See the stock. Track the trend.
Your reps are already in the store. ShelfPro turns the photo they take into on-shelf availability, share of shelf and planogram compliance in seconds — and tracks every number, visit after visit.
See ShelfPro in action
Left of the handle is the photo the rep took. Right of it is what ShelfPro reads from it: every pack, every gap, every price tag.
- Soft drinks9
- Juice & tea5
- Water4
- Out of stock3
- Price tags17
What this one photo turns into
- On-shelf availability
- 86%18 of 21 facings stocked, 3 flagged out of stock.
- Share of shelf · Aqua
- 19%4 of 21 facings, gaps included in the shelf.
- Price tags read
- 17One on promotion: Gold Reserve 2L at €1.99, down from €2.49.
- SKUs recognized
- 6Every facing matched to a product in the catalog, with a score.
The shelf, its boxes, scores and price readings are an illustrative fixture in the API's own data shape, drawn rather than photographed — no customer imagery is cleared for public use.
Category-wide figures from published research, not ShelfPro claims.
You cannot fix what you cannot see
Manual shelf audits are slow, inconsistent and generous. By the time the report reaches head office, the promotion is over and the sale is gone.
Manual audits are slow and generous
A rep under time pressure checks “compliant” and moves on. Counts vary by person and by day, and nobody can check them afterwards.
The data arrives after the promotion ends
A gap found in next month's report is a sale already lost. A weekly spreadsheet cannot rescue this week's display.
A photo in a chat group is not a dataset
Close-ups capture fragments. Share of shelf and competitor space need the whole fixture, in a form someone can actually count.
Three steps. No change to how your team already works
Capture
The rep photographs the shelf on the phone they already carry. Wide fixtures are captured in a single pan, with on-screen guidance for framing. It works with no signal — the store basement is not a problem.
Recognize
Our AI identifies every facing and every gap against your product catalog, then turns them into the KPIs your commercial team already reports on.
Act
Results are ready before the rep leaves the store. Missing stock gets flagged to the store manager on the spot, not in a report next month.
All your shelf KPIs from a single photo
Everything below comes out of the same capture. No extra visit, no separate audit program, no clipboard.
1. On-shelf availability
Which of your SKUs are missing, in which stores, right now — the single number most directly tied to lost sales.
2. Out-of-stock alerts
Gaps are flagged while the rep is still in the aisle, so the store manager can fix them during that visit rather than after the next one.
3. Share of shelf
Your facings against the category and against named competitors. The evidence you take into a retailer negotiation.
4. Planogram compliance
Whether the shelf actually looks like what you paid the retailer for — position, block, eye level, facing count.
5. Promotion execution
Did the display go up, in the right stores, on the right day? Measured while the promotion is still running.
6. Competitor activity
New SKUs, new packs and space grabs, spotted from the same photos your reps already take.
7. Field team performance
Coverage, visit quality and consistency by rep and by region — coaching data, not surveillance.
8. An evidence trail
Every number traces back to a timestamped photo from a known store and visit. Nobody argues about whether the display was really up.
9. Your reporting stack
Results flow into your BI or field tool through a documented API — or your team reads them in ShelfPro's own dashboards.
The stockroom has no reception — but the visit still gets done
Basements and back rooms kill the signal. ShelfPro captures, checks and queues every shelf on the device, then uploads in the background once the signal returns. The rep never waits for a spinner and never loses a visit.
- Capture and check framing with no connection at all
- Uploads resume on their own when the signal comes back
- Nothing is retaken, nothing is lost — every capture is kept until it is safely through
Store 104 · aisle 4
Shelf 1 of 3 · 10:42
2 captures queued
Uploads when signal returns
3 overlaps · 36 facings, each counted once
A long shelf is one scene, not a dozen overlapping photos
Reps pan along the fixture and ShelfPro joins the frames into one continuous shelf before anything is counted. Facings at the seams are counted once, and share of shelf is measured across the whole fixture rather than guessed from close-ups.
- No double counting at frame boundaries
- Share of shelf across the full fixture, not a sample of it
- The whole fixture on one review screen, gaps and all
What changes in the first quarter
Reps spend their visit selling
Counting facings by hand is the least valuable thing a rep does in a store. A photo takes seconds and is more accurate than the count it replaces.
Data you can defend
Every number traces back to a timestamped photograph. Nobody argues about whether the display was really up.
Problems surface same-day
Availability gaps reach the person who can fix them while the shopper is still in the store, not in a monthly deck.
It gets better as you use it
When your team corrects something the model got wrong, that correction trains the next version on your catalog.
Against the clipboard you're using today
Time per shelf audit
20–30 minutes
Under a minute
Consistency between reps
Varies by person and by day
Identical every time
When you see results
Next weekly or monthly report
Before the rep leaves
Evidence behind a number
A checkbox on a form
A timestamped photograph
Competitor visibility
Anecdotal
Measured from the same photo
Cost of covering more stores
Linear — more reps
Marginal
What would ShelfPro be worth to you?
Move the sliders to your own numbers. The math is shown, the assumptions are yours to change, and nothing you type is sent to us.
Your numbers
Change anything here. Nothing is sent to us.
Outlets in your visit plan
Your call cycle
Sell-out value
Category average is ~8% (NielsenIQ)
Estimated annual impact
€1,200,000
sales recovered from fewer empty shelves
€4,800,000 is currently at risk; this assumes you close 25% of it.
€66,880
field time released
3,040 hours a year — 19 minutes saved across 9,600 visits.
Reported separately on purpose: recovered sales and saved cost aren't the same line on your P&L. Apply your own gross margin to the first figure before comparing the two.
Before you book a call
- Do we need new phones or hardware?
- No. It runs on the Android and iOS phones your reps already carry. No scanners, no rigs, no shelf cameras.
- What if there's no signal in the store?
- Captures are stored on the device and upload themselves when signal returns. Reps never wait and never lose a visit.
- How long until it recognizes our products?
- A pilot trains a model on your catalog and typically runs for eight weeks, including the first rounds of field capture.
- What happens when it gets something wrong?
- Your team corrects it in the review screen in a click, and that correction becomes training data — so the same mistake gets rarer.
- Can it work with our existing field app?
- Yes. ShelfPro can run as your capture app or sit behind the one you already use, feeding results into your reporting stack.
- Where is our data held?
- In the EU. We're based in Zagreb and operate on EU infrastructure.
See it on your own shelves
Send us photos from a handful of your stores. We'll come back with the analysis and what it says about your availability — before you commit to anything.