How to Reduce Out-of-Stock Issues in FMCG Retail: A Data-Driven Approach

By Sufyan · 2026-05-10 · 4 min read

A biscuit brand we onboarded last year was losing roughly 14% of weekly sales to out-of-stocks. Not stockouts at the warehouse. Stockouts on the shelf. The product was sitting in 200+ stores across Karachi, just not where shoppers could grab it.

That's the painful part of OOS in FMCG. The inventory exists. The demand exists. The connection between them just breaks somewhere between the distributor's godown and the kiryana shelf.

I've been building Zivni for FMCG companies in Pakistan, the UAE, and a few markets in Africa, and if there's one problem every single sales head brings up in the first call, it's this one. So let me share what actually works — not theory, but the patterns we've seen across hundreds of distributors.

Stop guessing at the SKU level

Most brands track OOS at the brand or category level. That's already too late. By the time your category share dips, you've lost weeks of sales.

The shift has to happen at SKU-store-day granularity. Sounds obvious. Almost nobody does it.

Here's what I mean. A 200ml shampoo SKU might be flying off shelves in DHA Lahore but rotting in inventory in Multan. If your reporting only shows "shampoo category sold 8,400 units this week," you're flying blind. You need to know that SKU 4471 went OOS in 38 of 142 outlets on the south Karachi beat — and it happened on Tuesday, not Friday.

Once you have that data, three things become possible:

You can calculate true lost sales (velocity × hours OOS)
You can predict the next stockout before it happens
You can hold the right person accountable — the rep, the distributor, or the planner

We found that distributors using SKU-level OOS tracking through Zivni cut their out-of-stock rate from an average of 17% to under 6% within four months. Not because the software is magic. Because for the first time, someone was actually looking at the right number.

The four causes nobody separates

Here's the thing most teams miss — "out of stock" isn't one problem. It's four, and the fix for each is completely different.

Phantom inventory. The system says you have 40 cartons. The shelf is empty. This is usually a merchandising issue — stock sitting in the back room because the retailer didn't refill, or because your rep didn't push it forward during the visit. Fix: photo-based shelf checks and merchandising scoring on every beat visit.

Forecast failure. You shipped 100 units, demand was 180. Classic. Fix: secondary sales data feeding back into your demand planning. If you're forecasting based only on primary dispatch, you're forecasting based on what you pushed, not what shoppers actually bought.

Beat coverage gaps. The rep visited the outlet 11 days ago instead of 4. By the time they come back, the SKU has been out for a week. Fix: beat plan adherence tracking with GPS verification. We see compliance jump from around 60% to over 90% once reps know the visits are being logged accurately.

Distributor hoarding or skewing. The distributor is pushing the higher-margin SKU and quietly letting the low-margin one stock out. Happens more than people admit. Fix: independent secondary sales tracking that doesn't rely on the distributor's self-reported numbers.

If you treat all four as the same issue, you'll throw money at the wrong solution. I've watched brands buy expensive demand planning tools when their actual problem was a rep skipping 30% of his beat.

What a data-driven OOS program actually looks like

Look, I'll be honest — I used to think the answer was just better dashboards. Build a nice OOS heatmap, give it to the regional sales manager, problem solved. That was wrong.

Dashboards don't reduce OOS. Field actions do. The dashboard just tells you where to act.

So the workflow that actually moves the needle looks like this:

Every rep visit captures shelf state — either through a quick checklist or, better, an AI shelf photo analysis (we use this in Zivni and it cuts visit time by about 40 seconds while catching stuff reps miss).
OOS events get tagged with a reason code at the moment they're detected, not reconstructed later from memory.
The system auto-generates a replenishment task — sometimes back to the distributor, sometimes a direct alert to the brand's KAM.
A weekly review compares OOS rates against beat compliance, distributor stock cover, and SKU velocity. Whichever signal is leading the OOS spike gets the intervention.
On-shelf availability improvement gets measured per rep, per distributor, per region — and tied to incentives.

That last part matters more than people think. If your reps' bonus is tied to primary dispatch only, OOS will never go down, no matter what software you buy. Tie 20-30% of variable pay to on-shelf availability and watch what happens.

The unsexy truth

Reducing out-of-stocks isn't a technology problem. It's a measurement and accountability problem that technology makes solvable.

The brands winning this game in Pakistan and the Gulf right now aren't the ones with the fanciest AI. They're the ones who decided that OOS is a number worth tracking weekly at the leadership level — same as revenue, same as gross margin. Once it's on the CEO's dashboard, it gets fixed. Before that, it just gets discussed.

If you're sitting on an OOS rate above 10% and you don't know which of the four causes is dominant, that's where I'd start. Not with new software. With one week of clean data on where, when, and why your shelves are going empty.

Then you can decide what to buy.