The run reads 92% complete.
The two sizes missing
are half your sales.
A run with a hole in it is a lost sale you never see — because the completeness number is counted in sizes nobody buys.
Illustrative · not a client store
The AI that
Size-run completeness, pairs per hour and full-price sell-through by door — on one board, narrated by Iris.
The missing size. The second item.
The whole run.
Footwear fails quietly. Nothing goes wrong on the shop floor — a customer simply asks for a nine, hears “not in your size”, and leaves without ever appearing in a report.
A run with a hole in it is a lost sale you never see. This is the one that sees it — and fills it first.
Iris finds the gap the day it opens.
What actually sells together — bag with the shoe, belt with the formal — prompted while the customer is still at the till.
Iris knows what sells with what.
Every article, every size, every door — counted nightly so the gap shows up before the customer does.
Iris checks the run so nobody has to.
Every door
on one board.
Rules fire. Models predict.
Iris explains.
| Rule | Threshold | Now | State |
|---|---|---|---|
| Gross margin | ≥ 46% | 48.6% | clear |
| Attach rate | ≥ 31% | 34% | clear |
| Size-run completeness | ≥ 88% | 91.8% | watch |
| Article forecast accuracy | ≥ 88% | 92.2% | clear |
| Model | Predicts | Accuracy |
|---|---|---|
| Run integrity | hole risk by article | 92.7% |
| Attach propensity | pair-with likelihood | 90.4% |
| Article curve | 8-week sell-through | 91.1% |
91.8% complete sounds like a rounding error away from fine. It is not: the count is dominated by sizes almost nobody buys, and the missing ones are UK 8 and 9 — 46% of every pair sold. The rule fires, the model names the nine articles, and Iris says the thing a buyer needs to hear on a Monday: fill the holes you already have before you touch this week's intake.
Pick one. It opens
the real screen.
Size-run completeness, pairs per hour and full-price sell-through by door.
Point of Sale
Footwear is the one category where the second item is genuinely expected — a bag with the shoe, a belt with the formal. At 132 pairs an hour the prompt has to be right the first time or it is noise.
The size is captured, not just the sale. What sold in which size is the raw material for run integrity, so the till is where the next stock decision starts.
Attach is measured as a rate, not a hope. 34% is on the board next to margin, which makes the prompt accountable rather than decorative.
Iris knows what sells with what. The pair-with prompt comes from a 90.4% propensity model on this estate's own baskets, not from a category rule somebody wrote once.
Balanced Scorecard
Margin, attach rate, size-run completeness and article forecast accuracy on one live board — with Iris naming the articles whose runs have quietly broken.
Size-run completeness earns a lens of its own. In footwear it is the operational metric — a broken run costs full-price sales that never appear as a miss anywhere else.
A high number can still be the problem. 91.8% is on watch because the count is unweighted; the same board carries the size mix that explains why.
The fourth lens audits the AI. Article forecast accuracy sits beside margin, because a buy is only as good as the curve behind it.
Numerus · CFO
Trial balance, P&L, balance sheet and cash flow on a ledger that ties out to the penny — across 118 doors whose stock value sits in single boxes rather than in bulk.
A broken run is a valuation problem. Nine sizes of an article without its two best is worth materially less than the count suggests, and the balance sheet should not pretend otherwise.
Markdown at 9.4% is read against the runs. Most footwear markdown is the tail of a broken run, so the two numbers belong in the same conversation.
Ask Iris for any number. Any door, any article, any size, any period — answered in the conversation rather than a week after close.
Nexus · Supply Chain
Footwear supply is not measured in units, it is measured in runs. An article is only sellable while the middle of its size curve is intact, which makes replenishment a completeness problem rather than a volume one.
Reorder on the hole, not on the total. An article can be well stocked in aggregate and unsellable in practice, and only a size-level view can tell the difference.
Filling a run beats buying new. Iris ranks the two against each other, which is how a nine-article fill ends up ahead of a whole week's intake in priority.
