Of the 431 venture-backed companies CB Insights traced to a shutdown since 2023, 70% ran out of capital, but CB Insights calls that "almost always the final cause of death, not the root problem". The root problem it finds most often is poor product-market fit, cited by 43%, and two-thirds of those companies were early stage. That is the MVP stage.
This page gathers the MVP numbers that can be checked, labels agency estimates as such, and says plainly where a number does not exist, such as the share of MVPs built in no-code.
How long an MVP takes, and what it costs
Only the agencies selling MVPs publish timelines and prices, so every figure in this section is a vendor estimate. Three independent 2026 guides still land in the same place.
| Scope | AddWeb Solution | CMARIX | Purrweb |
|---|---|---|---|
| Simple (one core flow, standard sign-in, no integrations) | 4 to 8 weeks, $8,000 to $20,000 | 6 to 10 weeks, $15,000 to $35,000 | 6 to 10 weeks plus 1 to 2 weeks of discovery, $25,000 to $35,000 |
| Medium (several roles, payments, a few integrations) | 8 to 12 weeks, $20,000 to $45,000 | 10 to 16 weeks, $35,000 to $80,000 | $35,000 to $45,000 |
| Complex, marketplace, fintech or AI | 12 to 18 weeks, $45,000 to $100,000 or more | 16 to 24 weeks, $80,000 to $150,000 or more | Add 2 to 4 weeks, $45,000 to $60,000 or more |
Source: AddWeb Solution, CMARIX and Purrweb cost guides, updated July to September 2026; agency estimates, not measured project data. AddWeb's ranges assume an offshore or hybrid team.
Purrweb's ten-week plan shows where the time goes: weeks 1 and 2 discovery, 3 and 4 design, 5 to 8 development, 9 and 10 QA and launch. Development is four weeks of ten.
The spread in prices is mostly team location. AddWeb calls $25,000 to $45,000 the median market-ready MVP and says a US in-house or US-agency build of the same scope runs two to three times its figures. CMARIX publishes the rate card behind that gap: $120 to $250 an hour in the US and Canada, $100 to $180 in Western Europe, $50 to $100 in Eastern Europe, $40 to $80 in Latin America and $20 to $70 in India.
After launch. Maintenance runs 15% to 25% of the build cost a year (CMARIX) or about 20% (AddWeb). AddWeb adds 20% to 40% for hosting, integrations and post-launch iteration.
What Y Combinator and Stripe say about speed
In his Startup School talk "How to Plan an MVP", Michael Seibel told founders a lean MVP should be built "in weeks, not months" and that the spec should be timeboxed: pick a launch date, then include only what fits.
Among the 23,000 startups incorporated through Stripe Atlas in 2025, 20% took a first paying customer within 30 days, up from 8% in 2020, and among those monetising in their first three months the median time to first payment was 34 days, down from 38 in 2024. The caveat is selection: companies that incorporate through a payments company skew towards products that charge from day one.
How often startups fail, and when
| Horizon | Cohort (year ended March) | Survival rate |
|---|---|---|
| 1 year | 2024 | 77.9% |
| 2 years | 2023 | 65.9% |
| 5 years | 2020 | 51.4% |
| 10 years | 2015 | 34.7% |
Source: US Bureau of Labor Statistics, Business Employment Dynamics, table 7
Those are all establishments, not startups, and an exit includes a sale or a retirement. The "90% of startups fail" line is different data: Failory traces it to Startup Genome, whose 2019 report put it at 11 in 12.
The recent trend is upward. Carta counted 966 US startup shutdowns among its customers in 2024 against 769 in 2023, a 25.6% rise, and called even that an undercount (as reported by TechCrunch, January 2025). SimpleClosure's data in the same report puts 74% of closures since 2023 at pre-seed or seed, with seed alone at 41%: the MVP stage is where most companies end.
