Product · Featured Debate

5 guests 8 episodes 3,166 words

The MVP Trap: When "Good Enough" Isn't — and When Perfection Is the Enemy

Should you ship a scrappy MVP or hold out for a high quality bar?

Reid Hoffman famously said you should be embarrassed by your first release. Rahul Vohra spent years polishing Superhuman before letting anyone touch it. Cameron Adams held off launching Canva despite investors begging him to ship. Gaurav Misra ships a new feature every week at Captions. They are all successful. So who is right — and more importantly, when is each approach the winning bet?

The MVP debate is one of the most persistent arguments in product development, but after listening to how ten different leaders across Lenny's podcast and newsletter actually navigate this question, a more nuanced picture emerges. The real variable is not speed or polish. It is the cost of a bad first impression in your specific market — and whether you are innovating or implementing.

You have a product idea with real potential. Do you get it into users' hands as fast as possible — accepting rough edges, incomplete features, and potential embarrassment — or do you invest in a quality bar that creates an exceptional first impression, even if it means months of additional development?

The 5 Positions

1 / 5

Ship the MVP Fast — Contain the Liability of Being Wrong

Eric Ries · Gaurav Misra

Eric Ries — Creator of the Lean Startup methodology, founder of the Long-Term Stock Exchange. Ries coined the term MVP and has spent over fifteen years watching it get misunderstood, misapplied, and blamed for failures it did not cause.

Ries argues that MVP has been fundamentally mischaracterized. It is not a prescription to ship garbage. It is a descriptive concept — naming what already happens when startups test ideas. The core logic is risk containment: the most expensive mistake in product development is not a rough first impression, it is spending years building something nobody wants.

"People think that MVP is about a specific tactic. So, an MVP is like a bare bones, stripped down the thing that will crash your computer, but it's not anything to do with MVP. MVP is simply for whatever the hypothesis is that we're trying to test, what is the most efficient way to get the validation we need about whether a hypothesis is true or not?"
Eric Ries
▶ 00:28:22

When Linear founder Karri Saarinen asked whether MVPs still work given rising user expectations, Ries flipped the framing: if the quality bar is high, that means you can only afford to build one feature at that quality level, not fifteen. Either way, you are making a trade-off — and MVP forces you to confront which trade-off you are actually making.

"A lot of MVP is just about containing the liability of making a mistake. So yes, one of the things I really encourage people to do is go find 10 customers, not 10 million, but find 10 customers. Get them to use the product and have them tell you what's awful."
Eric Ries
▶ 00:29:22

He points out that in industries far more demanding than the iOS App Store — battery backup systems for data centers, deep-sea drilling equipment, jet engines — the cost of offering someone something they do not like is still just that they say no. He also draws a critical distinction between a product launch and a marketing launch: you can ship a product to ten customers without telling the world. Nobody will find out.

"I don't really care how high quality the initial MVP is because quality is defined by the customer. And if we don't know who our customer is, we literally don't know what the word quality means."
Eric Ries
▶ 00:32:06

Gaurav Misra — CEO and co-founder of Captions, early employee at Snap where he led the design engineering team; Captions has over ten million users and has raised over $100 million. Misra takes the speed argument further by reframing technical debt as strategic leverage.

"Our engineering goal is every engineer should ship a marketable product every week."
Gaurav Misra
▶ 00:00:21

In fast-moving markets like AI, speed is the primary competitive moat. Misra argues that technical debt is not a liability but a form of leverage — analogous to financial debt that lets you buy a house you otherwise could not afford. The constraint is interest payments: every piece of debt costs roughly one to two percent of daily engineering time in maintenance and bugs. Accumulate too much, and you spend eighty to ninety percent of your time just paying interest. Captions also maintains two parallel roadmaps: a public one (table-stakes features every competitor knows about) and a secret one (conviction bets where speed of execution is the differentiator).

When This Applies

when you are testing a genuinely new hypothesis, your market is forgiving of early-stage products, or the cost of being wrong for another six months exceeds the cost of a bad first impression. Captions (AI video), Vanta (SOC 2 as spreadsheet), and most pre-PMF startups fall here.

Invest in the Quality Bar — Create Super Fans From Day One

Rahul Vohra · Cameron Adams

Rahul Vohra — Founder and CEO of Superhuman, previously sold Rapportive to LinkedIn. Achieved a 58% "very disappointed" product-market fit score, well above the 40% threshold. Known for rigorous first-principles thinking about product-market fit measurement.

