What should a soft launch prove before you do a hard launch?

What should a soft launch actually prove?

That a stranger can find you, understand what you sell, and want it enough to act, without you in the room. A soft launch is not a smaller version of a hard launch. It is an experiment with a pass mark, and if you have not written the pass mark down, you are not running one.

I have watched a lot of founders do a quiet release, get some encouraging messages, and declare it a success. Then the real launch lands flat and nobody can explain why. The soft launch never tested the thing that was actually uncertain.

So this piece is about what to put under test, how to size it, and what result should stop you from going ahead.

Why do most soft launches prove nothing?

Because they are run with a friendly audience, by a founder who does the explaining personally. Warm contacts will be kind about a confusing product, and a founder in the room compensates for everything the website fails to say.

That combination generates the most misleading signal in go-to-market work. People sign up, they say encouraging things, and none of it transfers to a cold visitor who arrives from a search result with four minutes of patience.

The second reason is that nobody defined failure. If any outcome can be read as promising, you have built a ritual rather than a test. A real soft launch has at least one result that would make you stop and rework the offer.

What are the three questions worth answering?

Can the right person find it, can they understand it unaided, and will they take the step you actually care about? Those three cover almost everything a launch can get wrong, and each one fails in a different, diagnosable way.

Findability is about channel and message match. If nobody arrives, the problem is upstream of the product. Comprehension is about the page doing the founder's job. If people arrive and leave, the message is the problem. Conversion is about whether the offer is worth the price and the risk.

Separating these matters because the fixes are completely different. A distribution problem cannot be fixed by rewriting the hero section, and a comprehension problem cannot be fixed by buying more traffic. Most teams try the second fix for the first problem, because traffic feels like progress.

How small should a soft launch be?

Small enough that a bad result costs you nothing except time, and large enough that the result is not noise. In practice that usually means one channel, one audience segment, one offer and one page.

The temptation is always to launch to everyone quietly, which is the worst of both worlds. You burn the audience you will want for the real launch, and you get a blurred signal because three different kinds of people reacted to three different things.

Narrowing is also where you discover whether your ideal customer profile is real. If you cannot describe the one segment precisely enough to reach just them, the segment is probably too broad, which is a problem worth catching now rather than after the launch. I have written about the symptoms in signs your ideal customer profile is too broad.

How long should you run it?

Long enough to see the full cycle you are testing, and not a day longer. If your buying cycle is two weeks, a five day soft launch measures interest rather than intent, and interest is the cheapest signal there is.

Set the end date before you start. Open ended soft launches drift into being the launch, and then the hard launch never happens because the moment passed while you were still gathering data.

A related discipline applies to the channel you chose. Give it a fair run and then judge it honestly rather than extending indefinitely because you are not ready for the answer. I have gone through that decision in how long to give a channel before killing it.

What counts as a pass?

A pass is a number you wrote down before you saw any data. It can be crude. Ten qualified conversations, or five people who reached the pricing page and asked a specific question, or two customers who paid without a discount.

What matters is that it is falsifiable and that it measures commitment rather than politeness. Email signups are politeness. A calendar booking is commitment. Money is commitment with a receipt.

Be careful about measuring the wrong end of the funnel. A soft launch that produces a lot of traffic and no conversations has failed, even though the traffic number looks good in a slide. The whole point is to find out whether the offer works on strangers.

What do you do when the soft launch fails?

You find out which of the three questions failed, and you fix only that. The most common mistake after a disappointing soft launch is to change everything at once, which guarantees you learn nothing from the second attempt.

If nobody arrived, change the channel or the message you are using to attract people, and leave the product page alone. If people arrived and bounced, rewrite the page and run the same traffic at it. If people understood and did not buy, the problem is price, risk or proof, and those are three separate experiments.

Failing a soft launch is the cheapest possible outcome, and it is worth saying that out loud to whoever is anxious about it. The expensive version of this lesson is a hard launch with a budget behind it.

What should you build during the soft launch?

Only what the test needs. One page, one way to convert, and enough tracking to answer the three questions. Everything else is a distraction dressed as preparation.

I say this as someone who builds websites for a living and prices fixed fee, mostly between one thousand and ten thousand dollars. It is not in my commercial interest to tell people to build less. It is in my interest to have clients who are still around in two years, and the ones who over build before validating rarely are.

What is genuinely worth building early is the feedback path. A way to talk to everyone who converts and everyone who nearly did. That conversation is the actual output of a soft launch, and the metrics are just how you decide who to call.

When should you skip the soft launch entirely?

When you already have the answers. If you are launching a second product to an audience that already buys from you, through a channel you already understand, a soft launch is theatre that delays revenue.

The same applies when the product is small enough that shipping it is cheaper than testing it. Not everything deserves a launch process. A feature, a template, a small tool: put it out, see what happens, and spend the saved effort on something bigger.

The other legitimate skip is when you have run something structurally similar already. A design partner programme, for instance, can do the same job earlier and with more depth, and I have written about running one in how to run a design partner program before launch.

Across more than 70 projects for more than 25 clients over more than six years, the pattern I trust most is a modest soft launch result followed by a specific, named fix. Enthusiasm without a diagnosis has never predicted much. A soft launch tells you whether the offer works on a small number of strangers in one channel; it does not tell you whether the market is big or whether the channel scales.

What should you do next?

Before you plan anything else, write one sentence that says what would make you cancel the hard launch. If you cannot write it, you do not yet know what you are testing, and that is the real work of this week.

Then pick one channel, one segment, one page and one date. Write the pass mark next to it. Run it, and resist the urge to help the test succeed by personally selling to everyone who shows up.

If you want someone to pressure test your launch plan before you commit budget to it, reach out. I am happy to tell you which of the three questions your plan is actually answering, and which one it is quietly skipping.