November 16, 2026 · 7 min read · By Chaminda Delpagodage

Stick or Quit: How Long to Stay With a Niche Before You Change It

Quit a niche too early and nothing compounds. Stay too long and you waste a year. A practical way to tell a dip from a dead end, with kill criteria set in advance.

There are two ways to lose with a niche. The first is to leave too early, again and again, so that nothing you build ever gets old enough to work. The second is to stay too long in something that was never going to work, because leaving would mean admitting it.

Most advice picks one side. “Never quit” produces founders three years into a dead end. “Fail fast” produces founders on their ninth niche. The evidence says both extremes lose: in the Startup Genome Report’s study of more than 3,200 startups, those that pivoted once or twice outperformed both the ones that pivoted more than twice and the ones that never pivoted, raising 2.5 times more money with 3.6 times better user growth.

So the real skill is neither persistence nor flexibility. It is telling, while you are in the middle of it, which one the situation calls for. This post is a practical method for that. If you want the case against hopping first, read The niche slap.

The dip and the dead end

Seth Godin’s short book The Dip (2007) gives the cleanest vocabulary for this. Anything worth doing has a dip: a long stretch after the fun start where effort goes up and visible reward goes down. The dip is not a sign that something is wrong. It is what keeps most people out, which is why the far side of it is valuable.

A cul-de-sac is different. Effort goes in and nothing changes, this month or next year. There is no far side.

Godin’s advice is to quit dead ends quickly and without guilt, and to refuse to quit dips. The difficulty, which critics of the book point out, is that from the inside the two feel the same. Tired, underpaid and unsure is how both of them feel on a Tuesday.

Why you cannot decide in the moment

This is where Annie Duke’s Quit (2022) is useful. Duke, a former professional poker player who studies decision making, argues that people are poor judges of when to quit while they are in the situation. Sunk cost pulls one way. Frustration pulls the other. Neither is evidence.

Her answer is to make the decision before you need it. She calls the tool kill criteria: conditions, written down in advance, under which you will stop. The simplest form pairs a state with a date: if I have not reached a named result by a named day, I quit.

Kill criteria protect you in both directions. They stop you leaving in week three because a post flopped, since the date has not arrived. And they stop you staying in month nine out of pride, since the date has passed and the result is not there.

A stick-or-quit protocol for your niche

1. Define the niche tightly enough to test

“Founders” is not a niche. A testable niche names a buyer, a problem and an offer: bootstrapped SaaS founders, churn in the first ninety days, a teardown of their onboarding. If you cannot say what a win would look like, you cannot tell later whether you got one.

2. Set a minimum term

Choose a period long enough for a fair sample and short enough that you will actually hold to it. For a niche on X, ninety days of consistent posting is a reasonable floor. Reach on individual posts varies too much for two or three weeks to tell you anything. During the term, the niche is not up for discussion.

3. Choose your evidence before you start

Likes are the wrong measure. They reward breadth, and a niche is narrow by design. Better signals:

  • Replies and messages from the specific kind of person you are trying to reach.
  • Profile visits that turn into follows, which shows the profile matches the posts.
  • Inbound questions about the problem you say you solve.
  • Calls booked, trials started, money received.

4. Write the kill criteria down

One sentence, with a number and a date. For example: if I have not had ten conversations with qualified buyers by the end of February, I stop. Put it somewhere you will see it. The value comes entirely from having written it while calm.

5. Do the hardest part first

Duke borrows a rule from Astro Teller at Google’s X lab: if the plan is to have a monkey reciting Shakespeare on a pedestal, train the monkey before you build the pedestal. For a niche, the pedestal is the banner, the bio and the landing page. The monkey is whether anyone in that market will pay. Test that in the first weeks, not the last.

6. Change tactics freely, not the market

Committing to a niche does not mean repeating what is not working. Inside the term you should change hooks, formats, posting times, the wording of the offer and the price. What stays fixed is who it is for and what problem it solves. That is the line between iterating and hopping.

7. Review once, on the date

When the term ends, there are three honest outcomes:

  • Stay. The evidence is there or clearly building. Set the next term.
  • Narrow. Part of the niche responded and part did not. Hormozi’s advice in $100M Offers is to commit and then narrow further. Going from “founders” to “SaaS founders with churn” is not a hop.
  • Quit. The criteria were not met. Leave cleanly, write down what you learned, and start the next one with a new term and new criteria.

Good and bad reasons to leave

Reasons that count as evidence:

  • You hit the date and missed the result you named.
  • You have spoken to real buyers and they do not have the problem, or will not pay to solve it.
  • The market is real but you cannot reach it at a cost you can bear.

Reasons that are only mood:

  • A bad week of impressions.
  • Someone else’s screenshot from a different niche.
  • Boredom. Boring is often what the middle of something working feels like.
  • A new idea that has not yet had the chance to disappoint you.

For a worked example of a quit made on evidence rather than mood, I wrote up why I shut down CampfireX after 8 days: 200 clicks, zero customers, and a pre-pay test that said no.

Check what you already left

One more step before you choose the next niche. If you have changed direction several times, there is a fair chance that one of the earlier niches was working better than you noticed, and that you left it during a dip. The evidence for that is in your own posting history: which period earned the most views and replies for each post, and what you were talking about then.

That comparison is hard to do from memory, because memory keeps the mood and loses the numbers. It is the reason NicheSlap exists: it lays every niche you have tried on one timeline, shows which one was working, and, if you choose one, holds you to it for thirty days.

If the honest answer is that you never found a problem worth staying with, start earlier. IkigaiNiche helps you pick the person and the problem first.

Sources

  • Seth Godin, The Dip: A Little Book That Teaches You When to Quit (and When to Stick) (2007).
  • Annie Duke, Quit: The Power of Knowing When to Walk Away (2022).
  • Alex Hormozi, $100M Offers (2021).
  • Startup Genome, Startup Genome Report (2011).

Count your own niche changes.

Free. We read a year of your public X posts and show every niche you tried, and the one that was working.

Reading a year of posts takes a few minutes. We email you the link when it's done. No newsletter. Privacy · Terms

Read next