Most catalog audits fail because they ask the wrong question. "Which SKUs are profitable?" is not the question. The question is "where would more resources produce more profit?" — and the answer is almost always not the SKU you are currently spending the most on.
Across 30+ client brand rationalization audits we have run, one pattern holds with eerie consistency: the catalog is not balanced. 20% of SKUs drive most of the revenue and almost all of the contribution margin. 30-40% of SKUs lose money once shared costs are allocated properly. And another 20-30% of SKUs are quietly mediocre — not losing much, not winning much, just consuming team attention and ad budget that should be going somewhere else. By the end of this article you will know exactly how to calculate contribution margin per SKU, how the 5-bucket classification works, the kill criteria, how AWD changes the math, the quarterly audit sequence, the 5 mistakes brands make, and how to execute a kill list cleanly without disrupting your operations. The math is brutal but the lift is real — 15-25% revenue improvement within 90 days of a clean rationalization.
What SKU rationalization actually means
SKU rationalization is the systematic audit of every product in your Amazon catalog to identify the right resource allocation per SKU. The output is a decision per SKU: scale it, maintain it, optimize it, cut costs on it, or kill it. The goal is not catalog reduction for its own sake — the goal is profit optimization through better resource allocation.
The cleanest mental model: think of your catalog as a portfolio. Each SKU is a position. Some positions deserve more capital (ad spend, listing investment, team attention). Some positions deserve less. Some positions should be closed entirely. Treating every SKU equally is the equivalent of holding every stock at equal weight regardless of performance — mathematically suboptimal.
The opportunity cost framing
Every dollar of ad spend on a Kill-bucket SKU is a dollar not spent on a Scale-bucket SKU. Every hour of team attention on a marginal SKU is an hour not spent on a winner. Every cubic foot of FBA storage on a slow-mover is space that could hold faster-turning inventory. Rationalization is fundamentally about opportunity cost recognition.
It is not about catalog reduction
Some brands resist rationalization because they hear "catalog reduction" and assume it means losing breadth. That is the wrong framing. Many rationalizations result in more SKUs over time because the analysis identifies gaps where adjacent product opportunities exist alongside the cleanup of obvious losers.
It is not a one-time exercise
Markets shift. Categories evolve. Consumer preferences change. A SKU classified as Scale in Q1 might become Optimize in Q3 if a strong competitor enters the query. Quarterly rationalization keeps the portfolio dynamically aligned with current reality.
The Pareto math on Amazon catalogs
The 80/20 rule is the single most consistent pattern in Amazon catalog analytics. Across every audit we have run, the math lands within a few percentage points of the textbook distribution: 20% of SKUs drive 75-85% of revenue, and the same 20% capture 80-95% of contribution margin.
of revenue. Often 85-95% of contribution margin.
20% of SKU count → 80% of value
of revenue. Often 5-15% of contribution margin.
80% of SKU count → only 20% of value
Why the distribution holds
Amazon's algorithmic surface area rewards winners. The Best Sellers Rank reinforcement loop, the Amazon's Choice badge mechanism, the conversion rate compounding effects all push successful SKUs higher. Marginal SKUs do not get the algorithmic boost. The result is bimodal distribution — clear winners and a long tail of mostly-stagnant items.
What this means for resource allocation
If your Scale bucket has 10 SKUs and your full catalog has 50 SKUs, those 10 deserve approximately 80% of your ad budget, listing investment, photography refresh, A+ Content depth, and team attention. Most brands allocate closer to 40-50% to their winners because they spread resources evenly out of habit or fairness. That gap is where rationalization unlocks lift.
Reallocating 30% of ad budget from Kill-bucket SKUs to Scale-bucket SKUs typically produces 15-25% total revenue lift within 90 days — not from new sales, but because the Scale SKUs already convert at higher rates, hold Choice badges, and have stronger organic ranking. More budget on already-winning SKUs compounds harder than the same budget spread thin across the full catalog.
The 5-bucket SKU classification
The 5-bucket framework is the operational backbone of rationalization. Run every SKU through the matrix and assign it to one of these buckets. The bucket determines the action.
