How to validate a startup idea with a real landing page (and what it costs)
A landing page with real traffic behind it is the cheapest honest test of demand. Here is the whole method, step by step, with the budget it actually takes.
Oct 10, 2026 · 7 min · KPIndicator
Asking people whether they would use your product tells you whether they are polite. Putting a real offer in front of strangers and counting who tries to buy tells you whether there is demand.
A landing page test — sometimes called a smoke test or fake-door test — is the cheapest way to get that second kind of answer. You describe the product as if it exists, send real traffic to the page, and measure what visitors do. No product is built until the numbers say it should be.
Here is how to run one properly, and what it costs.
1. Write down the decision before anything else
Decide what result means build it, and what means don't, before a single visitor arrives. For example:
| Measure | Go if at least |
|---|---|
| Visitor → email signup | 4% |
| Cost per signup | under $40 |
| Paid deposits or pre-orders | 3 |
The exact numbers depend on your market and your economics — we cover how to pick them in how to set a go/no-go threshold. What matters is that they exist in writing first. A threshold set after you see the data is not a threshold; it is a justification.
2. Make one specific offer
The page should read like the real product launch, not a survey:
- One customer. "Bookkeeping for solo landscapers" beats "bookkeeping for small businesses."
- One promise. The outcome they get, in their words.
- A price. Showing a price is what separates curiosity from intent. Without one, a signup tells you very little.
- One action. Join the waitlist, book a call, or — the strongest signal — leave a small refundable deposit.
If you have two or three angles you can't choose between, run them as separate pages with the same traffic source. The ranking between them is often more trustworthy than any single number.
3. Build the page in a day, not a month
You do not need a designer or a developer. A landing page builder (Carrd, Framer, Webflow, Unbounce), a domain, analytics, and a form or checkout are enough. Tracking is the part people skip and regret: you need to know, per traffic source, how many people visited, how many acted, and what each action cost.
4. Buy the traffic — don't borrow it
Posting the page to your own network produces friendly numbers. Friends click, colleagues sign up to be nice, and the result looks better than the market will be.
Paid traffic from strangers who match your target customer is what makes the number honest. The usual channels:
- Paid social (Meta, LinkedIn) — good for reaching a specific job title or interest.
- Search ads (Google, Bing) — good when people already search for the problem.
- Cold outreach to a targeted list — good for B2B, slower to set up.
Click prices vary widely by industry and audience. As a rough guide, consumer audiences often cost well under $2 a click on social platforms, while B2B and professional audiences can run from a few dollars to well over $10. Run a small test spend first and use what it costs you, not a benchmark, to size the rest.
5. Get enough visitors to trust the number
A conversion rate from 40 visitors is noise. As a working rule, aim for a few hundred visitors per page before you read the result — more if your expected conversion rate is low, or if you are comparing several pages and the differences between them are small.
Run the test for at least a week so you are not reading one unusual day, and do not change the page halfway through. Every change resets the clock.
6. Read the result against the threshold — and accept a no
When the test ends, compare each number against what you wrote down in step 1. There are four honest outcomes:
- Go — it cleared the bar. Build a narrow first version for the people who signed up.
- No-go — it didn't. That is a complete, useful answer that just saved you the build.
- Pivot — one angle or audience worked and the others didn't. Re-test the one that did.
- More data needed — a genuine grey zone. Run a short follow-up that isolates the biggest doubt; don't decide on gut feel.
You can see what this looks like written up in our sample validation report.
What it costs
Doing it yourself, for one idea:
| Item | Typical cost |
|---|---|
| Domain | $10–20 |
| Landing page builder | $0–50 a month |
| Analytics | Free |
| Ad spend | $500–2,000, depending on click prices and how many visitors you need |
| Your time | 2–4 days of setup, then checking in daily for 1–3 weeks |
So a solid self-run test is usually somewhere between $500 and $2,000 in ad spend, plus a few days of your time. Compare that with building a first version, which commonly costs months and tens of thousands — and only then tells you the same thing.
The ad spend is not where most self-run tests go wrong. They go wrong on the parts that need discipline: no threshold set in advance, traffic from friends, too few visitors, and a page that changes every other day.
Or have it done for you
That discipline is what we sell. KPIndicator builds the page, buys and manages the traffic, agrees the threshold with you in writing before anything runs, and hands you a written go / no-go with the numbers behind it. You keep the domain, the analytics, and the leads.
We are currently taking a founding cohort of five clients: a full Market Test at $1,250 instead of $2,500, ad spend at cost, in exchange for permission to publish the results.
Apply for a founding place — or see what the finished report looks like first.