A pivot is a deliberate change to one core element of your business: the customer, the problem, the product, or the revenue model. You keep what you have learned and change one variable. A pivot is not giving up, and it is not a small tweak to your pricing page.
The signal to pivot is evidence, not mood. Look at three things: how many people you have spoken to, how many of them paid, and whether the ones who paid stayed. An example. After 12 months you have had 300 sales conversations, 40 free trials, and 3 paying customers, and 1 of those 3 cancels every few months. That is not a marketing problem you can fix with a bigger advertising budget. The offer does not match the need.
Pivots have types. You can keep the product and change the customer, for example moving from private users to companies. You can keep the customer and change the product, for example turning one popular feature into the whole product. You can keep both and change how you charge.
A worked example of the last type. An agency sells custom dashboards for 4,000 per project and closes about one project a month, so revenue is 4,000 per month and every project must be sold individually. It packages the same work as a self-service product at 79 per month. To reach the same 4,000 per month it needs 51 subscribers. That takes longer to build, but it does not stop when the founder takes a holiday. Writing both numbers down makes the decision discussable instead of emotional.
Two mistakes. Changing everything at once means you learn nothing, because you cannot tell which change worked. And calling every small adjustment a pivot hides the fact that nothing was really tested. Change one variable, set a deadline and a target, then judge the result.