High velocity + high contribution margin. The catalog winners that drive most of the profit.
Moderate velocity + healthy contribution margin. Steady performers that pay their way.
Positive CM but underperforming velocity. Listing optimization candidates — 90 days to turn around.
Thin or negative CM, moderate velocity. Save through cost discipline (sourcing, AWD, ad reduction).
Negative CM + low velocity + 12+ months on listing. Liquidate, discontinue, or convert to special-order.
How to sort SKUs into buckets
Run each SKU through two questions. Question 1: Is this SKU's contribution margin positive over trailing 90 days? Question 2: Is this SKU's velocity in the top 50% of the catalog or bottom 50%? The 4 combinations map directly to Scale (yes/top), Maintain (yes/middle), Optimize (yes/bottom), and Cut Costs / Kill (no, depending on path to recovery).
The Cut Costs vs Kill distinction
The hardest call is between Cut Costs and Kill. A SKU with negative contribution margin and decent velocity might be Cut Costs (save through sourcing renegotiation, FBA fee optimization, AWD migration) or might be Kill (no path to profitability without category shift). The test: is there a credible operational path to positive contribution margin within 90 days? If yes, Cut Costs. If no, Kill.
How to calculate contribution margin per SKU
Contribution margin per SKU is the foundational metric. Get this wrong and the entire rationalization fails. The formula is straightforward, but the execution requires discipline in cost allocation.
This SKU contributes $1,400/month to overhead and profit. Scale bucket candidate, assuming velocity is also strong. Now imagine the same revenue line with $2,400 ad spend instead of $1,680 — the contribution margin drops to $680, still positive but materially weaker. Now imagine $4,200 ad spend — the SKU goes negative.
The 6 cost categories to track
- COGS: production + inbound freight + duty + landed cost per unit
- Amazon referral fee: 8-15% depending on category
- FBA fulfillment fee: per-unit based on size and weight tier
- FBA storage fee: monthly cubic-foot rate (varies seasonally for FBA, flat for AWD)
- Returns: Amazon's customer returns processing + value of damaged returned units
- Allocated ad spend: direct ad spend (Sponsored Products on this SKU) plus brand-level spend allocated proportionally
Where most calculations break down
Two common allocation errors: (1) ignoring brand-level ad spend (DSP, Sponsored Brands, brand campaigns) and crediting all sales to organic, (2) ignoring returns at the SKU level and using a brand-wide return rate instead. Both errors flatter your worst-performing SKUs and disguise their true contribution margin. Reading your full P&L correctly requires SKU-level allocation of every shared cost.
The kill criteria checklist
Killing a SKU is irreversible (or at least expensive to reverse). The kill criteria should be strict to avoid premature kills, but firm enough to actually cut the losers. A SKU is a Kill candidate when it clears ALL of the following criteria:
- Negative contribution margin in trailing 90 days — not just a bad month, a consistent 90-day trend
- Under 30 units/month velocity — below the threshold where ad investment can drive lift
- 12+ months on listing — not a new launch still in optimization phase
- No path to profitability through cost reduction — AWD migration, sourcing renegotiation, FBA fee optimization all explored
- No path to profitability through velocity lift — listing optimized, A+ Content built, image testing run, ad campaigns optimized
- No customer would meaningfully miss this SKU — not a tentpole product for a customer segment, not a brand signature item
The 6-of-6 rule
All 6 criteria must clear before killing. A SKU that misses one criterion goes to Cut Costs or Optimize bucket instead. Premature kills happen when only 3-4 criteria are met and the team rushes to clean up the catalog. The 6-of-6 standard prevents that mistake.
Before killing a SKU, ask: would a customer be materially disappointed if this disappeared? If the SKU is one of your top 3 reviewed products in a sub-category — even at low volume — killing it might cost brand equity that outweighs the unit economics. The exception is rare but real. When in doubt, move the SKU to Cut Costs bucket for one more quarter and reassess.
