Before It Clicked

How Raj Suri Got the Idea for Presto — and Validated It by Working in a Restaurant

Raj Suri built Presto after one painfully slow restaurant bill split turned into an obsession. In this episode, he breaks down how he got the idea, why he worked inside a restaurant before building, how he found his first customer, and what founders get wrong about picking markets vs products. This is a sharp episode on idea formation, validation, founder conviction, and surviving when you’re too early. Timestamps:00:00 How Raj got the idea for Presto from a restaurant bill split01:48 What Presto became and why restaurants felt broken04:20 Leaving an MIT PhD to go all-in on startups07:35 Why he chose hardware instead of software09:24 How he thought about de-risking the form factor11:34 What “day two” looked like after the idea15:06 Why he actually worked in a restaurant as a waiter16:28 How to tell whether you’re early vs late to a market19:27 No competitors: huge opportunity or no market?22:10 Why founders need to uncover the real nuance themselves24:34 What Raj got wrong about the business28:00 Why he believes founders should start with product, not market30:50 When that advice breaks down33:25 How he raised early capital and found his first big customer34:35 Raj’s current process for generating and killing ideas36:25 A startup idea he killed and why41:49 Does he start with the product or the problem?44:18 Raj’s advice and current projects

10 Years of Near-Death to a $200M Uber Exit — Ryan Rzepecki (Jump Bikes)

Ryan Rzepecki had no hardware background and no software background. But in 2008, he saw the Paris bike share system, thought the $5,000-per-bike infrastructure cost was insane, and started building a smarter alternative — a smart lock on a bike, unlocked with your phone. What followed was a 10-year journey of product pivots, near-bankruptcy, and relentless grinding. His parents mortgaged their house. His girlfriend put in her career savings. He ran the "apology tour" to customers while shipping late hardware. And just when billions of VC dollars started flooding into micromobility, none of it flowed to him. In this episode, Ryan walks through the full founding story of Jump Bikes — from the first prototype to landing Nike as a sponsor, to pulling off five existential pivots in nine months, to almost going bankrupt weeks after his Series A, to the Stripe founders stepping in at a critical moment, and finally selling to Uber for a reported $200M. Ryan is now building MeshMap, an AR startup backed by a16z crypto and Solana's co-founders. — (00:00) Cold open: "I told my wife we might need to disappear to Argentina" (00:38) Intro: Who is Ryan and what was Jump? (02:28) Interview begins (03:40) The genesis moment: Paris bike share and the iPhone (06:10) First pivot: from smart lock to custom bike (07:50) A 3-sentence email to TechCrunch that changed everything (09:30) Selling bikes to cities on a subscription model (10:45) First customers: Buffalo, Sun Valley, SF Airport (12:50) Shipping hardware without a hardware background (15:15) The apology tour (17:23) Jump at the time of acquisition (19:30) Lime and Bird disrupt the market (21:00) The 5 existential moves in 9 months (26:13) Deploying in SF and almost dying (29:33) "A 10-year overnight success" (30:03) Why he noticed Paris bike share (34:03) Why this couldn't be lean-tested (40:59) Parents mortgage the house, girlfriend invests savings (43:26) "We always had just enough success" (46:42) Early competition (49:35) "Right to win" vs. domain expertise (52:26) First hires and never missing payroll (55:02) When Lime and Bird showed up (59:00) What Ryan is building now: MeshMap (1:07:43) Advice for founders

100 Customer Interviews to a $1B Exit to Stripe | Metronome's Story

Scott Woody left Dropbox with no idea, no code, and no product. What followed was months of deliberate exploration — a LIDAR startup, a nanny company, workflow software — before a random conversation with HashiCorp unlocked the wedge that became Metronome, a billing infrastructure company that just sold to Stripe for $1 billion. In this episode, Scott breaks down the exact methodology he used to find the idea: 100 customer interviews in 6 weeks, why universal pain isn't enough (it has to converge), and how an offhand comment about "the metering problem" changed everything. He also shares how he would ideate differently in 2026 — where AI has collapsed the cost of building software and made competition nearly instantaneous. If you're a founder in the early exploration phase, or thinking about what to build next, this is one of the most tactical conversations we've had on the show. What we cover: - Why Scott spent 6 months not writing a single line of code - The "would you pay us before we build it" filter that killed most ideas instantly - How 100 back-to-back customer interviews create a living, iterating pitch - Why diffuse pain is a trap — and what convergent pain looks like - The accidental HashiCorp conversation that surfaced metered billing - Why every major VC passed on Metronome's seed round - How to find your structural asymmetry in a world of instant competition - The "second-order wave" framework for finding whitespace in AI Timestamps: 00:00 Introduction — Scott Woody and the Metronome story 02:21 What is Metronome? 05:30 Scott's background at Dropbox 09:48 Why simple pricing changes took 3–6 months 12:20 Leaving Dropbox: LIDAR, nanny companies, going wide 14:33 The ideation methodology — 6 months before writing code 16:00 Why B2B enterprise? Building a moat from day one 18:13 How to generate ideas from your own pain 20:24 100 customer interviews in 6 weeks 26:28 The diffuse pain trap 29:00 The accidental HashiCorp conversation 31:50 How the interview shifts from learning to pitching 34:54 Was finding metered billing a real accident? 38:46 Did Metronome create the usage-based billing category? 40:29 Getting design partners to pay before a product existed 42:18 Why now? The role of tailwinds 45:46 Seed round: $5M with zero revenue 48:01 When VCs changed their tune 49:26 The thesis update — from niche to dominant model 51:21 How Scott would ideate differently in 2026 53:30 Structural asymmetry: the moat you need today 55:29 Final reflections

