Memorandum
- From
- Caroline Mercer via MarTech
- Date
- Filed
- Business·4 min to read
- Re
Cheap Branded Search Ads Often Take Customers Who Would Have Come Anyway
ReCheap Branded Search Ads Often Take Customers Who Would Have Come Anyway
A low customer acquisition cost in advertising often means a business is merely capturing demand that already existed. Incrementality tests help separate genuine new conversions from baseline traffic.
A low customer acquisition cost, or CAC, and a high return on ad spend, ROAS, often look like signs of effective marketing. However, both metrics credit conversions to channels rather than prove that a specific ad created new customers. A campaign with a cheap figure may only be intercepting demand that already existed, while a more expensive campaign sometimes delivers real additional growth.
A typical example is branded search with a CAC of about $15. Such customers mostly already know the company and buy from it, so ads on its own name often just take a visit they would have made anyway. By contrast, paid social advertising with a CAC of about $60 may reach people who otherwise would not have come. Channels that work closer to the purchase, such as branded search, retargeting and affiliate programs, usually look most effective in reports, although their true contribution may be smaller.
The author describes paid search testing results that illustrate this problem in practice. Cutting branded search spending by about 90 percent had almost no effect on the number of customers acquired. Paid advertising had also partly replaced organic search traffic, meaning people who would have visited the site for free anyway. After paid spending was reduced, organic results recovered most of that demand. An incrementality study for non-branded search showed a 55 percent increase in conversions, indicating the real contribution of this channel.
Standard attribution assigns a conversion to the advertising that preceded it. Incrementality, by contrast, measures which conversions the advertising actually caused. Without such an approach, platforms risk overstating the role of advertising for demand that would have brought the customer anyway. At the same time, incrementality tests require careful design and a sufficient number of conversions for the results to be reliable.
To set up a test, according to the material, one should start with a specific question about new customers. The methods named include platform lift studies, geo holdouts and controlled experiments. The key metric is incremental CAC, that is, spending divided by the number of additional customers. It should be compared with customer lifetime value, and thresholds should be defined in advance for when the budget is scaled up, left unchanged or cut.
The material also discusses the role of artificial intelligence in such tests. Large language models can select similar regions for geo holdouts, estimate the required duration and sample size, and monitor tests for data contamination or tracking errors. Tools such as GeoLift from Meta and Haus support similar workflows. AI can also prepare report drafts, but people should verify the results.
For small teams, this approach makes it possible to test advertising without a separate analytics department. The main takeaway is that spending on existing demand crowds out the ability to acquire new customers. The goal of such studies is to find investments that genuinely add customers and justify increasing the budget.
4