Single pairs are worth moving. The transfer engine works at pair level because in footwear the smallest unit of stock is also the one that completes a run.
Motus · Operations
The wall, the stockroom and the walk between them. Footwear operations are unusual in that most of the selling time is spent behind the shop floor, looking for a box.
Stockroom time is selling time. Every minute spent finding a box is a minute not spent with the customer who asked for it, and 132 pairs an hour is what that discipline buys.
The nightly count is an operating routine. Runs are verified on a schedule rather than discovered when a customer asks, which is the entire premise of the L1 layer here.
Iris watches the exception, not the routine. The doors whose runs are intact stay quiet; the twenty-three that are not are named.
Manus · Workforce
Footwear is a service sale. Somebody measures, fetches, fits and then suggests the second item — and an attach rate of 34% is a description of how well that conversation is staffed.
Cover follows the fitting, not the footfall. A door doing 132 pairs an hour needs people on the floor and people in the stockroom, and those are different shifts.
Attach is a coachable skill on the roster. The gap between a 25% door and a 40% door is almost never the prompt — it is who is standing there.
Hours, payroll and compliance on one record. Who worked, where, at what rate — without a second spreadsheet running alongside the season.
Every solution,
on the real screen.
Grouped by what they act on. Pick one and it runs.
How Iris enforces the rules.
Thresholds that fire on the shift the breach happens (the “Zen Rules” layer) — the run, the single pair and the box on the shelf. Three solutions run on this layer.
Nine articles are out of UK 8 and 9 across twenty-three doors — the two sizes that account for 46% of every pair sold. £38k of weekly sale sits behind those holes.
Size-run integrity
| Article | UK 8 | UK 9 | Doors hit | State |
|---|---|---|---|---|
| FW-3021 Chelsea boot | 0 | 1 | 23 | broken |
| FW-2870 Court sneaker | 2 | 0 | 19 | broken |
| FW-3144 Derby brogue | 3 | 4 | 11 | thin |
| FW-3210 Trail runner | 4 | 6 | 7 | thin |
| Other 5 broken articles | — | — | 34 | broken |
Nine articles are out of UK 8 and 9 — the two sizes that are 46% of everything you sell. The run looks 92% complete because the count is dominated by sizes nobody buys. Filling these runs first is worth more than this week's entire new-arrival intake.
Single-pair transfer engine
In most categories moving one unit between stores is not worth the handling. In footwear it is the whole game — one pair is what turns a dead article back into a sellable run.
Transfers are proposed at pair level, because the smallest movable unit is also the one that completes a run.
A pair is only moved out of a door where that size is genuinely surplus, so filling one hole never opens another.
Handling cost is weighed against the full-price sale the completed run unlocks, which is what makes a single-pair move defensible.
Moves are batched by route rather than by article, so twenty-three doors are served as a round rather than as twenty-three decisions.
Every accepted and rejected proposal trains the run-integrity model, which is why hole-risk sits at 92.7% rather than at a launch-day number.
Box, shelf & stockroom locator
The system can say the nine is in the building and still lose the sale, because nobody can find it. In footwear the stockroom is not storage — it is the shop floor's other half.
Stock is held to a shelf and box position, not just to a door, so “we have it somewhere” stops being an answer.
The walk is what costs the sale. A customer waits about ninety seconds before the fitting turns into a browse, and the locator is built around that fact.
Positions are corrected as a by-product of picking, so the map stays true without a separate maintenance task nobody has time for.
A box that cannot be found is functionally out of stock, and the nightly count treats it that way rather than flattering the run.
Accurate positions are what make the nightly run count trustworthy — a completeness figure is only as good as the shelf beneath it.
How Iris predicts the article.
What the numbers say to buy next (the “Zen Models” layer) — modelled early enough to act on, with the cause named and the move costed. Three solutions run on this layer.
