How to write a marketing hypothesis you can actually test

Every marketing team has more ideas than budget. The problem is rarely a shortage of things to try — it's that most ideas are launched in a form that can't produce a clear answer. A campaign goes live, numbers move a little, and three people read the result three different ways.

A hypothesis fixes that before any money is spent. It forces you to say what you expect to happen, how you'll know, and why you believe it. That last part matters most: the 'why' is what you learn, whether the test wins or loses.

We write every test in the same three parts:

• If we… — the one specific change you'll make (an audience, an offer, a page element).
• Then… — the measurable result you expect, in the metric that decides the test.
• Because… — the belief about your customers that explains why it should work.

“Better creative will improve results” can never fail, so it can never teach you anything. “If we lead with the customer's problem instead of our features, click-to-lead rate on cold audiences will rise, because cold audiences recognise their problem before they recognise our product” can fail — and that's exactly why it's useful.

A good test of specificity: could two people disagree about whether the hypothesis was proven? If yes, tighten it.

Pick a single number to judge the test, and pick it before launch. If you track ten metrics, one of them will look good by chance — and it's very tempting to declare victory on that one.

Where you can, choose a metric close to revenue: qualified leads or booked calls rather than clicks. Upstream metrics are fine as diagnostics, but they shouldn't decide what gets budget.

Once ideas are written as hypotheses, they're easy to compare. We score each one on how much it could move the deciding metric, how confident we are in the 'because', and how cheap it is to test. The top of that list is your next sprint.