Sales lift reporting is often treated as the gold standard in retail media measurement.
But in reality, it comes with a number of limitations that can materially impact how results are interpreted.
Without understanding how these reports are constructed, it’s easy to overstate performance—and make the wrong investment decisions.
Here are the key areas brands should evaluate carefully.
1. Lack of De-Duplication Across Channels and Households
One of the most common issues in retail media measurement is double-counting.
Sales are often tied to individual user IDs, not households.
This creates two challenges:
- Multiple household members purchasing the same product may each be counted separately
- A single shopper exposed across multiple channels may be credited multiple times
For example:
A shopper sees an ad on CNN.com, Walmart.com, and Facebook, then makes one purchase.
If each channel is measured independently, each may claim full credit.
Without cross-channel and cross-household de-duplication:
Performance can appear significantly higher than reality.
2. SKU Selection Can Significantly Influence Results
Another major factor is which SKUs are included in the measurement set.
Brands typically want visibility into two views:
- Featured SKUs (“hero cuts”) — to determine whether the campaign drove purchases of advertised products
- Total SKU set (“halo cuts”) — to understand broader brand impact
However, many retailers:
- Report only total (halo) performance
- Or fail to define a SKU set prior to launch
This creates a problem.
Halo reporting captures a wider range of sales—and can inflate perceived campaign impact.
Without transparency into SKU selection, it becomes difficult to understand what actually drove results.
3. Incomplete Attribution and Use of Extrapolation
Even the most advanced retail media networks cannot fully attribute all sales.
In many cases:
- Only ~80% of transactions can be directly tracked
- Cash purchases and other gaps limit visibility
To compensate, retailers often apply extrapolation models based on observed behavior.
While this helps fill gaps, it introduces risk.
Modeled sales can inflate results and reduce confidence in reported performance.
This is particularly important with smaller retailers, where extrapolation may represent a larger share of total reported sales.
Brands should:
- Review methodologies carefully
- Validate assumptions
- Request transparency around modeled contributions
4. Variability in Attribution Methodology (View vs. Click)
Not all attribution models are created equal.
Two common approaches include:
View-through attribution
- Credits a sale if an ad was served—even without engagement
- May include in-store purchases
- Typically results in higher reported performance
Click-through attribution
- Requires user engagement
- Produces more conservative, direct measurement
Each has trade-offs.
- View-through can overstate impact—especially with low-visibility placements
- Click-through can underrepresent influence—since many purchases happen without a click
The choice of attribution model can significantly change how performance is perceived.
The Bottom Line
Sales lift reporting can provide valuable directional insight—but it should not be taken at face value.
Results are highly dependent on methodology.
Variability in:
- De-duplication
- SKU selection
- Attribution coverage
- Measurement approach
Can all materially impact reported outcomes.
To improve confidence in results, brands should:
- Prioritize transparency in methodology
- Validate assumptions—especially around extrapolation
- Align on consistent frameworks across retailers
The goal isn’t just to measure performance—it’s to understand true incremental impact.