The Ecom Profit Box
11 PDF guides including the Why Traffic Isn't Your Problem — Conversion Is playbook — the conversion-side levers that determine whether a SKU lives in Optimize or Kill bucket.
Grab it free →SKU Rationalization Sprint
14-day catalog audit. Per-SKU contribution margin, 5-bucket classification, kill list execution, budget reallocation framework, 90-day reassessment.
Book a strategy call →How AWD changes the rationalization math
Amazon Warehousing & Distribution shifted the borderline math for hundreds of SKUs that previously looked like Kill candidates. The 30-60% storage cost savings from moving buffer inventory to AWD often turn a marginal SKU into a viable one.
Rerun the math with AWD-adjusted storage
Before killing any oversized or slow-moving SKU, model the contribution margin assuming buffer inventory is in AWD ($0.06/cuft/month) instead of FBA ($0.87-$2.40/cuft/month). For oversized SKUs sitting 6+ months in FBA storage, the math swing can be $500-$2,000 per SKU per year — enough to push a $200/month bleeder into a $300/month contributor.
The Cut Costs path expanded
AWD effectively expands the Cut Costs bucket. SKUs that would have been Kill in 2023 are now realistic Cut Costs candidates in 2026 because the storage savings are real and the AWD-to-FBA auto-replenishment removes the operational friction.
When AWD does not save the SKU
AWD does not solve velocity problems. If a SKU sells 5 units per month, moving it to AWD reduces its storage cost but does not change the fundamental fact that the SKU is not earning relevance through sales. AWD saves SKUs that are size-cost-burdened, not velocity-burdened.
The quarterly audit playbook (14 days)
Here is the 14-day sequence we run for client brands. Done once per quarter, this becomes a predictable rhythm that catches problems before they compound.
Days 1-3: Data pull
Export Sales and Traffic Report for trailing 90 days. Export FBA Storage Fees report. Export Returns Report. Pull contribution margin per SKU from Helium 10 Profits, SellerInvoice, A2X, or your accounting system. Consolidate into a single SKU-level dataset in Google Sheets or similar.
Days 4-6: Contribution margin per SKU
For each SKU compute the full cost stack: Net Revenue - (COGS + Amazon fees + FBA fees + storage + returns + allocated ad spend). Sort by contribution margin descending. Flag every SKU with negative CM in trailing 90 days.
Days 7-9: 5-bucket sort
Run each SKU through the velocity x CM matrix. Sort into Scale, Maintain, Optimize, Cut Costs, and Kill. Document the reasoning per SKU. Most catalogs land at 10-20% Scale, 20-30% Maintain, 30-40% Optimize/Cut Costs, 15-25% Kill candidates.
Days 10-12: Execute the kill list
For SKUs in Kill bucket: initiate FBA removal orders, mark inventory for liquidation, deactivate active ad campaigns, discontinue or convert to special-order. Use AWD as a parking spot for B-stock you want to liquidate slowly.
Days 13-14: Budget reallocation
Reallocate ad budget, A+ Content investment, and team attention from killed/optimized SKUs into Scale bucket. The compounding effect typically produces 15-25% revenue lift within 90 days of rationalization.
The 5 rationalization mistakes
Mistake 1: Treating revenue as the metric
Top-revenue SKUs are not necessarily top-margin SKUs. A high-revenue SKU with heavy ad spend, expensive FBA fees, and elevated return rate may be a Kill candidate even though it looks like a winner on the revenue dashboard. Sort by contribution margin, not revenue.
Mistake 2: Not allocating shared costs
Brand-level ad spend, DSP, and overhead all need to be allocated to SKUs proportionally. Brands that skip this allocation see inflated contribution margins on every SKU and miss the real losers. Allocate every shared cost to a SKU using revenue share as the default allocation key.
Mistake 3: Killing too fast
The 6-of-6 kill criteria exists to prevent premature kills. Most rushed kill decisions are reversed within a year as the team realizes they killed a SKU that had a path to recovery. Strict criteria protect against this.
