How to Lower TACOS Without Killing Sales

Every plan to lower TACOS runs into the same apparent trade: efficiency or sales, pick one. Cut ad spend and the ratio improves this month while the volume your ads were buying starts to leak. Defend volume and the ratio refuses to move. Most operators treat the two as opposite ends of a single dial, and most TACOS projects fail because they get run that way.

The pressure to pick efficiency is real. Amazon raised FBA fees by roughly $0.08 per unit on average in January 2026, added a 3.5% fuel and inflation surcharge in April worth roughly $0.15 to $0.35 more per unit, and pushed AWD storage and transportation rates up 19% to 22%. On the auction side, Ad Badger's benchmark data put average CPC at $1.18 in 2026, with average ACoS at 32.5% in January, and more than 70% of Amazon sellers now run ads, up from roughly 40% five years ago. A margin bar that cleared two years ago may be underwater at the same efficiency target today.

TACOS became the number leadership asks about because it is the one efficiency metric you cannot improve by shuffling spend between campaigns: the revenue side of the ratio includes the organic sales your ads are supposed to be building. And the trade-off it implies between efficiency and sales is only real in some parts of the account. Some spend is holding organic rank, and cutting it hands back sales on a delay. Some spend bought nothing for 60 days straight, and cutting it costs nothing. This post is about telling those apart: how to diagnose which side of the ratio is broken, which cuts are free and which are loans against next quarter, and a six-step order of operations for bringing TACOS down without handing back the volume. For the definition and the formula, our companion post on Amazon TACoS covers the basics, and the ACoS guide covers the ad-only side.

Quick Answer

Work the account in this order:

• Diagnose before cutting. Put ACoS next to TACOS over the same 90 days. If ACoS is falling while TACOS climbs, the problem is organic, and cutting spend makes it worse.
• Cut the spend that bought nothing. Pull 60 days of search-term data and negate targets with real click volume and zero orders. This is the one cut with no risk to organic rank.
• Fix your placement multipliers. A 300% multiplier on a $1 bid is a $4 CPC that no bid report shows you.
• Protect the keywords carrying rank. Test paid pullback on 20 or 30 organically ranked terms against a control, not across the whole account.
• Shift, then hold. Move the recovered budget toward terms where organic is absent or climbing, and rerun the cycle on a schedule.

Budget one to two quarters for a TACOS improvement you can defend, not one to two weeks.

Why most TACOS reductions do not survive the next quarter

TACOS is a fraction with a spend side on top (the numerator, total ad spend) and a revenue side underneath (the denominator, total revenue including organic). The two behave nothing alike. Spend is under your direct control and responds within a day of a bid change. Revenue responds over weeks, and only indirectly, through rank, review velocity, and conversion rate.

The spend-side trap

Almost every TACOS reduction project pulls the spend side, because it is the lever that moves before the next reporting cycle. Cut 20% of ad spend and TACOS improves this month whether or not anything about the business got better. That is the spend-side trap: a TACOS drop bought purely by cutting spend is a TACOS drop you give back next quarter, when rank decays and the organic sales on the revenue side quietly shrink underneath you.

Sponsored placements dominate above-the-fold results on most high-intent searches, and paid impressions contribute the sales velocity that feeds organic rank on that term. Pull the paid impressions on a term you were partly holding with them, and the organic position carrying your revenue side softens over the following weeks. The efficiency win lands in month one. The revenue loss lands in month three, by which point nobody connects the two.

A revenue-side reduction is the one that holds

The version that holds is a revenue-side reduction: take out spend that was never buying anything, keep the spend that is buying rank, and let organic growth shrink the ratio. Same arithmetic, opposite durability. Spend removal is the opening move. The strategy is getting organic to carry a larger share of the same revenue, which is slow, unglamorous, and the only reason a TACOS number stays where you put it.

Read the direction of ACoS before you touch a bid

Before any cut, put ACoS and TACOS side by side over the same 90 day window and note which way each is pointing. The four combinations mean different things and want different first moves:

The two climbing-TACOS rows are where accounts get misdiagnosed, so they are worth a closer look.

TACOS up while ACoS is down: organic is not keeping pace

This is the case operators misdiagnose most often, because the ads look like they are working. Run the arithmetic on an illustrative account:

Ad efficiency improved. TACOS got worse. Cutting the ad budget here makes the ratio worse, because you would be removing the only volume that is growing. The problem sits below the ad account: organic rank slipped, a competitor took a position, reviews stalled, the listing lost relevance, or a hero SKU went out of stock long enough to lose its place. Fix that, and TACOS corrects without touching a bid.

TACOS up while ACoS is also up: the auction or the conversion path

When both climb together, ad money is buying less than it did. Either CPCs in your categories rose, your match types are pulling looser traffic than they were, or conversion rate on the destination page fell. This is the case where spend removal is the right first move, because there is waste to find.

Start with the spend that bought nothing

Zero-order spend is the cleanest cut available, and it is usually larger than anyone expects. In one account we reviewed, roughly 44% of 60 days of ad spend produced zero sales across 13,000 targets. Nearly half the budget was buying impressions and clicks that never converted once in two months.

A target with zero orders across a 60-day window at meaningful click volume is not seeding organic anything, so there is no rank contribution to protect. It is the one place where cutting spend carries no revenue-side risk, and it is where a TACOS project should spend its first two weeks.

The mechanics are ordinary and worth doing properly:

• Pull 60 days of search-term reports across every campaign type, Sponsored Products included.
• Sort targets by spend and filter to zero attributed orders.
• Set a click floor before negating. Ten or more clicks with zero orders is a decision; three clicks is noise.
• Add exact negatives at the level where the traffic enters, and keep a log so every negation is reversible.
• Rerun the sweep on a schedule. Waste accumulates continuously, and a one-time cleanup is back where it started within a quarter.