Superhuman's approach was the deliberate anti-MVP. They manually onboarded every single user through one-to-one concierge sessions. At peak, about twenty people were doing manual onboarding. You could not use the product without going through the experience.

"In those early days we insisted on one-to-one concierge onboarding, and it was absolutely the right thing to do. You couldn't use Superhuman unless you went through the onboarding experience."
Rahul Vohra
▶ 00:48:12

The logic has two parts. First, the user metrics were excellent across the board — engagement, retention, PMF score, NPS, and virality all significantly beat industry benchmarks.

"I think you'll significantly beat your industry benchmarks if you go to the effort of one-on-one onboarding your early customers. It becomes so powerful to have that early cohort of super fans when it comes to things like building a brand."
Rahul Vohra
▶ 00:49:49

Second, there was a practical efficiency argument that turns the conventional wisdom upside down. Manual onboarding was actually the lean choice — it just did not look like it.

"I saw other companies often competing spend almost half their engineering dollars on self-service flows for products that ultimately did not find product market fit."
Rahul Vohra
▶ 00:50:32

By using people instead of code for activation, every engineer could focus on finding product-market fit, not building onboarding flows for a product that might not work. The crucial caveat: Vohra always made money on onboarding. This was not a charity program. And he is clear about when to stop — when personality types who resist one-on-one interaction become important customers.

Cameron Adams — Co-founder and CPO of Canva, which grew to $2.3 billion ARR and profitability, making it bigger than Figma, Miro, and Webflow combined. Organic word-of-mouth was the primary growth engine.

"The Lean Startup book came out, so that was all anyone talked to us about... They were like, 'Just get something out the door, as crappy as it is just to get in front of users.' I think, for us, the product is the experience and giving people a great experience is an intrinsic part of the product."
Cameron Adams
▶ 00:30:59

Adams resisted the Lean Startup consensus and held off launching. His reasoning: Canva's growth depended on people being excited enough to tell others about it. A crappy product that got the job done but did not create joy would not have generated the word-of-mouth that drove growth for years. The critical enabling condition: the Canva founders had deep domain expertise from their previous company, Fusion Books, so they were not building in the dark.

When This Applies

when word-of-mouth from early users is your primary growth engine, your product is in a premium or trust-intensive category, or you have funding to substitute dollars for engineering effort on activation. Superhuman (premium email), Canva (design tools), and consumer products where joy drives sharing fall here.

Build Beyond MVP — Design for the Hardest Case From Day One

Jeremy Henrickson

Jeremy Henrickson — Former VP Product at Rippling, previously at Coinbase during its scaling phase. Led Rippling's expansion from a handful of products to dozens of product lines.

"I think in our case specifically for Rippling, a minimum viable product would do a disservice to both our customers and to the very team that was building it. And the reason I believe that is that when you design a minimum viable product, you're optimizing for speed. And in that set of optimizations, you are minimizing the deeper product thinking about what can fully differentiate our product."
Jeremy Henrickson
▶ 00:18:16

The deeper problem Henrickson identifies is architectural: MVP shortcuts compound. If you only think through simple cases, you make a set of architectural assumptions. Then you build on those assumptions for six to nine months, with dozens or hundreds of decisions stacked on top.

"It's extremely difficult to unwind those decisions once you've built them into the product."
Jeremy Henrickson
▶ 00:19:38

Rippling's global payroll exemplifies this: instead of copying their US system for one country, they built for six very different countries from the start. Adding subsequent countries became far easier. The approach does not mean shipping the hardest case — it means building the foundation so you can get there later without unwinding everything.

When This Applies

when you are building known product categories on a shared platform, when architectural decisions compound, or when unwinding early shortcuts would cost more than building right the first time. Rippling (compound startup), enterprise infrastructure, and platform companies fall here.

Replace MVP with "Minimum Lovable Product

Jiaona Zhang

Jiaona Zhang — SVP of Product at Webflow; formerly at Airbnb, Dropbox, and WeWork; lectures on product management at Stanford. Her breadth across consumer marketplaces, productivity tools, and design platforms gives her an unusually wide view of when different quality bars are appropriate.

