SKU-Level Sales Analytics for FMCG Distributors: How to Spot Dead Stock Before It Kills Your Margins

Last month a distributor in Sharjah showed me his warehouse. 4,200 SKUs. And when we pulled the numbers together, 617 of them hadn't moved a single case in 90 days. Sitting there. Blocking cash. Some of them 60 days from expiry.

He had no idea. Not because he's careless — he's actually one of the sharper operators I've worked with. But his ERP dashboard showed him totals. Category totals. Brand totals. Warehouse-wide inventory value. Everything looked fine from 30,000 feet.

The problem lives at SKU level. Always has.

Why Aggregate Reporting Lies to You

Here's the thing about FMCG distribution — you can have a brand growing 12% year over year and still be sitting on dead stock inside that same brand. The hero SKUs pull the average up. The dogs hide underneath.

I've seen this in Karachi, Riyadh, Manchester, Houston. Doesn't matter the market. The pattern's the same. A distributor carries 8 variants of a shampoo. Two variants do 70% of the volume. Three do most of the rest. And three are dying slowly while the P&L looks green because the category average is fine.

That's the trap.

Aggregate reporting was built for accountants, not operators. It tells you what happened last quarter. It doesn't tell you which specific 250ml lavender-scented SKU is about to cost you AED 40,000 in write-offs because it hasn't moved in Deira for 74 days and expires in October.

SKU level sales analytics is a different beast. You're looking at each individual item as its own little business. Velocity, days of cover, outlet penetration, reorder frequency, and — the one most distributors ignore — the trend of the velocity itself. Not just "is it moving?" but "is it moving less than it was moving six weeks ago?"

That second question is where dead stock gets caught early.

The Four Signals That Predict Dead Stock

I'll skip the theory. Here's what actually works based on what we see across Zivni customers managing 500 to 15,000 SKU catalogues.

Signal 1: Velocity decay over rolling 4 weeks. Not month-on-month. Weekly rolling. If a SKU sold 42 cases in week one, 38 in week two, 29 in week three, and 22 in week four — you don't wait for the monthly report. That's a dying SKU. Even if the monthly total still looks acceptable versus last month.

Signal 2: Outlet drop-off. If a SKU was being ordered by 118 outlets last month and only 74 this month, something changed. Maybe a competitor launched. Maybe a planogram shifted. Maybe your reps stopped pushing it because their incentive changed. Either way — investigate before the inventory ages out.

Signal 3: Reorder gap widening. For every SKU, there's a natural reorder rhythm. A tea brand in a busy grocery might reorder every 9 days. When that gap stretches to 14, then 19, then 22 — that's your early warning. Most distributors catch this only when the retailer stops ordering entirely. Too late by then.

Signal 4: Shelf share collapsing at the outlet. This is where AI shelf photo analysis matters. If your rep is snapping shelf photos on every visit and your facings dropped from 6 to 2, the sales decline is already baked in. You just haven't seen it in the numbers yet.

These four signals — layered together — give you a dead stock probability score per SKU. Honestly, we didn't build this at Zivni on day one. I used to think dead stock was mostly an ordering/forecasting problem. Buy less, stock less. But after watching customers still get burned even after they cut orders, I realised it's actually a detection problem. The stock's already in the warehouse. The question is whether you can move it before it becomes worthless.

What to Actually Do When You Spot One

Spotting dead stock is only half the job. The other half is what most distributors mess up — the intervention.

Here's the playbook I share with our customers doing FMCG dead stock management properly:

First, segment the SKU. Is it slow because demand died, or slow because distribution died? Two totally different problems. If 90 outlets used to carry it and now only 40 do, that's a distribution problem — push it back into outlets with a targeted rep drive. If 118 outlets still carry it but each is ordering less, that's genuine demand softening. Different play.

Second, price the risk. A SKU with 14 months of shelf life sitting on 8 weeks of cover is not an emergency. Same SKU with 5 months of shelf life is a fire. Sort by risk, not by volume.

Third, run a scheme — but a smart one. Blanket 15% off across a dying SKU is lazy and expensive. Better: targeted bundling with a fast-moving SKU, or a bonus-case offer only in the outlets that used to carry it. Distributor sales analytics done right will tell you exactly which outlets those are, by rep, by beat.

Fourth — and this is the one people skip — feed the learning back into purchasing. If SKU X died in Q2, don't reorder the same quantity in Q3 because "the brand said so." Push back with data. Your distributor sales analytics is leverage in that conversation with the principal. (Yes, I know I'm not supposed to use that word. But you know what I mean.)

One of our customers in Muscat cut their dead stock write-offs by roughly 63% in five months just by running this loop weekly instead of quarterly. Nothing fancy. Just SKU-level visibility plus a Monday morning review with the sales manager and the warehouse manager in the same room.

That's really the whole trick. The tech matters — you can't do this on spreadsheets once you cross 500 SKUs, and Zivni or something like it earns its keep here — but the discipline of looking every week is what actually saves the margin.

So — how many SKUs are dying in your warehouse right now that you don't know about?