When to pivot or quit a startup idea

Why this decision is so hard to make well

Most founders do not quit too early or too late because they lack information. They struggle because the decision is tangled up with identity, time already spent, what they told friends and family, and the fear that success was just one more month away. Each of those is understandable, and none of them is evidence about the idea.

The way out is to make the decision partly in advance, while you are calm and before the sunk costs pile up. You cannot predict everything, but you can commit to what you will look at and what results would trigger a rethink.

Set kill criteria up front

Kill criteria are specific, measurable conditions, agreed before you start a phase of work, that would make you stop or change direction. They work because they are written when you have nothing to defend.

Good kill criteria have three parts: a metric, a threshold and a date. An illustrative example: 'By the end of March, at least eight of the twenty businesses in our pilot are paying, and at least half of those are still active after a month.' Or for an earlier stage: 'After fifteen interviews, at least ten describe the same workflow and can name what they currently pay to deal with it.'

Choose criteria that test your riskiest assumption, set them at levels that would genuinely change your mind, and share them with someone who will hold you to them, such as a co-founder, adviser or a friend who is not invested in the idea. Write down your personal limits too: how much money and how many months you are willing to spend.

Run the review like a meeting, not a mood

Put the checkpoint in the calendar when you set the criteria, and treat it as a formal review even if you are a solo founder. Before the meeting, gather the numbers and a short summary of what customers said, then compare them line by line against what you wrote. Do this on a normal working day, not the evening after a bad sales call or the morning after a good one.

Write the outcome down in a sentence: persevere, pivot or stop, and why. If you decide to persevere, set the next criteria and date there and then. A decision recorded in writing is much harder to quietly reverse a week later when your mood changes.

Persevere, pivot or stop

The difference between persevering and drifting is direction. Slow progress on the metric that matters is a reason to continue. Lots of activity with no movement on that metric is not.

Types of pivot

A real pivot keeps something you have validated and changes one major thing. Changing everything at once is not a pivot; it is a new idea, which is fine, but should be treated as starting again with fresh validation.

• Customer segment — Same product, different buyer. Often the most productive pivot, because it keeps what you built.
• Zoom in — One feature becomes the whole product.
• Zoom out — The product becomes one feature of something larger that solves the problem more completely.
• Problem — Same customers, a different and more painful problem you discovered while talking to them.
• Business model — Same value, different way of charging, for example usage-based rather than subscription, or services rather than software.
• Channel — Same product, different way of reaching people, such as direct sales instead of self-serve.
• Technology — Same problem and customers, a different way of delivering the result.

Emotional traps

• Sunk cost — 'We have already spent a year on this.' Time already spent is gone either way; only future time is the decision.
• Moving the goalposts — Lowering the threshold after missing it. If you change criteria, say so explicitly and say why.
• One more feature — Believing the missing piece is always the next thing to build.
• Anecdote over pattern — Holding on because of one enthusiastic customer while the rest drift away.
• Identity — Feeling that quitting the idea means failing as a person. Stopping a failing experiment is a competent decision.
• Survivorship stories — Remembering founders who persisted and succeeded, and forgetting the many who persisted and did not.

How to stop well

If the answer is to stop, do it cleanly. Tell customers early and honestly, give them time and help to move their data, and refund where it is fair. Settle obligations to suppliers and anyone you employ, and close any company or accounts properly, taking professional advice where needed.

Then write down what you learned: about the market, the customers, the channels and yourself. Many founders find that their next, better idea came directly from something they discovered while working on the one they stopped.

Questions