Four angles beat forty creatives

The standard advice for Meta is to ship more creative. It is not wrong, exactly — creative is the main lever left — but it gets applied as raw volume. Forty assets go into a campaign, three get delivery, and the account learns almost nothing it can carry into next month.

The reason is that most of those forty assets are the same claim wearing different clothes. Different crop, different colour, same promise. When they lose, you have not falsified anything; you still do not know whether the promise was wrong or the execution was.

Volume is not a hypothesis

An angle is a claim about why someone buys. “You will not be sold to on the call.” “We can see you this week.” “Fixed price, quoted before we start.” Each of those is falsifiable: run it, and the market tells you whether that reason moves people.

A variation is a different rendering of the same claim. Variations are how you exploit an angle that works. They are not how you discover one, and running twenty variations of an angle nobody cares about is the most common way to spend a test budget on nothing.

Creative volume exploits an angle. It never discovers one. Decide which job you are doing before you brief anything.

Where angles actually come from

Almost every angle worth running is already written down somewhere the business owns. Review replies name the objection that nearly lost the sale. FAQ pages exist because the same question kept arriving. Sales notes record the sentence that closed it.

This matters for more than convenience: an angle sourced from your own site is one you can substantiate. If the claim ever gets challenged — by a platform reviewer, by a customer, or by your own legal team — there is a page behind it rather than a copywriter’s guess.

One concept per angle, then iterate

Run each angle as a single strong concept first. If an angle earns delivery and holds its cost per result, then produce variations — that is when volume pays, because you are exploiting something you have evidence for.

Keep the angle named in the ad set or the naming convention, not just in a brief nobody reopens. Six weeks later the only question that matters is “which claims worked,” and that answer should be readable from the account rather than reconstructed from memory.

Knowing when you are out of angles

Running out of angles is a real state, and it has a signal: new concepts stop changing the cost per result in either direction. At that point additional spend is buying more impressions of a message the audience has already answered.

The right response is not a higher daily budget. It is either new evidence — a new offer, a new proof point, a new segment — or a shift of the marginal pound to the platform that still has headroom. That is a budget decision, and it belongs to whoever owns the split.