When operating margin drops week-over-week, the instinct is to look at everything at once. You pull the revenue number, scan the ad spend, glance at inventory. Nothing obvious. Then someone says “sessions are down” and someone else says “conversion rate looks soft” and the meeting runs 40 minutes without a clear answer.
There is a better structure. After 143 weekly management meetings reviewing Amazon-channel performance, one thing became clear: there are exactly four factors that explain nearly every operating margin drop. They are, in order: sessions, conversion rate, ACOS, and out-of-stock status.
You work through them sequentially. The order matters because each factor is upstream of the next.
Sessions
Sessions are the first place to look. If fewer people are seeing your listings, everything downstream suffers — conversion, ad revenue, organic rank. Session drops can come from several places: a rank loss from a previous OOS event (rank damage shows up in sessions one to three weeks after the stockout itself), a price increase that made the listing less competitive in search, or an Amazon-side traffic shift that you didn’t cause and can’t directly fix.
The important distinction is whether the session drop is acute or gradual. A sudden 20% drop in a single week points to a ranking event or a suppression issue. A gradual 5% decline week-over-week usually means BSR is slowly eroding — often from inconsistent ad spend or an OOS event a month ago that the algorithm is still penalizing.
Conversion rate
If sessions are stable but revenue is down, conversion rate is next. This one is more treacherous because conversion problems have multiple causes that look identical in the data. The usual suspects: the listing has a content issue (flagged ASIN, weak images, a title change that removed core keywords), a competitor dropped price and made your product less attractive in comparison, or a packaging damage problem caused by SIPP auto-enrollment that is generating negative reviews before you have noticed any returns spike.
Conversion rate drops that originate from listing settings rather than listing quality are particularly hard to catch. The listing looks fine. The images are fine. The copy is fine. But Amazon is shipping the product in its own packaging without poly-bag protection, it’s arriving damaged, and customers are leaving one-star reviews about crushed boxes. That is not a customer complaint problem. It is a settings problem. It shows up first in conversion data.
ACOS
If sessions and conversion rate are both healthy, the margin problem is in ad efficiency. ACOS trending upward while revenue holds steady means you are paying more to drive the same sales — usually because bids are too high relative to conversion rate, keyword targeting has drifted into irrelevant search terms, or rank mode campaigns are burning budget on placements that are not converting at the expected rate.
The key thing to understand about ACOS in the root cause tree is that it is not the same problem as ad spend volume. A team that cuts ad spend to improve margin will often improve ACOS while simultaneously destroying sessions and organic rank. That is the wrong intervention. The question is not “how much are we spending” but “how efficiently are we spending it.” Those two questions lead to very different corrective actions.
Out-of-stock status
This is last in the diagnostic order but is often the root cause of the other three. When a bestseller goes OOS, BSR drops. When BSR drops, sessions fall. When sessions fall, conversion rate data gets noisy. When you have also cut ad spend — a common reaction to low sales during an OOS period — ACOS looks artificially clean because there is nothing to report. The whole picture distorts.
Here is the real example. In a June 2024 management meeting for Wargames Delivered, the team analyzed a week where operating profit had dropped significantly. After working through the four-factor tree, the attribution came out this way: 40% of the decline was from out-of-stock SKUs that had lost rank and were generating no revenue. The remaining 60% came from ad spend that had been cut in the preceding period, which compounded the rank loss that the OOS events had already started.
Both causes were found in a single review session. Without the structured approach, the team would have likely focused on one or the other — probably the ad spend, because that is the more visible control lever. Addressing only one would have left half the problem untouched.
Why the order matters
The tree runs in the direction of causality, not just correlation. Sessions are upstream of everything. Conversion rate affects ACOS because Amazon’s algorithm factors conversion history into placement decisions. ACOS affects margin directly. OOS events create the conditions that make all three worse simultaneously.
Working through them in order stops teams from treating a downstream symptom as if it were the root cause. It also stops the most common failure mode: cutting ad spend when the real problem is inventory. That response improves one line on the dashboard while actively making the underlying situation worse.
Operating margin drops are rarely mysterious. They are usually one of these four things, in one of these four places, or a combination of two of them interacting. Start at sessions. You will usually find your answer before you reach the bottom.
