Landing Page MVP

For a long time the accepted answer was that you can't know whether an idea is any good until you build it. So teams spent six months building, shipped, and only then discovered that nobody wanted the thing. Six months and five salaries bought a single line of information: people don't want this.

Planning meetings stalled at the same spot every time. Someone would ask whether customers really wanted it, and the answer was that they'd know once people could try it. Trying it required building it, building it took six months, so the question got postponed instead of answered. The team didn't build because it was confident. It built because there was no other way to find out.

This method reverses the order. Instead of shipping a product, you ship a description of the product and watch whether anyone raises a hand. Having nothing to give the people who do raise their hands isn't a flaw in the setup — it is the setup. Nothing was built, so being wrong costs a week and a few hundred dollars in ads.

Build one page, send traffic to it, and compare the share of visitors who press the button against a threshold you wrote down in advance. There are five steps.

1. Write the hypothesis you want to test as one sentence. A hypothesis is a guess phrased so it can come out true or false. "If meeting notes get written up automatically, team leads at companies of ten or more will pay $8 a month" names a group and a condition, which is what makes a verdict possible later.
2. Build one landing page. A landing page is a standalone page built to produce a single action. One line on what the product does, two screenshots, a price, and one button carrying a call to action (CTA) is enough. Leave out the nav bar and the company story. If there are several things to click, you can't tell what the click meant.
3. Send people to it. Search ads, community posts, an email blast — whatever gets you at least a few hundred visits. At 50 visits you can't separate one sign-up from three, and a single enthusiastic friend swings the whole result.
4. Tell everyone who presses the button that the product isn't ready yet, and collect a contact. Don't fake a checkout that errors out. Say you'll notify them at launch and take an email address.
5. Compare the conversion rate against your threshold. Conversion rate is the share of visitors who take the target action, so 40 sign-ups out of 1,000 visits is 4 percent.

Write the threshold down before the page goes live. Decide afterward and 2 percent becomes "not bad for a first try" while 0.5 percent acquires an explanation of its own. If you can't find a credible industry figure, use the number from another sign-up page your team already runs, since at least that one was measured on the same kind of traffic. It's tempting to A/B test two headlines, but at a few hundred visits the split won't separate them.

It fits when nothing has been built yet, the product explains itself in one sentence, and you can buy or borrow an audience. All three conditions matter. If half the product already exists, shipping it beats describing it, and if the pitch needs five sentences, a low conversion rate won't tell you whether people didn't want it or simply didn't follow it.

The limits follow from what actually got measured. What the page measures is interest in a description. It says nothing about whether the people who signed up will pay and stay. Of 400 people on a waiting list, perhaps 20 reach a paid plan, so this is not evidence of product-market fit — a different thing is being measured.

Where the traffic came from moves the number too. The same page posted to a community that knows you converts at 12 percent and converts at 2 percent from search ads. The first figure measures your relationships, not your product. To see how people react while actually using the thing, move on to a concierge MVP, where a person delivers the service by hand, or a Wizard of Oz MVP, which looks automated from the outside.

"Let's put up a landing page first and see how many sign-ups come in."

A proposal to measure demand before building.

"We got 1.2 percent. The threshold was 3, so I think we drop this direction."

The number missed a line set in advance, so the idea gets shelved.

"Four hundred people on the waiting list doesn't mean four hundred users."

Drawing the line between signing up and actually using something.

At Pagedeck the team ran this on an idea for automatic meeting notes. Mia wrote the hypothesis and the threshold, Ravi built the page in two days and bought search ads. Over one week: 1,200 visits, 52 sign-ups, 4.3 percent against a threshold of 3. The verdict was pass.

Nobody started building that afternoon, though. All 4.3 percent establishes is that the pitch sells; whether automatic meeting notes would be any good was still unknown. The following week Mia contacted ten of the people who had signed up, took recordings of their meetings, and wrote the summaries by hand. What the page opened was the next experiment, not a decision.

The most common mistake is reading a conversion rate above the threshold as proof the product will work. The number cleared the bar, so the conclusion feels earned. But what cleared the bar was the pitch, not the product. It means people liked the sentence "your meeting notes write themselves," and if the notes that come out are mediocre those same people leave in week one. Left uncorrected, this reading sends a reassured team straight back into a six-month build.

The other reaction is that advertising a product you don't have is dishonest. A page that looks like a storefront invites that reading. Two things draw the line: whether money changed hands, and whether the state of the product was hidden. Disclose right after the click and take no payment, and you're collecting early interest rather than deceiving anyone. Cross that line into real charges and you get refunds and apology emails, and the trust you lose costs more than the number you gained.