Why they fail: three datasets, three answers
| Reason | Share of the 385 companies with an identifiable cause |
|---|---|
| Ran out of capital | 70% |
| Poor product-market fit | 43% |
| Bad timing or macro conditions | 29% |
| Unsustainable unit economics | 19% |
Source: CB Insights, 431 venture-backed companies shut down since 2023, published March 2026; companies could cite several reasons
The dataset you read decides the answer. CB Insights' sample is venture-backed, so capital dominates. Failory's own interviews with more than 80 founders of failed, mostly bootstrapped startups put marketing first at 56%, then team 18%, finance 16% and tech 6%. Wilbur Labs' February 2026 survey of 200 US tech founders ranks competition and shifting market dynamics at 45%, technology or product issues 44%, hiring missteps 30% and running out of money 25%, down from 38% in its 2023 survey.
Read together: the venture-backed die of money, the bootstrapped die of distribution, and almost nobody dies of code. For an MVP budget, that argues for spending less on the build and more on the weeks of customer conversations around it.
How often startups pivot
Pivoting is the norm. In Wilbur Labs' 2026 survey, 81% of founders had pivoted from their original idea at least once and 57% had made a major pivot or several. The sample is 200 founders with a stated margin of error of plus or minus 6.9 points, so even the low end means roughly three in four MVPs do not ship the idea they started with.
The Startup Genome report, summarised by Steve Blank in May 2011, is the only study linking pivot count to outcomes: across more than 650 early-stage internet startups, those that pivoted once or twice raised 2.5 times more money and grew users 3.6 times faster than those that pivoted never or more than twice, and solo founders were 2.3 times less likely to pivot at all. Blank called the methodology "still shaky" and the sample is fifteen years old, so trust the direction more than the multiples.
What activation looks like after launch
| Measure | Figure | Source and sample |
|---|---|---|
| Activation rate, all products | Average 34%, median 25% | Rachitsky and Timen, 500+ responses, 2022 |
| Activation rate, SaaS only | Average 36%, median 30% | Rachitsky and Timen, 2022 |
| Activation rate, B2B SaaS | Average 37.5%, median 37.04% | Userpilot, 62 companies, 2025 |
| Month-one retention | Average 46.9%, median 45.25% | Userpilot, 83 companies, 2025 |
Sources: Lenny's Newsletter activation survey, October 2022; Userpilot SaaS Product Metrics Benchmark 2025
Two caveats. Userpilot's figure is vendor-sourced from its own customers, whose average revenue is $63.9 million; an MVP with 40 users should expect to sit below it. And activation depends on the milestone. Rachitsky and Timen define it as the earliest onboarding action that predicts long-term retention, and call the 60th percentile good and the 80th great. Pick the milestone before launch.
No-code versus code: what can and cannot be measured
There is no primary survey of how many MVPs are built with no-code tools; vendor pages claim 70% of new businesses use low-code or no-code without naming a source. The nearest proxy is Clutch's August 2025 small-business survey: of the 83% of US small businesses with a website, 41% run on DIY builders such as Wix and Squarespace, 34% on WordPress or Shopify and 12% are custom-built; 45% were built by an outsourced partner and 37% in-house. Those are marketing sites, not applications, so adjust downwards for anything with user accounts, payments and roles.
Our own position, stated as the bias it is: we build marketing sites in Webflow and applications in Next.js with Supabase, because a founder who pivots, which the data says is most of them, needs a data model and sign-in that survive the pivot. The trade-off is the discovery and design weeks above, against a landing page that can be live in the "days, not weeks" Seibel describes. We scope before we quote, and the scoping call often ends with a recommendation to validate on a landing page first.
What this means for scoping an MVP budget
Set the timeline before the feature list: six to ten weeks for a simple product, and anything past a quarter is no longer minimum. Budget for the pivot, since four in five founders change the idea, by paying for a data model and sign-in that can be reused. Hold back 20% to 40% of the build cost for the months after launch, where the failure data says the money runs out. Define activation before the first user arrives and read your cohort against the 25% to 37% band weekly. Our web app development service page explains how we scope it.
