A tour operator spends £5,000 on a new channel. The platform reports 200 bookings at £25 each, comfortably below the £35 the team had agreed it could afford. The obvious next step is to increase the budget.

Before you do, ask what happened to total bookings. If the operator would have sold 120 of those 200 bookings anyway, the campaign added 80. That puts the cost per additional booking at £62.50. The same spend that looked like a reason to scale now gives you a reason to hold.

You won’t know that 120 is the right number just by opening another dashboard. To get a credible answer, you need to design the test before the campaign starts.

Start with the decision you’ll make

“We want to see if this channel works” is too loose to guide a test. Start with the budget decision you’ll face at the end. For the tour operator, that might be: “Should we keep spending £5,000 a month on this channel to fill weekday departures that usually have spare seats?”

That gives you a hypothesis you can challenge: the channel will generate enough additional weekday bookings to justify the spend. At a ceiling of £35 per additional booking, £5,000 needs to deliver at least 143. Set that ceiling using what the business keeps after refunds and the variable costs of serving those customers. If the campaign sends people towards Saturday departures that already tend to sell out, those bookings have very different economics. Decide upfront which departures count.

Give yourself a fair comparison

Once you know which bookings matter, work out how you’ll estimate what would have happened without the campaign. If you can randomly split an eligible audience into a group that can receive the ads and a group that can’t, compare booking rates across those groups. If that isn’t practical, you might test in some comparable markets and hold others back. Check that booking trends were reasonably similar before the test, and watch for differences in school holidays, promotions and availability that could skew the result.

Use your booking system to count the agreed outcome across both groups. The channel dashboard is still useful for seeing which ads ran and what the platform credited to them. The booking comparison answers the question you’re spending money to resolve: did running the campaign improve the outcome for the business?

Check whether the test can settle the decision

The budget tells you how much you can afford to risk. The volume of bookings and how much it normally varies tell you whether the test is likely to give you a useful answer. If weekday bookings routinely swing by more than the gain you need to detect, a short test may leave you unable to distinguish a worthwhile result from an ordinary fluctuation.

Check that before you commit. A longer test, more comparable markets or a different time of year might give you a clearer read. If none is practical, run a smaller pilot with a narrower purpose. You can learn whether the channel reaches the intended audience, whether the creative gets a response, and whether bookings can be tracked reliably. That may be enough to decide whether a larger test deserves funding.

Agree what would make you stop

Write down the spend limit, end date and how long you’ll allow for bookings to come through. Agree when you’d pause early, for example if tracking breaks, the campaign reaches the wrong audience or the departures you’re trying to fill sell out. Then agree what you’ll do with the result.

If the evidence supports enough extra bookings at an acceptable cost, put the next tranche of budget behind the channel and check whether the economics hold as spend rises. If a worthwhile result looks unlikely, stop. If the result could plausibly go either way, call it inconclusive and extend the test only if you can explain what the extra spend or time will resolve.

The most useful line in the test brief is the one you write before seeing the results: “Here’s what would make us decide against funding this channel.” If you can’t finish that sentence, the campaign may already have won its budget.