"The reason I care so much about minimal lovable product is because I do think in a world where there are so many different options, it's hard to just feel like, 'Hey, use this thing. It barely meets a quality bar.'"
Jiaona Zhang
▶ 00:17:10

Zhang's key nuance is that the required lovability depends on what you are competing against. If you are replacing a manual spreadsheet workflow, your product without fifteen additional features is already lovable. If you are building for designers, the bar is much higher.

"There's a world where your quality bar, your 'quality bar,' or let's call it your polished bar, can be a little bit lower because the reality is the thing that you're competing against or you're replacing is literally a manual workflow."
Jiaona Zhang
▶ 00:17:40

The practical advice: do five things with a high degree of polish rather than fifteen things where every part feels clunky. Lovability comes from two sources — things working well (baseline) and what Zhang calls "pixie dust" (small touches that exceed expectations).

"People I think would actually respect this idea of 'You've given me minimal lovable in five areas as opposed to minimal viable in 15 areas.'"
Jiaona Zhang
▶ 00:20:45

When This Applies

when you are implementing a known mechanic (notifications, onboarding flows, leaderboards) where users already have strong expectations from existing products. Duolingo (gamification mechanics), and any product adding established engagement patterns fall here.

Context-Dependent Quality — Match the Bar to the Domain

Paul Adams · Shweta Shrivastava · Jorge Mazal · Katie Dill

Several voices converge on the idea that the right answer depends entirely on context — and that the tension between speed and quality is permanent, not resolvable.

Paul Adams — CPO of Intercom for over ten years; previously global head of brand design at Facebook, user researcher at Google. Intercom's shipping principle literally evolved on the wall.

"We have a principle called Ship to Learn. And, we've actually changed it since. It's over on the wall here. Ship fast, ship early, ship often is what it says now."
Paul Adams
▶ 00:14:09

But Adams is unusually honest about what this costs:

"There's a real tension there, a real trade-off, where people have these high standards, which we encourage. We encourage them to ship fast, and learn, and make mistakes. It's a constant tension that we're navigating."
Paul Adams
▶ 00:14:48

Shweta Shrivastava — Senior Director of PM at Waymo; previously CPO at Nauto, Head of PM at AWS. Leads product at a company where the MVP bar is literally a matter of life and death.

"The MVP bar itself for safety is extremely high for us. So, the core product management philosophy also, getting an MVP out there and then iterating with the real world deployment. It applies, but it's just a different bar on that MVP."
Shweta Shrivastava
▶ 00:10:42

Waymo reframes the debate: MVP is about the minimum viable version. In safety-critical domains, "viable" means "will not kill anyone." The bar is non-negotiable. But the philosophy of shipping the minimum and iterating still applies.

Katie Dill — Head of Design at Stripe; previously at Airbnb and Lyft. Offers a structured framework for thinking about quality levels.

"There are levels of quality. There is the, does the thing work? Does it provide some sort of value proposition? It like executes on its job. That's baseline quality."
Katie Dill
▶ 00:11:39

Beyond baseline, Dill identifies progressively higher levels: error-free and well-rounded, then exceeding expectations and doing something the user was not even seeking. The right level should be based on user expectations — and the quality bar keeps evolving, keeps rising. This question never goes away.

Jorge Mazal — CPO of Duolingo, led the product team during exponential growth. Writing in Lenny's Newsletter, Mazal identifies a hidden variable most teams miss: whether you are innovating or implementing.

"Your first minimum viable product (MVP) version of an existing, widespread mechanic is not the time to get clever and innovative."
Jorge MazalLenny's Newsletter

If you are implementing a known mechanic (leaderboards, notifications, streaks), your MVP should be best-in-class industry standards. If you are inventing something new, ship rough and learn. Duolingo's leaderboard took months even as an MVP because the existing pattern had to be executed at a high level before they could innovate on it.

When This Applies

when the quality bar differs across features and user segments, and different parts of your product warrant different quality levels. Stripe (15 essential journeys), Intercom (speed vs. craft), and mature product organizations managing the tension across teams fall here.

Evidence from the Archive

Lean Startup / LTSE

Company that built a beautiful product but had a JavaScript error on the signup button — 0% signups, and they could...

