It is Q4 and marketers are building 2027 media plans, reviewing results and negotiating budgets. Yet the conversation still turns on search, social, CTV, programmatic and retail media. Somewhere between the audience strategy and the budget discussion, the customer gets lost.
Channel reporting still matters: costs, inventory performance and buying efficiency have to be understood. But those decisions should support the audience strategy, not replace it.
Start with the customer opportunity
The article suggests a practical reset. If the business goal is profitable sales, think in three audience groups before touching a media plan:
- New customers: acquire profitable buyers; watch qualified reach, customer acquisition cost, new customer volume and expected customer value.
- Occasional customers: increase profitable purchase frequency; watch repeat purchase rate, purchase frequency and profit from additional purchases.
- Loyal customers: retain value and grow spending where possible; watch retention and customer value, and avoid paying for purchases that would happen anyway.
Each group needs a different strategy, media mix and investment level.
A 6:1 ROAS shows credit, not cause
Within each segment, the mix can span CTV for reminders, social for relevant offers and search for ready-to-shop customers. But channel-level reports can mislead. The source gives a familiar example: search reports 6:1 ROAS, social reports 4:1 and CTV delivers strong reach but limited attributable sales.
Those numbers do not prove the media mix is right for any segment, and a 6:1 search ROAS does not mean the next search dollar will return 6:1. The author puts it plainly:
It tells you what search received credit for, not necessarily what it caused.
The useful questions are whether you reached the intended customers, how much frequency and duplication you created, and whether purchases grew profitably. The sharper test is what would happen if you shifted, reduced or removed spending in one platform.
Plan the test before launch
That is hard to reconstruct later. Consistent audience definitions, campaign identifiers, exposure data, costs and outcome reporting should be agreed before launch. Where possible, holdout or geo tests can estimate incremental contribution.
Then look across segments. The same customer may be reached by multiple campaigns, and not all overlap is waste. The challenge is deciding whether combined investment creates additional value or just adds cost.
In a retailer example from the piece, search looked strongest in the channel report. But audience analysis raised a different question: how much branded search spending reached loyal customers who would buy anyway? The retailer kept the total budget flat and tested whether reducing branded search hurt incremental sales before redirecting money into occasional-customer segments with greater growth potential.
One question for every budget review
The next time someone recommends more search because it delivered 6:1 ROAS, the first question should not be how much more to spend. It should be: for which audience and which segment?
Source: MarTech
