Rivet & Field
Your busiest weekday hours are your weakest converting hours. Shoppers are stopping at the hardware wall, waiting, and leaving, and it is visible on camera.
Four findings. Four actions. One week to act.
We watched the store for three days, 1,247 visitors, 27 hours of footage, seven mapped zones. Each finding ends with a specific action, an owner, and a metric that tells you whether the change worked.
Snapshot
| Zone | Name | Visitors | Draw rate | Median dwell | Register-reach |
|---|---|---|---|---|---|
| Z1 | Entry & decompression | 1,247 | 100% | 0:11 | n/a |
| Z2 | Front promo endcap | 886 | 71% | 0:04 | n/a |
| Z3 | Workwear apparel | 673 | 54% | 3:10 | 31% |
| Z4 | Footwear | 474 | 38% | 5:22 | 34% |
| Z5 | Hardware & accessories wall | 362 | 29% | 1:24 | 11% |
| Z6 | Clearance / rear | 112 | 9% | 2:05 | 22% |
| Z7 | Checkout & queue | 342 | 27% | 2:48 | n/a |
| Store | 1,247 | n/a | 7:40 | 27.4% |
Draw rate is the share of the 1,247 visitors who entered that zone at least once. Register-reach is the share of that zone's visitors who later reached the counter line, a stand-in for conversion, not a measure of it.
One zone breaks the pattern. Z5 draws well and converts badly, a third of the rate of every other shopping zone. That single row is the origin of the highest-priority finding.
Shoppers stop at the hardware wall, wait, and leave, nobody is standing there
On Thursday and Friday afternoons, shoppers stood at the hardware and accessories wall for more than a minute at a time, looked around, put items back, and walked out. No associate entered that zone during almost the entire block.
Of 362 visits to Z5, 128 lasted longer than 60 seconds, the threshold at which a stop stops being a glance. 71 of those fell inside the Thursday and Friday 15:00–17:00 block, the busiest weekday window. Across those same 240 minutes, an associate was inside the Z5 polygon for a total of nine. Staff were most often tracked near the door and the counter.
The hovers are not random. 44 of the 128 ended with the shopper's route turning back toward the door without entering the checkout zone at all.
If sustained hovers at Z5 reached the register at the store-wide rate of 27.4%, that would be 35 register visits from those 128 shoppers instead of 14, a gap of 21 over three days, concentrated in two afternoon hours. That figure is arithmetic, not a forecast: it assumes hardware shoppers are as convertible as apparel shoppers, which cannot be verified without POS.
How we'll check. Re-measure Z5 hovers, hover-to-register rate and staff-minutes-in-zone over four weeks against this baseline, roughly 8–10× this audit's sample, enough to separate a real change from noise.
The Saturday lunch queue crosses five minutes, and people leave the line
For most of the Saturday lunch hour the checkout line ran longer than five minutes, and seven people who had joined it walked back out without reaching the register.
Wait is estimated from queue depth and observed service rate, not measured per person. One register was open for the full window; a second counter position exists but was unstaffed.
The cost is not only the seven baskets. It is the share of the 96 people who walked in during that hour who saw the line on their way past the counter and adjusted their behaviour accordingly.
How we'll check. Count queue abandonments and minutes-above-threshold on the next four Saturdays against this baseline of seven and 31.
The promo endcap is in the six feet where nobody is looking yet
886 people walked past the seasonal endcap. Fifty-three stopped. The median time spent in front of it was four seconds, the time it takes to walk past something.
The endcap sits roughly six feet inside the door, inside what retail layout practice calls the decompression zone, the stretch of floor where people are still adjusting to being inside and are not yet shopping. Traffic past it is excellent; attention is close to zero. The comparison that makes this actionable is the store's own: the front table fifteen feet further in stopped 19% of the people who passed it during the same window. Same store, same shoppers, same days, three times the stop rate.
This is the clearest example in the report of a question a POS cannot answer. Sales for the promoted line were soft; the sales data tells you the promotion underperformed. The camera tells you why, it was never seen.
How we'll check. Stop rate at the relocated fixture against the 6% baseline, over the following three weeks.
Two-thirds of your shoppers never reach the back third of the store
The median shopper turns around 31 feet in, just past the footwear zone. Visitors who did reach the rear behaved normally once there, with a 2:05 median dwell and a 22% register-reach rate, which suggests the rear is not repellent so much as unvisited. Nothing in the trajectory data shows people reaching the back and retreating; they simply never go.
Whether that matters depends on what is back there, clearance may be exactly where you want low-attention, high-margin-recovery stock. But if the intent is for clearance to be discovered, it currently isn't.
What to expect: unknown, and we would rather say so. How we'll check: three weeks of baseline first, because three days is not a baseline for a question this structural, then three weeks with the category moved.
Privacy: what the system does not do
The design decisions below are deliberate. They limit what we can offer you, and they are the reason this deployment sits well clear of the laws that make camera analytics risky.
Want this for your store?
A behavioral audit runs on the cameras you already have. No new hardware, no installation. One week of footage, findings in about five days, and you only pay after you've read it.
Book the report