The Merchandising App That Actually Improves On-Shelf Availability

By Sufyan · 2026-04-28 · 4 min read

Last month I was sitting in a kiryana store in Gulberg, Lahore, watching a merchandiser from a top-3 biscuit brand do his job. He opened his app. Took four photos of the shelf. Tapped a checkbox. Left.

The shelf had two SKUs out of stock. Nobody fixed them. Nobody told the distributor. The photos uploaded somewhere, into some dashboard, and that was it.

This is the dirty secret of most merchandising software. It collects data. It doesn't fix anything.

And honestly, when we started building the merchandising module inside Zivni, I got this wrong at first too. I thought the answer was better photos, sharper AI, fancier dashboards. Turns out the real problem is much more boring — it's the gap between seeing a problem and acting on it.

What on-shelf availability actually costs you

A study by IHL Group pegged global OSA losses at $1.2 trillion a year. But forget global numbers for a second. I'll give you a Pakistani one.

A mid-size beverage distributor we work with in Karachi was losing roughly 14% of potential weekly sales to out-of-stocks at the shelf — even when the SKU was sitting in their warehouse. Not the factory. Their own warehouse. The product existed. The shopkeeper wanted it. The shelf was empty. And the merchandiser who visited that morning had marked the visit "complete."

That's the actual problem. OSA isn't a measurement problem. It's a closing-the-loop problem.

So when people ask me what makes a good merchandising app, I don't talk about photo quality first. I talk about what happens in the 30 minutes after a gap is detected.

The four things a merchandising app needs to actually do

Most shelf monitoring tools do one thing well — capture. The good ones do four.

1. Detect the gap automatically. If your merchandiser has to manually count facings and type SKU codes, you've already lost. They won't do it on visit number 11 of the day. Our AI shelf analysis runs on the photo the rep already took — it pulls SKU counts, share of shelf, planogram compliance, competitor facings. The rep doesn't do extra work. The system just reads the picture.

Tell someone who can fix it. This is where 90% of merchandising software falls apart. A gap is detected and... it sits in a report. Nobody pings the distributor. Nobody pings the order-booker assigned to that beat tomorrow. In Zivni we route detected OOS straight into the order management flow — so the rep visiting that outlet next gets a pre-filled suggested order with the missing SKUs at the top.
Track whether the gap actually closed. If a Pepsi facing was missing on Monday, was it back on Friday? If not, why? Was the SKU out at the distributor too? Was the shopkeeper refusing to stock it? Was a competitor paying for that slot? You need the second visit to talk to the first one.
Show the pattern, not just the snapshot. One empty shelf is a story. Forty empty shelves across the same town on the same SKU is a supply chain problem. Most merchandising software treats every visit as isolated. That's useless for a brand manager trying to figure out whether their Hyderabad depot is the bottleneck.

Why "taking photos" became the whole industry

Here's the thing — image recognition got cheap around 2019. Suddenly every field sales platform added "AI shelf analysis" to their feature list. FieldAssist, BeatRoute, us, Salesforce, everyone. And for a while we all sort of pretended that having the feature was the same as solving the problem.

It's not.

I used to think shelf analysis was the killer feature. Then I sat with a regional sales manager at an FMCG company in Dubai who showed me his dashboard. Beautiful charts. Share of shelf trending nicely. He scrolled, smiled, then said: "But my sales in Sharjah are down 8% this quarter and I have no idea why."

The data was there. The connection wasn't.

So we rebuilt how merchandising data flows inside Zivni. Shelf gap detected → routed to the rep's next-visit task list → tied to a suggested order → tracked until the SKU shows up on the next photo. If it doesn't show up after two visits, it escalates to the area manager. If it doesn't show up after a week across multiple outlets, it flags the distributor.

Boring plumbing. But that's where OSA actually improves.

What to ask before you buy any merchandising software

When distributors and brand managers ask me how to evaluate a merchandising app — whether it's ours or someone else's — I tell them to skip the demo for a minute and ask three questions:

After a gap is detected at 11am, what specifically happens by 5pm the same day?
Can the same app place an order, or does my rep have to switch tools?
Can I see, for any given SKU, the number of outlets it was missing from this week vs. last week?

If the answer to any of those is "well, you'd export the data and..." — walk away. You don't need another dashboard. You've got fifteen of those already.

Look, I'm biased. I built Zivni and I obviously think we do this better than the alternatives, especially for FMCG companies in Pakistan and the UAE where margins are thin and a 3% OSA improvement is the difference between a good quarter and a flat one. At $5 per user per month it's not really a budget conversation either.

But even if you don't pick us — pick something that closes the loop. The photo isn't the point. What happens after the photo is the whole game.

When was the last time you actually checked what your merchandising app does in the six hours after it spots an empty shelf?