Start here
Should I build an MVP or validate first?
Updated 19 August 2026 · Thresholds read live from the verdict engine
Short answer
Validate first when the build is longer than a fortnight and the audience can be bought with ads. Build first when you can ship in a weekend, when you are the user, or when the value only shows up after someone lives with the product. A fake door test costs a few hundred dollars and a week. An MVP costs two to four months of your life.
What each path actually costs
The comparison people run in their head is wrong because it prices the MVP at zero when they build it themselves. Your time is the most expensive input you have, and it is the one you cannot get back.
| Path | Calendar time | Cash | What you own at the end |
|---|---|---|---|
| Fake door test | 1 week | $250 to $600 of ads | A conversion rate, a list of emails, and three tested angles |
| Concierge run | 2 to 4 weeks | Your labour | A handful of paying customers and a very clear spec |
| Solo-built MVP | 6 to 14 weeks | Hosting and your salary | Working software and an unanswered demand question |
| Agency-built MVP | 8 to 16 weeks | $15,000 to $60,000 | Working software, a maintenance bill, and the same question |
The asymmetry is the point. A test that costs $400 can save fourteen weeks, and fourteen weeks is roughly a quarter of the year you have for this idea.
When building first is the right call
- The build is shorter than the test. If you can ship something usable in a weekend, do that. The build produces both an answer and a product, and the test only produces the answer.
- You are the user and you will use it anyway. The tool pays for itself in your own time, so demand from others is a bonus rather than a condition.
- The value only appears after two weeks of use. Habit trackers, journalling apps, note tools and anything social look weak against a paywall and strong in retention. A fake door will underrate them.
- You cannot buy the audience. Surgeons, school heads and procurement teams do not click ads. Talking to twenty of them beats any funnel you could build.
When validating first saves you a season
Validate first when all three of these are true: the build is measured in months, the audience is reachable through ads or a community, and you have a price in mind. Those conditions cover most consumer app ideas.
The test is also the cheapest way to find out that your promise is wrong rather than your idea. Founders routinely discover that the third angle converts at three times the first, which is a discovery worth having before you name the product, design the onboarding and write the App Store copy around the losing angle.
Worked example
The four-month lesson
A developer spent four months building a scheduling app for personal trainers, launched it, and found that trainers already used a booking tool their gym paid for. The competing product was not another app, it was an existing arrangement nobody wanted to change.
A $300 test would have surfaced that in five days, because the ad would have been shown to trainers and the paywall would have converted at close to zero. The cost of finding out on day five is $300. The cost of finding out in month four is the four months.
The middle path most people skip
Between a landing page and a working product sits the concierge run: deliver the outcome by hand, for ten people, for money. No software at all. You take requests over email or WhatsApp, do the work manually, and charge something.
It answers the question a fake door cannot, which is whether people keep coming back. Delivery businesses, meal planners and financial tools have all started this way. The trade is speed: ten manual customers takes two to four weeks and cannot be scaled to produce a percentage, so pair it with a paid test if you want both signals.
| Question | Fake door | Concierge | MVP |
|---|---|---|---|
| Will strangers commit at this price? | Yes | Partly | Yes |
| What exactly do they want built? | No | Yes | Partly |
| Will they still use it in a month? | No | Yes | Yes |
| What does acquisition cost? | Yes | No | Yes |
| Time to an answer | 1 week | 2 to 4 weeks | 2 to 4 months |
The numbers
If you take the validation path, this is the bar the result gets held to. It comes straight from the verdict engine.
- Funnel starts before any verdict
- 30
- Below this, don't build
- 1.0%
- At or above this, build
- 2.0%
- Visitors before a segment counts
- 50
A funnel at 2% or better is permission to start building. Anything below the kill line is permission to keep your fourteen weeks and spend another $300 on a different angle.
How VerifyToLaunch does this
VerifyToLaunch is built for the validate-first path, so it does the week-one work: the funnel, the ad angles, the creatives, the tracked links and the verdict maths. It replaces the landing page you would have built by hand, and it keeps the thresholds fixed so you cannot quietly move the bar once you have seen the number.
It cannot tell you anything about retention, and it is the wrong tool if your build is a weekend or your buyers are twenty named companies. In those cases the manual paths on this page are better and cost less.
Common questions
- Is a landing page test the same as an MVP?
- No. An MVP is the smallest product that delivers the outcome, so it can measure whether people keep using it. A landing page test measures whether strangers will commit at a price. They answer different questions and cost an order of magnitude apart.
- How long should an MVP take?
- If the plan runs past twelve weeks it is not minimal, it is a product. Cut scope until the first version delivers one outcome for one audience, or run a paid test first and let the result decide what to cut.
- What if the test passes and the MVP flops?
- That happens, and it usually means the promise was better than the product. Buying intent proves the offer lands. Retention is a separate battle and you fight it after launch, with the customers the test brought you.
Read next
- How do I validate an app idea?The whole method: pick a commitment that costs the visitor something, buy cold traffic, count who commits.
- How do I test an app idea before building it?A seven-day plan, the four artefacts you need, and which test suits which kind of app.
- What does validating an idea cost?The bill line by line, and what $150, $400 and $1,000 each buy in certainty.
- How do I find out if people will actually pay?Three ways to collect a paying signal before you build, and the CAC maths behind the rate.
Buy the week back
Describe the idea and get the funnel, the angles and the verdict without building a page. If the number comes back under the kill line, you have saved a quarter of the year for the price of a few days of ads.
One-time licence, 30-day money-back guarantee