Retail media used to be a Sponsored Products problem. Neat campaign, neat ACOS, neat little report. Very tidy, very suspicious. In 2026, budgets are spread across search results, product detail pages, Sponsored Brands, category shelves, video, display and offsite retailer audiences. The question is no longer “which campaign has the best ROAS?” The better question is: which placement deserves the next euro after margin, stock, returns and organic cannibalization are included?
This article supports the retail media analytics pillar and connects placement decisions with marketplace advertising, profit analytics, contribution margin, TACoS vs ROAS, marketplace analytics and the practical audit in how to find wasted retail media spend.
Why placement mix matters more than average ROAS
Average ROAS is a comfort blanket. Search placements can look efficient because shoppers were already close to buying. PDP conquest placements can look expensive because they interrupt a competitor at the exact point of decision. Video may look weak in a seven-day window but support category discovery over six weeks. If all three are judged by the same ROAS target, the budget will drift toward what is easiest to attribute, not what is most profitable.
| Placement | Commercial job | Metric that matters | Guardrail |
|---|---|---|---|
| Search results | Harvest high-intent demand | TACoS + contribution margin | Watch organic cannibalization |
| Product detail page | Conquest or cross-sell | Margin after CPC and returns | Cap when price gap is weak |
| Sponsored Brands | Build category and brand demand | Cohort margin and repeat rate | Use longer windows |
| Display / video | Retarget or launch demand | Incremental sales and stock cover | Frequency and margin caps |
Start by assigning every placement a job
A placement without a job becomes a budget sponge. Label each placement as harvest, defend, build, test or fix. Harvest placements must hit margin thresholds. Defend placements must prove the position is worth defending. Build placements get patience, but not infinite patience. Test placements get a written end date. Fix placements do not get more spend until content, price, stock or return issues are solved. The labels sound simple because they are. The discipline is the spicy bit.
Connect placements to SKU economics
Placement reporting only becomes useful when it is joined to SKU economics. For every placement, pull attributed sales, ad spend, CPC, conversion rate and orders. Then add COGS, marketplace fees, fulfillment, returns, discounts, stock cover and current price position. A PDP placement with 4.5x ROAS can still be weak if it sells a low-margin SKU with a high return rate. A brand placement with 2.1x ROAS may be valuable if it increases repeat purchase on a high-margin range.
Use TACoS to catch paid-credit theatre
If search spend rises and TACoS rises with no improvement in total sales, the placement is probably collecting credit for demand the product already had. That is paid-credit theatre, and the actors are very committed. Compare placement spend against total SKU sales, organic rank and conversion rate. If paid sales increase while total sales stay flat, cap the placement and move budget to a role with clearer incrementality.
The weekly placement review
- Split spend by placement. Search, PDP, brand, category, video, display and offsite.
- Attach SKU economics. Margin, fees, returns, stock, price and Buy Box status.
- Label the job. Harvest, defend, build, test or fix.
- Apply thresholds. Contribution margin, TACoS, return rate, stock cover and content readiness.
- Move budget. Increase only where placement role and SKU economics agree.
Budget rules by placement role
Once roles are assigned, translate them into rules finance and advertising can both understand. Harvest placements can scale while contribution margin stays above the SKU threshold and stock cover remains healthy. Defend placements need a written reason: which rank, term or shelf position are we protecting, and what margin would we lose if we stopped? Build placements get a longer window, but they still need leading indicators such as new-to-brand share, store visits that convert later, or category-level TACoS improvement. Test placements should have a budget ceiling and end date from day one. Fix placements are the easiest: if content, price, reviews or availability are weak, spend waits. That sounds strict because it is. Retail media budgets behave better with boundaries.
How FiveX helps
FiveX connects retail media placement performance with SKU profitability, inventory, returns, pricing and marketplace data. Instead of arguing over campaign screenshots, teams can see which placements create contribution margin and which ones simply look charming in the ad console. Charming is nice. Profit is nicer.
FAQ
What is retail media placement analytics?
It is the analysis of ad performance by placement type, connected to SKU margin, TACoS, returns, stock and incrementality.
Why is placement ROAS not enough?
Because ROAS does not show organic cannibalization, stock risk, return cost or contribution margin.
Which placement should scale first?
Usually high-intent search or proven PDP placements, but only when SKU margin and stock cover are healthy.
How should brand and video placements be judged?
With longer windows, cohort margin, repeat purchase and category TACoS, not only same-week ROAS.
Can FiveX compare placements across marketplaces?
Yes. FiveX normalizes retail media placement data across channels and connects it to marketplace profitability.