This would clearly be a disaster for television ad executives. That’s because employing such metrics would miss the whole point of a TV ad, which is to stoke consumer demand and influence brand perceptions. After all, it takes time to sway the hearts and minds of consumers who will be buying these products in the future.
When it comes to search engine advertising, marketers tend to take the opposite approach. We measure what we can measure directly and reallocate budgets accordingly.
Except this standard operating procedure misses something big: the value of the non-branded search term.
Let me explain. If a consumer searches for a specific brand such as “Four Seasons Hotels,” clicks on an ad and books a reservation, we calculate the success of that ad based on a “cost per reservation,” a form of cost per acquisition (CPA). Similarly, if a consumer conducts a non-brand (or category) search such as “luxury hotels,” then clicks on an ad for “Four Seasons Hotels,” we can make that same calculation.
When comparing the value of brand and non-brand campaigns simply through the CPA lens, brand campaigns always win. When I speak with most search marketers, that part of the budget is considered “always-on.”
Non-brand campaigns are often characterized as “expensive” and “hard to make work” -- especially in mobile, when the number of visible ads on a screen shrinks to two.
Yet just like television ads, data shows non-brand search ads have huge and largely untapped potential to sway those hearts and minds of future buyers.
For digital transactions such as online hotel reservations, moving from a last-click attribution model -- in which marketers only credit the last click or interaction before a purchase is made -- to a model that assigns credit to every consumer interaction en route to a purchase can dramatically shift a marketer’s opinion on the value of non-brand advertising.
For example, a study by one travel site found that its non-brand search campaigns were worth an additional 43% after moving from a last-click attribution model. This estimate doesn’t explicitly take into account that many of the consumers involved with the study previously did not prefer a brand, but now preferred the site as opposed to a competitor.
My own team and I have found the same phenomenon to be true for “online-to-offline” purchase scenarios. In categories such as financial services and home services, we discovered that tens of thousands of consumers placed a phone call from a non-brand search campaign to a sales extension, then later made a subsequent phone call from a brand search campaign to a sales extension.
In most cases, this second call was placed more than three months later. But the point is this intent was seeded by the initial non-brand search campaign. Like a TV truck ad, these non-brand campaigns were meaningful in ways that typically go unnoticed in the current context of search advertising measurement.
In the months and years to come, marketers will need to move beyond last-click attribution models to better recognize the long-term effects of non-brand search marketing.