What to Do After a Failed Launch: Diagnose the Funnel Before You Decide
After a launch that got little response, do not delete it, repost it or buy reach. Wait a day, then find where the funnel broke: nobody saw it, people saw it and left, people tried it and did not return. Each failure has a different fix. Within a week you can decide to change the message, redirect to a different audience, or stop with the lessons written down.
What did “failed” actually mean?
Quiet launches feel identical from the inside and have different causes. Write down which of three happened, because the next move depends on it. Nobody saw it. People saw it but did not click or understand. People tried it and did not come back. A launch can fail one stage and succeed at the others, which is useful, because it tells you where to look.
Also check whether it failed against a fair standard. A new product with no audience posting in one place usually collects a handful of visits. That is ordinary, not a verdict, and the answer on why nobody visited a launch goes through the usual causes before you assume the product is the problem.
What should you do and not do in the first 48 hours?
- Do not delete the listing. It is a public page with your name on it, and a quiet page costs nothing to leave up. An edit is better than a deletion.
- Do not repost the same message. Repeating it in the same places looks like spam and teaches you nothing new.
- Do not buy reach. Paid placement sells position. If the sentence did not land with the people who saw it, more people will not fix that.
- Do not rewrite the product. One result is not a mandate for a rebuild.
- Do write down the numbers, link by link, while you still remember what you did. Then close the laptop for a day.
The pause is practical, not sentimental. A plan written while disappointed tends to be either a panic or a dramatic quit, and both are bad decisions made on one data point.
Where did the funnel break?
| Stage | What you saw | Likely cause | Test this week |
|---|---|---|---|
| Reach | Very few visits to any link | Too few people saw a post, or it went to the wrong place | Send the link directly to ten people who have the problem and count |
| Click | Many saw it, few clicked | The tagline or title did not say who it is for | Rewrite it and ask five people what they think it does |
| Understand | Clicked, left within seconds | The page did not explain itself | Watch three people open it silently and note where they pause |
| Try | Read the page, did not start | Too much friction, or a promise they did not believe | Remove the sign-up from the first step; add a real result |
| Return | Tried once, never came back | It worked, but not well enough to matter | Ask the people who tried it what they did instead the next day |
Put your own numbers into the directory traffic calculator to find which stage lost the most people. The launch metrics to track page explains how to count each stage. If the break is early, the fix is cheap. If it is at the last stage, the fix is about the product, and it is worth knowing before you spend months.
What does a seven-day repair look like?
- Day 1. Write the retrospective: what you did, what happened, what you assumed. The launch retrospective template gives a structure.
- Day 2. Talk to three people who tried it, or three who would have been right for it. Ask what they were doing when the problem came up, not whether they like the product. See getting feedback on an MVP.
- Days 3 and 4. Change one thing at the stage where the funnel broke. Only one, so you can read the result.
- Day 5. Show the changed version to ten new people who have the problem, personally and without asking for votes.
- Day 6. Compare against the first attempt, stage by stage.
- Day 7. Decide, using the rules below.
How do you decide: fix, redirect or stop?
| What you found | Decision | First step |
|---|---|---|
| People who see it understand it and a few try it; the break was reach | Fix | Pick one new place and reach people there, with a tagged link |
| The product works for a group you did not target | Redirect | Rewrite the page for that group; see the relaunch playbook |
| People try it once and say it is nice, but nobody returns | Change the product | Ask what they used instead; build the one thing that replaces it |
| After a clear sentence and the right people, nobody cared | Stop, or park it | Keep the notes and the page; start something else |
How do you start again without repeating the mistake?
Pick a smaller question next time, and ask it with fewer people. Before the next launch, find ten people who have the problem, and see whether any would use a rough version. The validation planner prepares the questions. When you are ready to publish again, a free listing such as the launch board costs a form and an emailed link, so you lose little if it is quiet again.
Keep your first launch in your records. It shows you can finish and publish, and the lessons make the second version better. The first 100 users playbook is a sensible next step when you are ready to try again.
Frequently asked questions
Should I delete a launch that got no response?
No. A quiet page costs nothing to keep, and deleting it removes a public record of what you finished. If the wording is wrong, edit it. Use the repair week to learn why it was quiet and decide whether to change the message, redirect to another audience or stop.
Is a launch with few visits a failure?
Not necessarily. A new product with no audience that posts in one place usually gets a small number of visits, which is ordinary. Judge it by stages: did people see it, understand it, try it and return? Find the stage that lost the most people, because that is where the fix is.
Should I relaunch right away?
No. A relaunch without a substantive change repeats the message and the result. First find where the funnel broke, change one thing at that stage and test it on ten new people. Relaunch only when you have a different reason to try it, a different audience or new proof.
How do I know when to stop working on a product?
When, after a clear sentence and the right people, nobody beyond your circle used it twice, and a changed price, user or promise did not alter that. Park it with the notes, not in shame. Pick a smaller question for the next product and ask ten people before building.
Written by Post your build ideas with AI assistance and edited for this site. Where a page names another product, the facts came from that product’s own pages on the date shown above; terms change, so check them before you rely on them. About the publisher.