Retail Audit · UAE & GCC
Retail Audit & Merchandising, powered by image recognition
We read MSL compliance from a photograph.
Your shelf data should not arrive three weeks after the shelf changed. Our auditors photograph the bay. Our recognition engine reads every facing, every price tag, every competitor SKU. You get store-level Must Stock List compliance, availability, share of shelf and pricing while there is still time in the quarter to act.
One photograph, twelve KPIs
This is the whole data-capture step
The auditor raises the phone, frames the bay and shoots. Recognition runs against the image and every tracked SKU is identified, counted, priced and scored against your Must Stock List before the auditor has walked to the next aisle.
What we measure
Twelve shelf KPIs from one set of photographs
| KPI | What you learn |
|---|---|
| MSL compliance | Whether every SKU that store is obliged to carry was actually on shelf, scored per store and per cluster. |
| On-shelf availability | Which of your SKUs were physically present, store by store, visit by visit. |
| Out-of-stock and voids | Which listings are missing entirely versus temporarily empty, and where the pattern repeats. |
| Share of shelf | Your facings as a percentage of category facings, against your fair share of category value. |
| Share of visibility | Total brand presence including secondary displays, gondola ends and cooler space. |
| Planogram compliance | Scored against the planogram you supplied, not an auditor’s impression of it. |
| Price compliance | Shelf price read from the tag, checked against your RRP and competitor pricing. |
| Promotion compliance | Whether the mechanic you paid for actually ran, in that store, in that window. |
| POSM and display | Whether the material went up, and whether it went up correctly. |
| Competitor activity | New SKUs, price moves and space gains in your category, captured in the same visit. |
| Stock condition | Near-expiry flagged at shelf before it becomes a returns conversation. |
| Photo evidence | Every claim above, backed by a timestamped, geotagged image. |
A note on MSL, because it gets confused with OSA
On-shelf availability asks: of everything we track, what was on shelf? It is a portfolio health measure. A 91% figure sounds strong until the 9% missing turns out to be your three highest-margin lines.
MSL compliance asks a harder question: of the SKUs this specific store agreed to carry, what was actually there? That is a contractual measure, and it is the number you can take into a retailer meeting.
MSL varies by store cluster, so a single national figure tells you almost nothing actionable. We score each outlet against the list that applies to it, then roll it up by chain, cluster and region, so you can see whether the gap is a distributor problem, a store-execution problem, or a listing that was never activated.
Planogram compliance
Is the shelf built the way you agreed it should be?
A planogram is a promise: the layout, the facings and the block your brand paid for. We score the shelf you actually have against the planogram you signed off, store by store, and flag every place the aisle and the agreement stop matching, from a dropped facing to an out-of-stock gap to a competitor creeping into your block.
How it works
Four steps, and only one of them involves you
We build your product library
You send your SKU list, pack shots and Must Stock List by cluster. We train recognition against your portfolio and your named competitor set, and keep it current so a Ramadan pack change does not break your data.
Our auditors photograph the shelf
Trained field staff work your outlet list on an agreed cycle. The app checks image quality on the spot and records GPS and timestamp on every visit. Offline capture is supported for basement supermarkets and traditional trade.
The images are processed and QC’d
Recognition runs against the images, and our quality-control team reviews the output. Anything ambiguous is human-verified before it reaches your dashboard.
You get the data where you want it
Dashboards with drill-down from national to store level, scheduled reports, or a feed into your own BI. Your team gets alerts on the exceptions, not a monthly PDF nobody opens.
Why SMRC
An agency that runs the technology, not a platform that sends you an invoice
Buying an image-recognition platform means you own the rollout: recruiting the field team, chasing store permissions, maintaining the SKU library, and carrying the licence cost whether you use it heavily in March or barely in August. Working with SMRC means you buy the answer. We own the fieldwork, the recognition, the quality control and the reporting, and you pay per audit cycle.
Field infrastructure that exists
A standing retail panel across the UAE and Oman, staffed by a team that includes ex-MEMRB and ex-GfK researchers. No recruitment lag.
Store access
Relationships with modern and traditional trade across the UAE and Oman that take years to build and cannot be licensed.
Judgement on the data
Recognition says share of shelf dropped four points in Sharjah. A researcher tells you whether it is a distributor issue, a competitor promotion or a reset, and what to do.
Arabic and English fieldwork
Across nationalities and store formats, in a market where that is not optional.
Benchmarks
What image recognition typically delivers
Published across the image-recognition category. Platform providers publicly report SKU recognition accuracy in the 95 to 98% range, in-store audit-time reductions of up to 50%, and shelf-image processing in under 60 seconds. These are category indicators published by vendors, not independently audited findings, and actual performance varies by category, pack complexity and store conditions.
We benchmark against your existing manual process during the pilot and show you the delta on your own data.
Coverage
Where we audit, and what we audit
Markets
United Arab Emirates and Oman as core markets with a standing panel, and wider GCC and MENA coverage on a project basis.
Channels
Categories with panel coverage
FAQ
Questions we get asked
How is an image-recognition retail audit different from a normal one?
The fieldwork looks similar: an auditor visits on an agreed cycle. What changes is the capture. Instead of counting facings and prices by hand, the auditor photographs the shelf, and SKU identification, facing counts, share of shelf, price reading and planogram scoring are generated from the image. That removes manual counting error, cuts turnaround, and leaves photographic evidence for every data point.
What is the difference between MSL compliance and on-shelf availability?
On-shelf availability measures how much of your tracked range was physically present. MSL compliance measures how much of the range a specific store is obliged to carry was present. OSA is a health measure; MSL is a contractual one, and it is the number that gives you standing in a retailer conversation. Because MSL varies by cluster, we score each outlet against its own list, then roll it up by chain, cluster and region.
How accurate is it compared with a human auditor?
On facing counts and SKU identification, recognition is consistently more accurate than manual counting, because it does not fatigue and does not estimate. Accuracy depends on image quality, pack similarity and how well the SKU library is maintained. We quality-control the output and human-verify anything ambiguous before it reaches your dashboard.
Can you audit traditional trade, or only modern trade?
Both. Traditional trade is harder, with smaller shelves and less predictable layouts, but it is where a lot of GCC volume sits and our panel covers it. Offline capture means connectivity is not a constraint.
Do we need to buy software?
No. You buy audit cycles. The recognition technology, fieldwork, quality control and reporting sit with us.
Which markets do you cover?
UAE and Oman as core markets with a standing panel, and wider GCC and MENA coverage on a project basis.
See what your shelf actually looks like this month
Send us your outlet list, SKU range and Must Stock List. We will come back with a scope, a cycle cost, and a sample of the output on your own category.