Article sell-through forecast
| Article | Day 10 | Projected | Plan | Call |
|---|---|---|---|---|
| FW-3021 Chelsea boot | 18% | 79% | 74% | ahead |
| FW-2870 Court sneaker | 16% | 71% | 72% | on plan |
| FW-3144 Derby brogue | 8% | 44% | 70% | −26 pts |
| FW-2996 Khussa flat | 22% | 84% | 76% | ahead |
| Other 28 articles | — | 69% | 71% | on plan |
The shortfall is one category, not one article — formal projects 44% while everything else is at or ahead of plan. Office footfall has not returned to these doors. Cutting the formal balance and moving it into boot, which is running ahead, lands the season on plan.
Fit & return-reason learning
A return in footwear is almost always a fit statement, and it is almost always thrown away. Read properly it is the cheapest product research a retailer owns.
Return reasons are captured as structured data at the counter, not as free text nobody reads afterwards.
Fit behaviour is learned per last and per supplier — a nine that runs small is a property of the mould, not of the customer.
What is learned is published back to the wall and to the site, so the next customer is told before they buy rather than after.
Persistent fit problems become supplier conversations with evidence attached, which is the only version of that conversation that changes anything.
Returns distort run integrity as they come back in, so the two systems read from the same record rather than arguing about stock.
Release & resale-pressure demand
Some footwear demand is not retail demand. When resale pressure is on an article, the queue is buying to flip it — and a retailer who cannot tell the two apart buys the wrong quantity and sells to the wrong customer.
Resale signal is modelled alongside retail demand, because a release with pressure behind it behaves nothing like a normal launch.
Sell-out speed is separated from sell-out depth — an article gone in an hour to forty buyers is a different outcome from one gone in a day to four hundred.
Per-customer limits and member rules are informed by the pressure reading rather than applied uniformly to every release.
The size curve of a hyped release is not the size curve of a staple, and buying it as if it were is how a run breaks on day one.
The model feeds the same article curve as everything else, so a release is planned inside the season rather than beside it.
Ask Iris · the run
Not a screenshot — the actual agent, reading the same estate the rules and models write to.
Twelve core POS solutions,
in every sector.
The sector solutions sit on top of these. They are not an upsell and they are not configured per customer — every Zentallio retail deployment ships with all twelve.
Implementation,
run by Iris.
You configure a sector playbook, not a custom project. Iris applies it herself — agentically, from day one.
Onboarding is agentic. Iris connects the article file, the size-curve definitions and the door list directly — no manual data mapping.
Last season's sizes are imported, not retyped. Historic sales by size carry in, which is what makes the size-weighted view meaningful in week one rather than week twenty.
Live in weeks, learning from day one. Go-live is a configuration, not a project plan.
Support starts
with Iris.
The first line of support is agentic — Iris resolves most of it herself. Our engineers pick up from there.
Layer 1 — Iris, 24/7. Configuration questions, anomalies and routine issues resolved directly, instantly.
Layer 2 — our engineers. Anything Iris can't close escalates automatically to a Zentallio engineer.
No blank tickets. Every escalation arrives with Iris's own diagnosis — engineers start from an answer.
The run, intact.
the night it opens
by what sells
buying
not thirty-two articles
Where a sale disappears
without leaving a trace.
A completeness KPI counted in sizes rather than in sales, so 91.8% reads healthy while the two sizes that are half the business are gone.
A customer asks for a nine, hears “not in your size”, and leaves — a lost sale that never appears in any report.
A season read as thirty-two disappointing articles when it is one category that stopped coming back.
Runs counted nightly against sales-weighted thresholds, single pairs moved to fill them, and every box held to a shelf position.
Hole risk by article at 92.7%, pair-with likelihood at 90.4%, 8-week article curve at 91.1% — so the cause is named, not guessed.
“Why, and what do I do?” — fill those runs first, because they are worth more than the new intake.
The same spine,
a different floor.
Footwear & Accessories is sector five of nine in Fashion Retail. The layers, the six products and Iris herself are the same ones running across ten Food & Beverage sectors.
on your shoe wall.