Two cautions. Zero orders at three clicks is noise, and negating on it will strangle discovery. And a term with zero orders but a high add-to-cart rate deserves a look at the listing before you kill it, because the auction may be fine and the page may not be.

Protect the keywords that are carrying your rank

Once the obvious waste is gone, the cuts get consequential. Whether a target is safe to cut comes down to two questions: is it converting, and is it carrying organic rank. This matrix holds even when a spend-reduction target is bearing down on you:

Most of the rows are clean calls. The third one is not.

The complication nobody wants to test

On a term where you already hold page-one organic, pulling paid sometimes costs you nothing and sometimes costs you organic position too. Both outcomes are common. It depends on how much of the term's sales velocity your paid placement was contributing, how contested the term is, and how far ahead of the next competitor you were sitting. General advice cannot tell you which case you are in, and neither can we.

So treat it as a test, not a policy. Take 20 or 30 organically ranked terms across a mix of SKUs, hold the rest as a control, pull paid on the subset, and watch organic position and total unit velocity for four to six weeks. Our post on Amazon search ranking covers what to watch on the rank side. If the subset holds, widen it. If it slips, you learned that for the price of 30 keywords instead of your whole account, which is the point of running it small.

Check your placement multipliers before you cut another bid

One self-inflicted TACOS problem is worth ruling out before any further reduction: placement multipliers running above what anyone in the account remembers setting. The instinct when costs rise is to force position with a large multiplier. A 300% multiplier on a $1 base bid produces a $4 CPC. Base-bid optimization cannot see that number, so spend climbs and efficiency degrades while every bid report you look at shows a $1 bid.

Be precise about which problem you are solving. Placement optimization is the right tool when the goal is efficiency across a placement mix. It is the wrong tool when the goal is holding a specific position regardless of short-term efficiency cost, because an efficiency-seeking system can settle into a conservative loop and surrender the position you were paying to defend. Decide which you want per campaign, then set the multiplier to match.

Two ways to move TACOS, side by side

How to reduce TACOS on Amazon, in order

• Step 1: Diagnose the direction. Put ACoS and TACOS side by side over 90 days and find your row in the direction matrix before touching a bid.
• Step 2: Cut zero-order spend. Run the 60-day search-term sweep with a click floor and negate what converted nothing.
• Step 3: Fix placement multipliers. Find anything running above what the account remembers setting and match each multiplier to a deliberate goal.
• Step 4: Sort what remains. Classify every meaningful target with the cut-or-protect matrix above.
• Step 5: Test paid pullback on a subset. 20 or 30 organically ranked terms against a control for four to six weeks, then widen or reverse.
• Step 6: Shift budget as organic strengthens. Re-evaluate targets against their current organic position on a repeating cadence, easing paid where rank holds and reinforcing where it is absent.

Step 6 is the one that makes the reduction durable, and it is also the one that fails most often.

Where a careful manual pass runs out

Steps 1 through 5 are a manual project. A good analyst can do them well in a spreadsheet, and plenty do.

Step 6 is where the manual pass runs out, and the reason is arithmetic rather than skill. Shifting spend as organic strengthens means re-evaluating every target against its current organic position, its recent trend, and its paid contribution, on a repeating cadence. The earlier example held 13,000 targets in one account. A careful weekly pass gets through the top few hundred before the window closes, and the remaining several thousand hold their previous bid by default, so the shift you designed only ever happens on the loud end of the account. The top few hundred drift out of date in the meantime, because organic position on those terms moved again while you were working the list.

That is a throughput problem, not a judgment problem, and it does not get better by trying harder.

Where ads automation fits

This is the throughput problem Trellis' ads automation exists to solve: applying the logic you would apply by hand to every target rather than only the ones you reach. Four pieces of it matter for a TACOS project:

• Bid decisions across the whole account, on a daily cadence. The several thousand quiet targets get the same re-evaluation as the loud few hundred, so the shift you designed happens everywhere instead of only at the top of the list.
• Operator-set guardrails per campaign. Floors, ceilings, and the boundary between your rules and algorithmic bidding stay yours to set, so the system works inside the strategy rather than replacing it.
• Harvesting and negation on a cadence. The zero-order sweep from Step 2 becomes a scheduled workflow instead of a calendar reminder someone owns.
• A record of what changed and why. When the TACOS number moves, the explanation is a log entry rather than a reconstruction.

The honest limit. None of this decides your target TACOS or which SKUs deserve rank investment. The guardrails and the logic per campaign are operator-set. That is control, not less work, and choosing well still takes an hour with the catalog in front of you. Better records also shorten the explanation of what happened. They do not prove the change was right.

Conclusion

The trade between efficiency and sales is real only where spend is carrying rank. Everywhere else the cut is free, and finding the line between the two is the whole project. TACOS is a ratio, and the lever nearly everyone pulls is the one that moves fastest and holds worst. Cut the zero-order spend, because there is usually a lot of it and it is buying nothing. Fix the multipliers, because a $4 CPC behind a $1 bid defeats every other change you make. Then stop cutting and start shifting: protect the terms carrying rank, test pullback on a subset rather than the account, and move budget toward terms where organic is still absent. A TACOS number built that way is one you can still report in two quarters.

If you want the formula and the definitional side, start with our post on Amazon TACoS, and run your own numbers in the ACoS and TACOS calculator. If the throughput problem in Step 6 is the one you are stuck on, book a walkthrough and we will look at your search-term data with you.