:49 – 1:30:35
The owner’s delusion: designing for real humans, not captive audiences
Stewart closes with “the owner’s delusion,” the tendency to overestimate how much users care and how much attention they’ll devote. He uses bad restaurant websites and confusing product naming to argue for empathy, external feedback, and continuous re-grounding in user reality.
- •Restaurant website anti-pattern: ignores the 5 things users actually need
- •Owners overvalue their own context; users are busy and ready to bounce
- •Naming/UX failures stem from assuming shared understanding
- •Practical fix: name the bias, get outside feedback, and re-check from a ‘regular person’ lens
