A Series of Three Incrementality Tests Prove Brand Search to Be Ineffective

This brand was spending a significant amount of money on Google brand search and wanted to understand if this tactic was driving incremental customer activations. They also wanted to understand if these results changed due to seasonality or if brand search was more effective in new expansion markets.

‍

The brand decided to run a series of 3 incrementality tests - one during slow season with all markets, one during peak season with all markets, and one specifically for new markets where they had not yet launched their product. They designed the tests to be 2-cell experiments with a holdout with the structure listed below:

Test 1: Slow Season

Test 2: Peak Season

Test 3: New Market Launch

The initial test ran for 2 weeks and showed the slightest bit of lift - 1% lift at a CPA that is highly inefficient for the brand. The second test ran during peak season and resulted in no lift to the business at all. Finally, the third test ran in new markets with less established awareness and also resulted in no lift. After this series of tests the brand felt comfortable turning brand search off entirely and reallocating those media dollars to other incremental channels.