6 Pivots to $1.4b: PostHog Origin Story (and How They'd Do It In 2026)

PostHog just raised $75M Series E at a ~$1.4B valuation—but the real story is what happened before it clicked. In this episode, PostHog co-founder James Hawkins walks through the 6 pivots that came before PostHog: why early-stage is brutally hard, how they decided what to build next, and the exact moments that told them they were finally onto something. We cover: * Why 0→1 is harder than $1M→$100M * The pivot framework they used to stop thrashing * Why selling to developers changed everything * The launch + distribution tactics that worked * How the “MVP bar” changes in an AI world James also shares what PostHog is building next—and how it could turn product analytics into actual code changes. Chapters: 00:00 Intro05:41 The 4.5-year preparation before quitting13:16 Why traditional validation is dangerous territory17:45 Targeting developers: the bluntest stakeholders26:58 The NFL child analogy: when to pivot vs. persist28:14 Why the engineer retention tool failed32:37 Improving the process: why your website is the sales team36:03 The idea maze of territory management tools44:25 Building in 2026: the magic bar for new products47:54 How to use competition as a validation shortcut56:14 The PostHog aha moment: inverting the meeting01:02:18 The Hacker News pre-mortem launch strategy01:10:34 The pivot from open source to 90% cloud revenue01:16:53 Future: code editors that build products for you

From $15M to $1.8M and Back: How Dover Found PMF Twice (George Carollo)

Dover is one of those YC companies that looks “obviously successful” from the outside. But the real story is way messier: they hit PMF with a product called Autopilot, scaled fast, and then the 2022 hiring crash basically deleted their market overnight. In this episode of Before It Clicked, Sunny Rekhi sits down with George Carollo (cofounder of Dover) to unpack what it actually feels like to lose product-market fit, run experiments inside a ~30-person company, and claw your way back to growth. Today, Dover is a lean, profitable business doing ~ $6M in revenue, supporting ~50 recruiters on its marketplace with a ~10-person core team. What you’ll learn: - How Dover grew Autopilot from ~$1M → $3M → $6M → $15M… and what “PMF” looked like in practice - The moment the market flipped in 2022 (and why churn felt like a “light switch”) - What happens when you’ve raised $23M, hit ~85 people, and suddenly need to reinvent the company - Why Dover launched a free ATS (now used by ~1,000 companies) and a bunch of “micro-products”… and why none of it fixed the core issue - The email that sparked Dover’s second act: the Recruiter Marketplace, and the “why now” behind fractional recruiting in an AI-driven world - George’s biggest meta-lesson: we only hear the “iterate forever” stories when they work Chapters 00:00 Cold open — “This might be real… then layoffs”00:39 Intro: Dover found PMF twice03:19 Dover today: ~$6M revenue, (near) profitable, lean team05:13 2019: becoming recruiters to learn the problem07:31 “Saturated market” + the leap of faith08:48 No strong “why now” (and why that’s okay)13:49 Opportunity cost & timeboxing16:47 YC twist: Swap Space → recruiting17:42 Building Autopilot from a service19:20 PMF as a spectrum (referrals as the signal)21:34 The service→product trap: who owns outcomes?24:02 Autopilot growth: $1M → $15M24:34 2022 layoffs: losing PMF overnight29:02 The first layoff + “line in the sand”29:52 Trying to save Autopilot (self-serve, pricing experiments)30:56 Free ATS launch (and why)36:59 The “spray & pray” era (6 micro-products)38:20 The email that sparked the Marketplace42:21 Marketplace “why now”: AI leverage + spiky hiring45:10 Why agencies break (incentives + cost)47:07 Founder-led bet: team-of-one experiment (Apr→Sept)53:38 Killing Autopilot once Marketplace worked57:27 Selling the pivot internally (metrics + deck)01:03:20 Survivorship bias: “we only hear the ones that work”01:07:00 What’s next for Dover01:08:39 Hiring / recruiting partners