Mistake 4: Killing too slow
The opposite mistake: keeping known losers because of emotional attachment, hope for a turnaround that never comes, or fear of catalog reduction. Set a hard rule: any SKU classified as Kill for two consecutive quarters must be executed in the third quarter at the latest.
Mistake 5: Not reallocating the freed budget
Killing SKUs without reallocating the budget produces cost savings but not revenue lift. The compounding effect comes from pushing the freed ad budget, listing investment, and team attention into the Scale bucket. Skipping this step turns rationalization into pure cost cutting, which loses 80% of the upside.
How to execute the kill list cleanly
Killing SKUs operationally is more complex than it looks. Here is the order of operations to avoid stranded inventory, broken parent-child variations, and review fragmentation.
Step 1: Pause active ad campaigns
Pause all Sponsored Products, Sponsored Brands, and DSP campaigns targeting the SKU. Pause first, kill later — this stops the bleeding immediately without irreversible action.
Step 2: Initiate FBA removal orders
For SKUs with remaining FBA inventory, choose between: (1) liquidation through Amazon's removal service at $1.00-$1.40 per unit, (2) return-to-warehouse with reverse logistics costs to recover units, (3) destruction at $0.30-$1.00 per unit. Choose based on the unit value and your ability to resell elsewhere.
Step 3: Handle parent-child variations carefully
If the Kill SKU is a child variation of a parent listing, removing it can disrupt the parent's review aggregation and ranking. Either leave the variation as out-of-stock indefinitely (no removal needed), or consolidate variations carefully before delisting.
Step 4: Discontinue or convert to special-order
For SKUs you want to remove from active selling: change the listing status to discontinued. For SKUs you want to keep available without inventory commitment: convert to special-order with extended lead times. Most brands choose discontinued for true Kill candidates.
Step 5: Update brand catalog documentation
Update internal documentation: master SKU list, supplier sheets, photography asset library. Future team members should not be confused about whether a discontinued SKU is "still active." Clean records prevent future re-launches of dead SKUs.
How Evolve Media runs rationalization sprints
SKU rationalization is one of the highest-ROI exercises we run for client brands. The math reveals itself quickly and the lift compounds for months.
14-day rationalization sprint
Full data extraction, contribution margin per SKU calculation, 5-bucket classification, kill list documentation, AWD-adjusted modeling for borderline SKUs, execution plan, and 90-day reassessment cadence. Output is a brand that knows exactly which SKUs deserve more resources and which deserve less.
Quarterly rationalization cadence
Once we have run the initial sprint, ongoing quarterly audits become lightweight (3-5 days) because the data infrastructure is already in place. The discipline becomes part of the operating rhythm.
Integration with broader Amazon strategy
Rationalization integrates with Choice badge optimization, Brand Story module work, Attribution tracking, and AWD inventory strategy. The same SKU-level data feeds all of these. Once the dataset exists, the leverage compounds across every other Amazon optimization lever.
The 7 Things to Remember About SKU Rationalization in 2026
- 20% of SKUs drive 80% of revenue, and often 85-95% of contribution margin — the Pareto distribution holds with eerie consistency on Amazon catalogs
- 30-40% of most $1M-$10M brand catalogs has negative contribution margin once shared costs (ad spend, brand overhead, returns) are allocated correctly
- 5-bucket framework: Scale (more budget), Maintain (hold steady), Optimize (test and refresh), Cut Costs (reduce fees), Kill (sunset). Sort every SKU into one bucket per quarter
- Kill criteria require ALL 6: negative CM 90 days + under 30 units/month velocity + 12+ months on listing + no cost path + no velocity path + no customer would notice
- AWD changed the rationalization math: SKUs that were Kill candidates due to storage costs are often viable Cut Costs candidates once buffer inventory moves to AWD at $0.06/cuft/month
- The compounding lift comes from reallocating budget, not from cost cutting. Rationalization without budget reallocation loses 80% of the upside — redirect the freed resources into the Scale bucket
- Quarterly cadence: 14 days per sprint initially, 3-5 days per quarter once the data infrastructure is in place. Catches problems before they compound damage