What is a pivot, and when should you make one?
A structured change of direction that keeps what you have learned while changing a fundamental assumption. Eric Ries's definition is useful precisely because it excludes the two things people mistake for pivots: giving up, and drifting.
The types, roughly:
Customer segment pivot. Same product, different customers. Frequently prompted by discovering that an unintended group uses it most enthusiastically — and this is the most common successful pivot, because the evidence arrives unasked.
Problem pivot. Same customers, different problem. You understand the audience and were solving the wrong thing for them.
Feature to product pivot. One feature is used far more than the rest, and becomes the whole business. Several well-known companies began as a component of something else.
Platform pivot, from application to platform or the reverse.
Business model pivot — the product works, the way you charge does not.
Channel pivot, changing how you reach customers.
Technology pivot, delivering the same value a different way.
When a pivot is warranted: sustained lack of retention despite fixing the obvious problems; sales conversations that consistently stall at the same point; a market that turns out to be far smaller or less willing to pay than assumed; and — the positive case — a strong signal from an unexpected direction that you keep ignoring.
When it is not:
Boredom. Founders tire of a problem long before customers do.
Insufficient time. Most things take longer than expected, and abandoning at the point of ordinary difficulty is not a pivot.
Chasing whatever is fashionable, which is the most expensive version.
How to do it well:
Keep what you learned. A pivot that discards all accumulated knowledge is a restart, and you would be better off with more runway and a clean sheet.
Change one major variable, so you can tell what happened.
Decide in advance what would constitute evidence, and set a review point — deciding while exhausted and nearly out of money produces bad choices.
Count the runway honestly. Pivoting with three months left is rarely enough to reach evidence.
Tell investors early. They have seen many pivots and generally respond better to a reasoned case than to a surprise.
General information, not business advice.