Company that built a beautiful product but had a JavaScript error on the signup button — 0% signups, and they could not tell if the value prop was wrong or the button was broken

Eric Ries · Creator of the Lean Startup methodology, founder of the Long-Term Stock Exchange ▶ 00:28:22

Superhuman

Superhuman achieved a 58% 'very disappointed' score on the Sean Ellis PMF survey with manually onboarded early users

At peak, 20 people were doing manual onboarding — a surprisingly small team for the impact it created

Rahul Vohra · Founder and CEO of Superhuman ▶ 00:48:12

Captions

Technical debt runway as a mental model: too much debt means 80-90% of time goes to paying interest

Every engineer at Captions ships a marketable product every week

Gaurav Misra · CEO and co-founder of Captions ▶ 00:00:21

Webflow

Airbnb's mobile app revamp: pre-populated templates as 'pixie dust' that created lovability

Webflow released membership and logic features at minimal viable rather than minimal lovable, relying on the ecosystem to add the lovable layer

Jiaona Zhang · SVP of Product at Webflow; formerly senior director at WeWork and longtime PM at Airbnb and... ▶ 00:17:10

Canva

Canva reached $2.3B ARR and profitability — bigger than Figma, Miro, and Webflow combined

Canva held off launching despite repeated investor pressure, because they knew the problem space from Fusion Books

Cameron Adams · Co-founder and CPO of Canva ▶ 00:30:59

Waymo

1.35 million deaths per year from traffic accidents, mostly from driver error, sets the mission bar

Waymo chose fully autonomous driving from day one — not driver assist — making the MVP bar necessarily higher

Shweta Shrivastava · Senior Director of Product Management at Waymo; previously CPO at Nauto, Head of PM at AWS ▶ 00:10:04

The Synthesis

The MVP debate is actually a misframed question about three distinct variables:

First Impression Cost

When does a bad first impression actually matter?

Innovating vs. Implementing

How do you know which quality bar to apply?

Architectural Shortcut Debt

Why are MVP shortcuts more expensive than they seem?

Five Things Brilliantly

What is the practical synthesis of MVP vs. quality?

For most early-stage startups testing a new hypothesis, the cost of a bad first impression is near zero. But for products where word-of-mouth is the primary growth engine, a bad first impression means losing the super fans who would be your unpaid marketing team. The cost compounds.

If you are building a novel solution to an unmet need, ship fast and learn -- you do not even know what quality means yet. If you are implementing an established mechanic in a mature category, your minimum must be best-in-class. Many teams conflate these two situations and apply the wrong playbook.

MVP shortcuts compound technically. The first version's architectural assumptions become the foundation for six months of subsequent decisions. If you are building a platform or compound product, early shortcuts are extraordinarily expensive to unwind later.

Instead of MVP vs. quality, ask whether you are at minimal viable or minimal lovable -- and for which specific features. Do five things brilliantly instead of fifteen things adequately. That is the real discipline.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

What is the cost of a bad first impression in your market?

Question 2

How well-understood is the product category you are entering?

Question 3

What matters more right now: learning speed or user delight?

Notable Absences

The Bottom Line

Lenny's Newsletter reinforces the MVP-done-right approach in "How to validate your startup idea": Vanta's founders built a zero-code MVP (a spreadsheet) to validate SOC 2 demand before writing any code. Christina Cacioppo's advice: "Do all the product development best practices... Nobody wants to do them because they seem tedious, they're less fulfilling than building." And Miro's approach, documented in "How Miro builds product," shows how mature organizations operationalize both speed and quality through systems like the "Mona Lisa principle" and design reviews.

Jiaona Zhang offers a practical synthesis: instead of MVP vs. quality, ask whether you are at minimal viable or minimal lovable — and for which specific features. Do five things brilliantly instead of fifteen things adequately. That is the real discipline.

Sources

Eric Ries — "Reflections on a movement | Eric Ries (creator of the Lean Startup methodology)" — Lenny's Podcast, October 29, 2023
Rahul Vohra — "Superhuman's secret to success: Ignoring most customer feedback, manually onboarding every new user, obsessing over every detail, and positioning around a single attribute: speed | Rahul Vohra (CEO)" — Lenny's Podcast, March 23, 2025
Gaurav Misra — "How to win in the AI era: Ship a feature every week, embrace technical debt, ruthlessly cut scope, and create magic your competitors can't copy | Gaurav Misra (CEO and co-founder of Captions)" — Lenny's Podcast, March 27, 2025