The model is the question, the tactic is the answer

The most expensive mistake in growth is not a bad tactic. It is a good tactic aimed at the wrong stage, run with conviction, for a quarter. A founder who believes their problem is volume will buy more traffic. If the real leak is in activation, every new visitor simply leaks faster, and now the leak is also more expensive. The tactic worked, the campaign converted, the dashboard turned green, and the business did not move. That is the signature of running inside a model nobody chose.

A growth model is the lens that decides which numbers you are allowed to care about. It does not tell you the answer; it tells you which answers are even on the table. This is why the model is the question. Pick the acquisition lens and you watch cost-per-lead and click-through. Pick the retention lens and those numbers fade into the background while churn and expansion move to the front. The lens is upstream of every tactic, every dashboard, and every metric you call important.

Sequence is a law, not a preference. Strategy before tactics gets dismissed as consultant-speak, so let me put it plainly: it is a sequencing law, because each growth phase creates the preconditions for the next. You cannot meaningfully optimise referral before you have retention, because there is no one left to refer. You cannot scale acquisition into a product that does not activate, because you are filling a bucket with a hole in it. Running a later phase before an earlier one is the primary cause of wasted spend and stalled growth. The order is causal, not stylistic, which is why finding your one growth bottleneck is the move that makes every later decision obvious.

This is freeing, not constraining. It means most of the tactical decisions you agonise over do not actually need a decision, because the model already answers them. Once you know you are in an activation-first phase, you stop debating whether to spend on paid search; the answer is obviously not yet, because more top-of-funnel into a broken activation step is the textbook waste. The model collapses a hundred small arguments into one, and it does so before you have spent a euro proving the obvious the hard way.