Giving Supervisors a True Picture of the Shelf

A merchandising team is only as good as what's actually on the shelf — the right facings, real availability, products where they should be. But the people checking the shelf were also the people being measured on it. Reports were self-graded, inconsistent, and impossible to trust at scale.

The challenge

• check_circleSupervisors had no objective view of share of shelf, availability or planogram compliance across stores.
• check_circleMerchandiser visits were self-reported — there was no proof a store was actually serviced.
• check_circleProduct life-span and expiry tracking on the shelf was manual and patchy.

Why manual audits fell short

• check_circleA clipboard audit is a snapshot, biased by who walked the aisle and how carefully they looked.
• check_circlePhotos were taken but never analysed — they sat in phone galleries.
• check_circleKPIs were defined inconsistently store to store, so nothing was comparable.

You can't manage a shelf you only ever see through someone else's summary of it.

What we built

PetalKube extended its Merchant Eye approach. A merchandiser — or a fixed camera — captures a shelf image, and object detection identifies every product, then computes the KPIs supervisors actually care about: share of shelf, on-shelf availability, facings versus planogram, competitor presence, price and promo compliance, and product life-span. The same capture verifies which merchandiser serviced which store, and when.

How it works

• looks_oneCapture — a shelf photo from a phone or fixed camera, taken during a normal visit.
• looks_twoDetect — every SKU on the shelf is identified and counted, including competitors.
• looks_3Compute — share of shelf, facings, gaps and out-of-stocks are calculated instantly.
• looks_4Compare — results are checked against the planogram and pricing rules; expiries are flagged.
• looks_5Verify — the visit is logged with store, time and merchandiser, confirming coverage.
• dashboardSurface — supervisors get a live KPI dashboard and alerts for stores that need action.

The outcome

Supervisors stopped managing on anecdote and started managing on data. Share of shelf became a number they could trend, gaps were caught the same day, and field activity was finally verifiable — turning merchandising from a trust exercise into a measured operation.