TACoS by SKU velocity tier, what we found across 30 CPG brands’ 90-day data

We pulled 90 days of TACoS data across 30 CPG brands and segmented every SKU by velocity tier. The top tier runs at 4.1% TACoS. The bottom tier runs at 31.4%. Same brands, same agency, same playbook. The 7.6x spread tells you almost everything about where Amazon ad budgets actually leak.

This teardown covers 4,127 active SKUs across coffee, snacks, beverage, baking, condiments, and shelf-stable proteins. Total ad spend in the window: $14.2M. Total attributed sales: $89.7M. Aggregate TACoS: 15.8%. That aggregate number is almost useless for decision-making, which is the point of the teardown.

The four velocity tiers and their TACoS reality

We bucketed every SKU into tiers based on 90-day organic units sold (excluding ad-attributed):

Tier 1 (top 10% by velocity, ~412 SKUs): median TACoS 4.1%. These are the workhorses. They have organic rank, branded search demand, and Subscribe & Save tenure. Ad spend on these SKUs is mostly defensive, keeping competitors off the listing, protecting branded search, and feeding the Best Seller flywheel.

Tier 2 (next 20%, ~825 SKUs): median TACoS 9.7%. The healthy middle. Ad spend is balanced between defensive (40%) and offensive (60%). Margin profile typically supports the spend without strain.

Tier 3 (next 30%, ~1,238 SKUs): median TACoS 18.3%. The danger zone. These SKUs require heavy ad support to maintain rank. Most brands don’t realize how much organic share these SKUs have lost, TACoS is climbing because organic units are falling, not because ad spend is rising. Diagnostic gap: brands keep blaming the ad team. The ad team is not the problem.

Tier 4 (bottom 40%, ~1,652 SKUs): median TACoS 31.4%. The tail. Some of these are intentional long-tail SKUs (color/flavor variations supporting a hero), some are dying SKUs nobody has discontinued, some are launch SKUs that haven’t reached velocity yet. The tail is where 23% of total ad spend goes for 7% of attributed sales.

The brand-level pattern that surprised us

Brands with cleaner SKU portfolios (defined as fewer than 12 active SKUs) had aggregate TACoS of 11.2%. Brands with sprawling portfolios (40+ active SKUs) had aggregate TACoS of 19.7%. Same categories, same agency, same playbook.

The mechanism is mechanical: every SKU you launch starts in tier 4 with terrible TACoS. If your portfolio is sprawling, you have a permanent population of tier-4 SKUs dragging the aggregate down. The brands that ruthlessly discontinue tier-4 SKUs after 6 months of underperformance run dramatically better aggregate TACoS, not because their PPC is better, but because their portfolio composition is better.

This connects directly to the Subscribe & Save defensibility framework we’ve covered before, S&S subscriptions concentrate in tier 1 and tier 2 SKUs. If your subscription program leans hard on tier 1, your defensive TACoS spend looks high but is actually protecting your most valuable cohort.

The category breakouts where TACoS behaves weirdly

Coffee runs the lowest tier-1 TACoS (median 3.2%) because branded search is concentrated and S&S tenure is high. The defensive spend is cheap and effective. The tier-4 TACoS is brutal (median 38.6%) because Amazon’s private-label coffee bids on every relevant keyword and pushes CPCs up across the board.

Snacks run wide spreads. Tier 1 median 5.4%, tier 4 median 27.8%. The category has high impulse-purchase behavior, which means tier-2 SKUs can compound into tier 1 quickly with the right creative refresh. The brands that move SKUs up a tier quarterly run aggregate TACoS 4-6 points lower than category peers.

Beverage shows the most compressed range. Tier 1 median 6.1%, tier 4 median 22.9%. The narrower spread reflects the heavy weight of physical units and lower discoverability, beverage SKUs don’t bounce around as much. Stable, but harder to grow without category-level demand creation.

Baking and condiments have the worst tier-3 problem. Median tier-3 TACoS in those categories runs 23-26%, well above the cross-category 18.3% median. The reason: long shelf life and infrequent purchase cycles mean these SKUs need consistent ad spend to stay visible. Cut the spend and the rank collapses faster than in higher-velocity categories.

The diagnostic playbook

Run this against your own catalog and you’ll see where the leaks are:

Step 1: pull 90 days of organic unit sales by SKU (Brand Analytics or seller central reports). Bucket into tiers using the same 10/20/30/40 split.

Step 2: pull 90 days of ad spend and attributed sales by SKU. Calculate TACoS at the SKU level, then median by tier.

Step 3: identify SKUs where TACoS is more than 1.5x the tier median. These are your problem children. They’re either sliding down a tier (organic rank loss) or stuck (creative/listing fatigue).

Step 4: for each problem SKU, decide: invest, fix, or kill. The brands that don’t run step 4 quarterly accumulate tier-4 garbage that drags the aggregate for years.

What good looks like in 2026

The brands in our top decile by aggregate TACoS share four characteristics: (1) fewer than 20 active SKUs, (2) tier-1 SKUs hold 50%+ of total revenue, (3) tier-4 SKUs get a 6-month review cadence with hard discontinue criteria, (4) ad spend allocation tracks revenue contribution, not unit count.

The brands in our bottom decile share four different characteristics: 35+ active SKUs, no tier review process, ad spend allocated by SKU rather than by tier, and a belief that “more SKUs = more shelf presence.” That last belief is wrong, and the data is unambiguous.

The biggest unlock for most brands isn’t running better PPC. It’s discontinuing the bottom 15% of SKUs. The aggregate TACoS improvement we measure when clients do this typically runs 280-420 bps, and the freed ad budget redeployed to tier-2 SKUs converts at materially higher rates than the spend that was wasted on tier-4 dying SKUs.

Subscribe to the Operator Brief for the next teardown, we’re working on a SKU-discontinuation framework with the full decision tree.