e27 published a piece on 9 November 2026 called “Almost” is a strategy, not a stall. The author describes a founder they worked with who spent four months “about to” shut down a product line. The line had not worked in over a year. The numbers were not ambiguous, and she was not missing information.
The article’s account is that every week brought one more reason to wait. One more customer call to make first. One more pricing test to run. One more quarter to be sure. I could only read the feed summary, because the full article would not load for me. So I am working from that and nothing more.
The claim is that “almost” is a strategy
The framing is that waiting can be deliberate. Sometimes it is. If a decision depends on a fact you don’t have, getting the fact is sensible.
So the test is simple. What would the next customer call or pricing test tell her that a year of numbers hasn’t? According to the summary, nothing. She had the information. The extra steps did not add to it. They postponed the decision.
I’d call that a stall, whatever name it is given. A strategy has a point at which it ends. “One more quarter” has no such point, because there is always another quarter.
Why this happens to bootstrapped founders more than anyone
When you fund the business yourself, every month a failing product stays alive is paid for from money you earned elsewhere. That makes bootstrapped economics a filter. It tells you early what works and what doesn’t. Paying $53 per install was never going to work for AeroChat, so the answer had to be distribution that compounds, which for us meant AI search and not ads.
The filter only works if you let it filter. A founder who keeps asking for more evidence after the evidence is clear has switched the filter off. The reason is usually not data. Closing a product line means admitting it didn’t work. Waiting feels better than that.
I have shut things down myself, and I keep the failures in the story when I teach. It is more useful to people than a record of wins.
Set the kill line before you launch
Take a clinic that starts selling a new wellness package. Before the first booking, the owner writes down three things.
- The number of bookings per month the package must reach.
- The month by which it must reach that number.
- What the owner will do if it misses, which is to stop selling it.
This is written when the owner is hopeful and has not yet spent anything on the idea. It is much easier to be honest then than in month ten, when staff have been trained and brochures printed.
The line should be a number, a date and an action. “See how it goes” is not a line. Neither is “review when we have more data”. If the threshold is vague, you will bargain with it every week.
A marketing team can do the same with a campaign or a channel. A shop can do it with a product range. A service business can do it with a new offer. The structure is identical.
Let the dashboard count and keep the decision
Once the line exists, the counting is a chore, and AI is good at chores. I let Claude do the building, the counting and the checking, and I make the decisions. The same split works here.
Set up a weekly or monthly report that pulls the figures you named and compares them to the line. It should say plainly whether you are above or below, and by how much. It should not offer reasons to wait. No one has to feel bad about reading it, because the report simply states where you stand.
The report does not make the call. Stopping a product line depends on things a dashboard can’t see. Maybe the product feeds another one. Maybe a key customer depends on it. Maybe you have a reason to carry the loss for a while. Those are your judgements, and you are responsible for every decision that leaves your hands.
But if you choose to override the line, do it on purpose and write down why. That is different from drifting past it for four months while telling yourself you are gathering information.
What a business should do
Pick one product, offer or campaign that is running now. Write down the number it needs to hit, the date, and what you will do if it misses. If it has already missed, take that as the answer and don’t plan another test.
For anything you launch next, write the line first. Then have a report sent to you on a fixed schedule that shows the figure against the line. When it comes back below, decide that week.