There is a reason the last ten products you signed up for felt like the same product. They ran the same onboarding sequence, sent the same drip of emails, offered the same referral credit, and posted the same kind of founder story on the same platforms in the same tone. Their founders read the same threads, attended the same programmes and copied the same case studies, and they did it because the case studies were genuinely good. The problem is not that the tactics were wrong. The problem is that a tactic everyone runs is, to the person on the receiving end, indistinguishable from the tactic everyone else runs, and a product that is indistinguishable in how it reaches people will be indistinguishable in how it is remembered, which is to say not at all.

Borrowed tactics decay for a simple reason. Every one of them worked, originally, because it exploited a gap between what a customer expected and what the company did. The gap was the arbitrage. It closed the moment the tactic was written up and enough companies read the write-up, and after that the tactic produced the average result, which is what it produces for everyone now. Reading the same post as your competitors and running the same play is not a strategy. It is a decision to be average on purpose, and the market obliges.

Generating volume without a pointless brainstorm

The alternative is not to find the one clever idea. It is to generate a great many cheap ones, most of which will be bad, and to have a fast way of finding the few that are not. The instinct here is to hold a brainstorm, and brainstorms are where ideas go to be politely ignored: a room, a whiteboard, an hour, twenty suggestions of which the loudest three get discussed and none get tested. Volume needs a different mechanism. Ask everyone on the team, individually and in writing, for ten ideas that a competitor would never try, with the explicit permission to be strange. Fifty ideas arrive in a day, unfiltered by the social pressure of a room, and the strange ones are the interesting ones because strangeness is exactly what the shared playbook cannot supply.

The constraint that produces useful ideas is specificity. Not what would grow a company, but what would grow this company, with this product, for these users, given the odd things that are true about us and false about everyone else. A product with an unusual feature, a founder with an unusual background, an audience with an unusual habit, all of these are raw material for moves that nobody else could run, and the exercise is to mine them deliberately rather than to reach for the generic move because it is closer to hand.

The questions that kill an idea early

Fifty ideas need a filter that works in minutes, because the point is to avoid spending a sprint on something that a few questions would have killed. Would this be worth doing even if it only half worked. Can we tell within a week whether it is working. Does it depend on anything we do not control. Would we be embarrassed if a competitor copied it, or would we be pleased that they were now doing something that only suits us. And the sharpest one: does this idea exist because of something specific about our product and our users, or could it be lifted onto any product in the category without changing a word.

Reading the same post as your competitors and running the same play is not a strategy. It is a decision to be average on purpose, and the market obliges.

The filter is meant to be run fast and without sentiment. An idea does not need a defence; it needs to answer five questions, and an idea that needs ten minutes of argument to answer them is telling you something. The team that gets good at this can clear fifty ideas in an hour and be left with four or five worth a week each, which is a better use of the hour than any brainstorm has ever been.

Most ideas die at the last question, and they should. An idea that could be lifted onto any product is a borrowed tactic that has not been borrowed yet, and running it first buys a few weeks of edge before the write-up appears. The ideas that survive are the ones rooted in something true only here, and those are the ones worth a week of someone’s time.

Running small

The idea that survives the filter should be tested in days, not quarters, and this is where most companies fail through excess ambition. A growth experiment is not a product launch. It is the smallest version of the idea that can produce a signal: one email to one segment, one page for one audience, one feature flag for one cohort. If the smallest version needs a fortnight of engineering before anyone can see a result, it is not the smallest version, and the work is to shrink it until it is. Speed here is not about urgency. It is about the number of ideas a team can test in a year, which is the real determinant of whether it finds the ones that work.

Small tests have a further advantage, which is that they are cheap to be wrong about. A team that has spent a quarter on an experiment will find reasons to believe it worked. A team that has spent three days will read the result honestly, because nobody’s reputation is riding on it, and honest reading is the whole point of testing.

The discipline that makes this possible is a fixed and short review. Every week, what did we test, what did it show, what do we drop and what do we scale. Ideas that showed nothing are dropped without ceremony. Ideas that showed something get a slightly larger version. Nothing lingers, because lingering experiments consume the attention that new ones need, and attention is the constraint.

Shipping what competitors would never try

Over time this produces a habit that is the actual competitive advantage: a team that routinely ships things its competitors would consider odd, because those things are rooted in what is specific about the company and the competitors have nothing to root them in. Some of these will fail visibly, and the team has to be comfortable with that, because a company that only ships moves it is certain of is a company that only ships moves everyone else has already validated. The failures are the price of the successes, and the successes are the only moves in the market that cannot be copied without changing the product.

The playbook will keep circulating, and it is worth reading, as a catalogue of what has already been done. But the companies people remember are the ones that did something the playbook did not contain, because it could not: it was specific to them. Generate volume, filter hard, test small, review weekly, and ship the strange ones. That is how a company stops being one of the ten products that felt the same.