01
Capture the shelf
The app guides overlapping shots of each fixture, and joins them into a panorama only when the overlap is trustworthy.
Recognition, plan checks and reporting, built around the store visit.
01
The app guides overlapping shots of each fixture, and joins them into a panorama only when the overlap is trustworthy.
02
Each pack is matched to your catalogue, keeping its photo, crop, confidence and the candidates it was weighed against.
03
Presence, placement and POSM are compared with your plan or planogram. Gaps and wrong variants are flagged.
04
Facings, share of shelf and POSM, counted by rules you agree.
05
By store, brand or territory. Each result keeps the plan version it was scored against, so history stays comparable.
A fast result while the team is still in the store. A complete one for the whole category later. One visit record.
Store visit
Photos submitted and accepted
The service clock starts here.
+ 5-15 minutes
Store visit result
Phase 1 · provisional where coverage is incomplete
+ 8 hours to 30 days
Full category result
Phase 2 · labelled complete when done
Phase 1
5-15 minutes
from an accepted submission, depending on your service tier
Phase 2
8 hours to 30 days
from an accepted submission, depending on your service tier
Store visit results run at high priority and are never held behind full-category work, which is scheduled around them. How service levels are measured
Plan gaps with enough evidence become a ranked list of things to fix.
Where coverage is insufficient, the action is to capture or review, never to replenish.
Two assisted tools, for a product that was missed and a match that is wrong.
The reviewer marks the visible pack; Pharos offers ranked catalogue matches. It only works for a pack that can be seen.
Pick the right candidate from those already attached, or record that none fits.
Candidates attached to this detection
Correction saved · original match kept · queued for quality review
The original match is kept beside each correction, with who, why and when, and reviewed before it touches the catalogue.
Choose the plan detail per category and fixture.
The expected range, with facings and shelf, without a fixed grid.
A descriptive plan is checked across every photo from the visit, not against one fixed position. Pharos matches each SKU to what’s on the shelf by its real product identity or barcode first, so a corrected barcode never breaks the match later; only when a client turns it on for a specific SKU will Pharos also match by brand and product name, kept off by default so two pack sizes of the same product are never credited for each other.
For every SKU on the plan, Pharos reports whether it’s present, how many facings were found, whether that meets the minimum you set, and whether it’s in the right zone or fixture area when one applies; and, where you set a reference price, whether the shelf tag matches it. Anything found on the shelf that isn’t on the plan at all is reported separately, so an unplanned product is never silently folded into the count.
Illustrative example. Findings depend on your catalogue and the merchandising rules you agree.
An exact layout, with brand blocks and the sequence of products.
A planogram sets an exact layout: which product belongs in which position, shelf by shelf. Pharos first works out which physical shelf in a photo matches which shelf on the plan by what’s actually on it, not just how high up the photo it sits, so a photo that only covers part of a fixture, or a shelf split across two photos, still lines up correctly. When a plan describes only one brand’s block on a wider shelf, Pharos finds that block within the row automatically and checks only against it.
Within a shelf, every planned position is checked against the whole row at once, so one misplaced pack never throws the rest of the row out of step. Each position then gets one clear result: it matches, the wrong product is physically there, the right product is elsewhere on the shelf or on a different shelf, it’s genuinely missing, or the shelf simply wasn’t photographed, which is never scored as a failure. If the fixture itself doesn’t match what the plan describes, Pharos flags that the plan needs updating, rather than reporting the whole shelf wrong.
Scroll the diagram sideways to see it at full size.
Each material has a location, dates and visibility rule, and is reported as present, missing or unverified.
Each visit is scored against the cycle active on its date, so past results stay comparable after a new range goes live.
Mobile-first, three views, the same evidence.
Sees the plan, captures, gets the result in the store, and closes the actions.
Follows visits, open actions and completion across their stores.
Owns plans, cycles and campaigns, and reads patterns across the network.
No long manual coding. Setup is your catalogue, your plans and capture guidance, validated in a pilot.
Names, variants and pack sizes must mean the same thing everywhere. We prepare it with you.
A photo shows one visible moment: not backroom stock, not what was out of frame. Pharos works from store visits, not continuous monitoring.
Private-first. The same service runs on your infrastructure or on dedicated servers we operate.
You control access, residency and operating windows.
A private environment under agreed controls.
Optional, for peaks, only where your policy allows.
Every result records where each phase ran, including any fallback. Data & security