After managing hundreds of brands on Amazon, the failure pattern we see most often is not a crisis. It is a slow leak. TACoS goes from 9% to 11% to 13% to 16% over 14 months, never moving fast enough in any single month to trigger an alarm, and by the time anybody acts, the brand has burned through 14 to 22 points of margin without a single account meeting flagging it.
This is the case study of a $6.2M/year pet supplement brand we inherited in February 2026 with TACoS sitting at 18.4%, down from a healthy 9.1% in December 2024. We walk through exactly how TACoS drift happened, the 4 root causes our 30-day diagnostic surfaced, and the 90-day reset that pulled TACoS back to 11.4% without revenue loss.
If you are running $100K+/mo on Amazon and your TACoS has crept up more than 3 points over the last 12 months, this is for you.
The brand profile and the baseline drift
The brand: a 14-SKU pet supplement house, joint, gut, and skin formulas. Hero SKU running $190K/mo. Total monthly revenue at intake: $516K. Previous agency relationship: 26 months.
Here is the actual TACoS curve we reconstructed from billing and ad reports:
- Dec 2024: 9.1%
- Mar 2025: 10.4%
- Jun 2025: 12.1%
- Sep 2025: 14.3%
- Dec 2025: 16.8%
- Feb 2026 (intake): 18.4%
Revenue was flat in absolute terms across that window. So total ad spend nearly doubled to hold the same top line. That is the definition of a margin leak — invisible if you only look at revenue, brutal if you look at gross profit.
The previous agency’s monthly reports never flagged this because they led with revenue and ACoS on attributed sales. TACoS was on page 4. The board never read past page 1.
Why TACoS drift hides
Most TACoS drift hides for three structural reasons.
Reason 1: revenue stays flat or grows, masking the issue. If revenue is flat, nobody panics. If revenue is up 4% YoY, everybody assumes things are working. The fact that ad spend is up 60% to produce that 4% is not in the room.
Reason 2: agency incentives reward spend, not efficiency. Many agencies are paid as a percentage of ad spend. We covered this dynamic in detail in our percent-of-ad-spend pricing critique. The same logic applies here — the system slowly rewards bloat.
Reason 3: monthly reports use the wrong denominator. Attributed-sales ACoS holds steady while non-attributed organic revenue erodes underneath. TACoS captures both. Most reports do not.
The 4 root causes (from our 30-day diagnostic)
We ran a full account diagnostic across SP, SB, SD, DSP, and the organic SQP picture. Four root causes accounted for roughly 92% of the drift.
Root cause 1: branded defense spend ballooned (responsible for ~3.4 points of drift)
The previous agency had been running branded defense campaigns with no aggressive bid cap, on the theory that “defending branded queries is always profitable.” It is — until the brand’s organic share of voice on its own branded SERP is already 80%+ and the marginal click is cannibalizing.
We pulled the branded-defense campaigns. Branded sessions dropped 4.1%. Branded revenue dropped 0.7%. Net annualized savings: $94K. We reinvested 35% of that into non-branded prospecting, kept the rest on the P&L.
Root cause 2: long-tail keyword bloat in non-branded (responsible for ~5.8 points of drift)
Same pattern we showed in our 9M pet brand keyword pruning case study. 340 non-branded keywords running, 78 producing 91% of attributed revenue. The other 262 kept getting renewed because they showed marginal returns at the keyword level — but never net of cannibalization with the top 78.
We pruned to 84 keywords. Non-branded sessions fell 11%, non-branded revenue rose 4% (yes, rose), ACoS on non-branded dropped from 39% to 24% inside 60 days.
Root cause 3: auto campaigns no longer producing quality discovery (responsible for ~2.6 points of drift)
Auto campaigns had been running on autopilot for 22 months. We pulled the 90-day search-term report and found that 74% of auto-campaign spend was going to queries that already existed as manual exacts elsewhere in the account. The auto was effectively bidding against the brand’s own manual.
This pattern accelerated after May 2025 as Rufus (now Alexa for Shopping since the May 13, 2026 transition) started intercepting top-of-funnel discovery queries. Auto-campaign ACoS rose meaningfully across the category between mid-2025 and early 2026.
We rebuilt the auto strategy from scratch: 1 hero-SKU auto at down-only, low budget, harvesting only. All discovery work moved to a manual broad campaign with weekly negative-keyword reviews.
Root cause 4: SD remarketing budget had no ceiling (responsible for ~1.9 points of drift)
Sponsored Display remarketing was running with no daily cap. Every month the previous agency increased remarketing budget by 8-12% “because remarketing always converts.” But the remarketing pool was finite — they were re-serving the same fatigued audiences at rising vCPMs.
We capped SD remarketing at 12% of total ad budget, hard. ACoS on SD dropped from 28% to 14% inside 45 days.
The 90-day reset playbook
For brands sitting at our intake state (TACoS 16-20%, revenue flat), our default 90-day reset is sequenced:
Days 1-15 — Diagnostic and quick kills
- Pull 90-day search-term reports across all campaign types
- Compute true incremental contribution of branded defense campaigns
- Identify long-tail keywords below 0.5% revenue contribution and confirm they’re not strategic
- Audit auto-campaign overlap with manual exacts
- Kill obvious bloat — usually 15-25% of total ad spend on day 15
Days 16-45 — Restructure and rebuild
- Rebuild non-branded campaign structure with pruned keyword list
- Reset auto campaigns to harvesting-only role
- Cap SD remarketing at fixed percentage of total budget
- Move bid strategy on mature non-branded to up-and-down (after 21-day learning period)
- Begin weekly negative-keyword reviews
Days 46-90 — Defend and reinvest
- Reinvest 35-40% of savings into newly-identified prospecting opportunities
- Roll TACoS reporting to weekly cadence with monthly board summary
- Build the 8-metric monthly report we cover in our agency monthly reporting deep-dive
The result
By day 90 (early May 2026), TACoS landed at 11.4%, down from 18.4%. Revenue closed at $522K for the month — slightly up from intake. Gross ad spend dropped from $95K to $59.5K. Annualized margin recovery: roughly $426K.
The brand didn’t lose impressions in queries that mattered. It stopped paying for impressions in queries that didn’t.
TACoS benchmarks worth knowing
From the 80+ brands we audit each year, our category-blended healthy TACoS ranges (for brands in steady state, not launch):
- Consumables (supplements, food, beverage): 9-13%
- Apparel: 12-17%
- Beauty: 11-16%
- Home and kitchen: 8-13%
- Electronics accessories: 7-12%
- Pet: 10-14%
If you are 3+ points above the high end of your category band and revenue hasn’t grown meaningfully in 12 months, you have TACoS drift. Treat it as urgent.
FAQ
Is TACoS more important than ACoS?
Yes, in nearly all cases. ACoS only measures the attributed slice. TACoS measures whether your ads are actually growing your business.
How often should we review TACoS?
Weekly at the account level, monthly at the SKU level, quarterly with hero-SKU deep-dives.
What if TACoS is dropping but so is revenue?
That’s a different problem. You’re starving the account. The reset above assumes flat-to-growing revenue with rising spend.
Does TACoS drift happen in growing brands too?
Yes — actually more often. Growth covers the leak for longer. We see the worst drift in brands that grew steadily for 18-24 months and never reset the ad infrastructure they built at $1M revenue.
—
If you’re looking for a team that manages every lever — creative, advertising, and operations — Velocity Sellers works with brands doing $100K+/month on Amazon. Contact us for a free account audit.
