Why "number of visits" is a bad metric
In most companies, the headline merchandiser metric is: "how many stores did you visit today." The problem is that this measures activity, not outcome. It's easy to fake and impossible to tie to sales.
A merchandiser can "visit" 15 stores by popping into each for two minutes and changing nothing. Or work deeply in 8 and actually fix the display, replenish the shelf, log competitor moves. By the old metric the first is a "hero" and the second is "underperforming" β even though the second one creates the value.
Number of visits, kilometers driven, hours "in the field" β these look great in a report but don't correlate with sales. If a metric can be hit without creating value, it's a bad metric.
A good KPI answers three questions: was the merchandiser where they should be; what did they actually do there; and did the shelf get better? Below are 8 metrics that together give an honest picture. You don't need all of them at once β start with the first three.
1. Visit compliance (GPS-verified)
Basic but critical: the share of planned visits that actually happened β and happened in the right place. The key word is "actually." Without GPS verification you're not measuring visits, you're measuring tick-boxes.
A GPS check-in tied to the store's coordinates removes the most common manipulation β "reporting from home" or "from the car outside." This isn't about distrust: without verification, even an honest team gradually optimizes effort in the wrong direction.
2. Time on site (dwell time)
How long the merchandiser actually spent in the store. It works in tandem with the previous one: a 90-second visit isn't a visit, it's a check-in. Real shelf work physically can't take two minutes.
Don't read time on site in isolation (a big store needs more time than a kiosk), but anomalies stand out instantly: if the average visit to a chain supermarket is 3 minutes, nobody there is working the shelf.
Compare time on site to the norm for that store type, not "the average across all." And watch the link: does more time correlate with a better shelf? If not, the problem isn't time β it's what the merchandiser is doing with it.
3. On-Shelf Availability / OOS
The core "product" metric: is your product on the shelf? A merchandiser exists first and foremost to keep the shelf from being empty. So On-Shelf Availability (OSA) is a direct measure of the outcome of their work, not their activity.
It's captured automatically from a shelf photo: the system recognizes SKUs and sees what's missing. OSA is then calculated by store, by SKU, and by merchandiser β and you instantly see whose route "breathes" and whose shelves are systematically empty.
target OSA on key SKUs
If a merchandiser keeps OSA at 95%+ on their route, they're doing the main thing. If OSA drops, the other metrics are secondary: an empty shelf doesn't sell no matter how nicely it's arranged.
We covered how much empty shelves cost and how to measure them in our article on out-of-stock β we recommend reading it alongside this one.
4. Share of Shelf
How much shelf space your brand occupies relative to the whole category and competitors. Share of shelf correlates directly with share of sales: more facings, higher purchase probability. It's one of the few things a merchandiser can influence every day.
It's important to measure this via Vision AI rather than "by eye": manual facing counts are slow, subjective, and impossible to verify. A photo gives an objective number and proof.
5. Planogram compliance
Whether the product is placed as agreed with the retailer and brand: right positions, right order, right shelf zone. A planogram isn't bureaucracy β it's the result of agreements you often paid for (listing, shelf space, promo zone).
If the merchandiser doesn't hold the planogram, you're paying for space you don't use. The metric shows how well the real shelf matches the agreed layout.
6. Photo report quality
Not "did they send a photo," but is it usable for analysis: the full shelf in frame, not blurry, correct angle, before/after where needed. A bad photo = no data = no control.
This metric matters most during rollout: it disciplines the team and ensures every other metric (OSA, Share of Shelf, planogram) can be calculated at all. Vision AI helps here too β it tells the merchandiser right on site if the shot is unusable.
7. Issues fixed per visit
The most underrated metric. A merchandiser doesn't just record the state β they fix it: replenish the shelf, restore planogram order, remove expired stock, bring product out from the store's back room. This metric counts how many issues were found and how many were closed during the visit.
This is the metric that separates a "photo walk" from real work. And it's the one that correlates best with route-level sales.
8. Perfect Store Score
A composite metric that rolls everything into a single number per store. Perfect Store Score is the percentage of "perfect store" criteria met: key SKUs present, share of shelf, planogram compliance, promo materials in place, correct price tags, and so on.
The advantage is one clear number for the merchandiser, the supervisor, and the brand alike. It aggregates individual metrics into a KPI that's easy to target and track over time. For the pitfalls of such audits, see Perfect Store: Why 9 out of 10 Audits Give False Results.
Don't roll out all 8 at once β the team will drown. Start with three: Visit Compliance (with GPS), OSA, and Share of Shelf. That already gives an honest "was there / did the work / shelf improved" picture. Add the rest once the first three become a habit.
How to bring this into one dashboard (and what to avoid)
Metrics only work when the supervisor sees them in one place and in real time β not hand-assembled from Excel at month-end. A good field dashboard shows:
Three anti-patterns to avoid:
The bottom line: merchandiser KPIs exist not to punish anyone, but to make field work visible and manageable. When you measure outcomes (OSA, share of shelf, issues fixed) instead of activity (number of visits), the team starts working on what actually moves sales. And you finally see which stores and people create value β and which just tick boxes.