The Budget Call: Auditing a Quarter of Retail Media Spend
Objective: Given a quarter of sponsored ads performance data across two retail media networks, apply TACoS over ROAS, separate branded from non-branded performance, and decide where next quarter's budget should actually go.
You're the retail media manager at Nykaa, evaluating a quarter of Sponsored Products performance across Amazon and a grocery-style RMN before proposing next quarter's budget split to leadership.
Recompute the real picture using TACoS instead of the headline ROAS number, split branded from non-branded performance, and recommend which platform gets the incremental budget.
Before you start
What you'll need
Free path (everything below is enough to finish)
Native reporting for any active advertiser, no extra cost
Free, flexible enough for custom formula columns not in native dashboards
Paid upgrades (optional, faster/deeper)
The free path (Amazon Ads native reports plus Google Sheets formulas) is a complete quarterly workflow. A paid dashboarding tool like Triple Whale only saves time once you're doing this recalculation across more than two or three platforms every month.
Saves manual recalculation once running budget across 3+ platforms regularly
No access? Build the same TACoS formula manually in Google Sheets each quarter
The process
3 steps
Step 01 of 03
The lesson explains TACoS divides total ad spend by total revenue (ad-driven plus organic), capturing the organic sales lift that ROAS alone misses.
Leadership's slide says 'Amazon Sponsored Products: 7.8x ROAS, best channel this quarter.' The same product line's organic Amazon revenue also grew 40% this quarter. Is 7.8x ROAS the number to lead the recommendation with?
Procedure
- Pull total ad spend and ad-attributed revenue per platform from the quarterly export
- Pull total revenue (ad-attributed plus organic) per platform for the same period
- Calculate TACoS = total ad spend / total revenue for each platform
- Compare TACoS trend across the quarter's three months, not just the single headline ROAS figure
Platform Ad spend Ad revenue ROAS Total revenue TACoS Amazon $42,000 $327,600 7.8x $890,000 4.7% Grocery-style RMN $18,500 $96,200 5.2x $210,000 8.8%
Healthy
TACoS on Amazon sits at 4.7%, well inside the 5-15% healthy band, meaning ad spend is compounding into organic rank rather than being the only thing carrying sales.
Unhealthy
Leading the budget recommendation with the 7.8x ROAS headline alone, without the TACoS context, makes it look like Amazon is simply 'better,' when the grocery-style RMN's 8.8% TACoS is also healthy and reaching a different, less Amazon-saturated shopper.
What this means
A high ROAS with a healthy TACoS means the platform is compounding well. Don't let one number override the other; both are needed for a real budget call.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Leadership slide leads with ROAS only, no TACoS context | Add the TACoS column to the recommendation slide before it goes to leadership | 30 min |
Step 02 of 03
The lesson's Practical Playbook, Step 2 says branded keyword ROAS is misleadingly high because those shoppers were already looking for the brand, and incrementality should be measured on non-branded keywords.
Within Amazon's 7.8x ROAS, the search-term report shows 61% of ad-attributed revenue came from searches containing 'Nykaa.' What does the incremental picture look like once branded terms are set aside?
Procedure
- Export the Search Term Report for the quarter
- Tag each search term as branded (contains 'nykaa') or non-branded
- Recalculate ROAS separately for the branded and non-branded segments
- Flag the non-branded ROAS as the real measure of incremental customer acquisition
Segment Ad-attributed revenue Spend ROAS Branded (61%) $199,836 $15,750 12.7x Non-branded $127,764 $26,250 4.9x
Healthy
Non-branded ROAS of 4.9x is still solidly profitable, meaning the campaign is winning real new customers away from competitors, not just capturing brand-search demand that existed anyway.
Unhealthy
Reporting the blended 7.8x ROAS as 'the number' overstates how much new customer acquisition the campaign is actually driving, since branded searches were mostly going to convert regardless of the ad.
What this means
Blended ROAS answers 'did the campaign make money.' Non-branded ROAS answers the more important question: 'is the campaign winning customers we wouldn't have gotten otherwise.'
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| 61% of ad-attributed revenue is coming from branded search terms | Report branded and non-branded ROAS as two separate lines going forward, not one blended figure | 5 min |
Step 03 of 03
The lesson recommends starting on Amazon for high-search-volume categories like beauty, then adding a grocery-style RMN like Instacart for CPG-adjacent categories with genuine reach beyond Amazon shoppers.
Given Amazon's healthy but branded-heavy 4.9x non-branded ROAS and 4.7% TACoS, and the grocery-style RMN's 8.8% TACoS reaching a largely non-Amazon audience, where should the next $10,000 in incremental budget go?
Procedure
- Chart non-branded ROAS trend by platform across the quarter's three months
- Chart TACoS trend by platform across the same period
- Check whether either platform's TACoS is trending up (a warning sign) or flat/down (healthy)
- Recommend the split based on both trend direction and audience overlap, not the single highest ROAS number
Platform Non-branded ROAS trend TACoS trend Audience overlap with other platform Amazon 4.6x -> 4.9x -> 4.9x 5.1% -> 4.9% -> 4.7% High-intent search demand Grocery-style RMN 4.1x -> 4.6x -> 5.0x 9.4% -> 9.0% -> 8.8% Largely distinct shopper base
Healthy
Both platforms show flat-to-improving TACoS and rising non-branded ROAS, meaning incremental budget on either platform is currently compounding well rather than just buying more of the same demand.
Unhealthy
Putting all incremental budget on Amazon alone, since it already has the highest single ROAS number, while the grocery-style RMN's improving trend and distinct audience get starved of budget to test further.
What this means
When both platforms show healthy trends and one reaches a meaningfully different audience, the incremental dollar usually belongs on the platform with room to grow its audience overlap, not just the platform with the biggest existing number.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Both platforms trending healthy but budget concentrated entirely on the higher-ROAS platform | Recommend splitting the next $10,000 roughly 60/40 toward Amazon, with a meaningful test allocation to the grocery-style RMN to keep growing the distinct audience | half day |
Final deliverable
A budget recommendation memo with TACoS and non-branded ROAS by platform, and a specific dollar split for next quarter's incremental spend.
See a reference example
Chewy, quarterly retail media budget memo (excerpt) HEADLINE METRIC LEADERSHIP SAW Amazon Sponsored Products: 6.9x ROAS REAL PICTURE TACoS: 5.3% (healthy) Non-branded ROAS: 4.2x (the real acquisition number, branded search was 58% of attributed revenue) RECOMMENDATION Split next quarter's incremental $15,000 as 65% Amazon / 35% grocery-style RMN test, both platforms show flat-to-improving TACoS
Success criteria
You're done when you can:
- Recomputes TACoS instead of relying on the headline ROAS figure alone
- Correctly separates branded from non-branded ROAS and flags the branded inflation
- Recommends a specific budget split backed by trend data, not just the single highest number