Do Not Remove the Constraint That Defines Your Product

When growth slows, almost every product rule starts to look negotiable. A required step becomes friction. A narrow audience becomes an addressable-market problem. A deliberate limitation becomes something a competitor does not have.

This is often sensible. Products accumulate rules that no longer serve customers. But some constraints are not leftover process. They are the product.

That distinction matters in the current market. A dating app is removing the rule that helped define its interaction model. A social network is expanding its focus while its active user base is reportedly shrinking. At the same time, product teams are sharing stories about useful B2B AI products that still failed, hardware concepts that should not have been built and beta tests in real restaurants that exposed a different source of value than expected.

I see one pattern across these developments: teams are very good at finding friction, but less disciplined about identifying what the friction protects.

A signature constraint can do several jobs. It can set expectations, attract a particular customer, shape community behaviour or make the product easier to explain. Removing it may increase the number of actions people can take while reducing the reason they would choose this product over another.

That is the trap. More possible behaviour is not automatically more customer value.

Consider a marketplace that requires suppliers to respond within a set period. Suppliers will complain that the rule is inconvenient, and they may be right. Removing it could increase supplier sign-ups. But if the response requirement is also what makes the marketplace dependable for buyers, the team has not removed friction. It has transferred that friction to the other side of the market.

The same applies to B2B software. A product may enforce a structured approval step that users call slow. It is tempting to make the step optional. Yet that step may be the reason a finance or compliance team accepts the product at all. The visible inconvenience belongs to one user; the less visible assurance belongs to another.

This is why usefulness alone is a weak test. Many products are useful. Far fewer are clearly preferable. AI makes this problem more obvious because a useful capability can be copied, bundled or accessed elsewhere. If the product has no strong point of view about who it serves and how the work should happen, usefulness quickly becomes replaceability.

Before removing a defining rule, I want a team to answer three questions in plain language. What customer promise does this rule support? Which users selected the product because of that promise? What will preserve the promise when the rule is gone?

The third question is usually where the proposal becomes vague. Teams are often precise about the expected gain in conversion and imprecise about the replacement mechanism. They know what they want to remove, but not what must take over its function.

A small experiment can help, but only if it reaches the real environment. Controlled testing may show that customers complete a flow faster without a constraint. It may not show that the marketplace becomes less reliable, the community changes character or a business buyer loses confidence over the following months. The restaurant beta stories are a useful reminder: context does not merely validate a solution. It reveals what the product is actually valued for.

I would rather make a constraint adjustable than simply delete it. Test it with a specific segment. Replace the assurance before removing the step. Watch whether the customers who valued the original promise stay, not only whether new users act. Speak to people who chose the product precisely because it behaved differently.

Product strategy is partly the choice of which friction to keep. If every uncomfortable rule is optimized away, the result may be easier to use and harder to care about. Growth pressure makes simplification attractive. It should also make us more careful about deleting the reason customers arrived in the first place.