The figures in this piece are illustrative. They show the method, not a real account.

Your branded search campaign probably looks like your best performer. Is it?

Last-click attribution gives all the credit for a sale to the final touch before it.1 For branded search, that last touch is often someone who’d already decided to buy.

Two numbers, two stories

In this illustrative example, last-click gave branded search 71% of the conversions. A holdout test gave it 38%.

Figure 1. Last-click credited search with 71% of conversions. A holdout credited it with 38%.Illustrative figures

The gap is the credit last-click hands to search for sales that would probably have happened anyway.

How a holdout works

You switch the channel off for a random group of regions, people or time slots, and leave it on for everyone else. The difference in conversions between the two groups is what the channel really adds.2

The method is simple. The discipline is in the set-up: enough volume, a long enough period, and groups that are genuinely comparable.

What to do with the gap

If the holdout credit is far below the last-click credit, your brand budget is probably bigger than it needs to be. Test one market first, not the whole account.

Limits

  • One market and one period. Seasonality can move the result.
  • Small samples give wide intervals. Report the interval, not just the point estimate.
  • Switch off brand search and a competitor may start bidding on your name. Watch for it during the test.

Footnotes

  1. Platforms now offer data-driven attribution, but last-click is still the default view in many finance reports. ↩

  2. Region-based holdouts are often easier to run than audience-based ones, because they don’t depend on user-level tracking. ↩