AI-poweredDemand Forecastingfused withConsumer Insights
Stock allocation, markdown timing, the new season buy, per SKU, size and store
TDI tells you how many units to move now, when to mark down, and how many to order for next season, so you can sell more. All three products run on the same demand estimates, which means you stock what will sell, not what sold, for every SKU and size in every store. TDI collects point-of-sale data and returns automated directives into your ERP, WMS or any other open system you run, this store, this SKU, this size, this quantity, ready to action.
TDI Allocate sets how many units of each SKU and size each store should hold. A single adjustment can drift out of balance within a few weeks, so TDI corrects it in time, throughout the season.
TDI Markdown tracks every SKU and size in every store against expected demand. Before sales fall behind, it tells you when to reduce the price and by how much, without giving away margin too early.
TDI Order plans how many units of each SKU and size each store will need next season. Added up, it is the total order. Broken out, it is exactly what each store should receive, and when across the season.
Two days. We measure the potential upside and plan how many of each SKU size every store needs, and when.
You reallocate. SKU sizes move from the stores where they sit unsold into the stores that are short of them.
Ten days. We report the actual increase in units sold, and its financial impact, in absolute units against a baseline.
One sales extract. No system access, no integration, no personal data.
Talk to usA customer finds the SKU she wants. Her size is not there. Your staff go looking, in the stockroom, in another store, online. Often she buys nothing anyway. None of it reaches your data, because sales show what was sold, not what was wanted, so the next order repeats last time's quantity and the same size runs out again.
A US specialty apparel retailer: 4 SKUs across sizes 0 to 5, 23 stores, one 21-day selling cycle. We identified which sizes each store was missing. The retailer supplied them in full from elsewhere in its network of around 2,000 stores. The units sold. No additional buying was required.
On the SKUs in the pilot. Where you land depends mostly on how much of the right stock is available elsewhere in your network to move.
It records every size you sold. It cannot record the customer who wanted a 12, found none, and left. That is the number your planners have never had.
The units already exist somewhere in your network. The whole result comes from putting them where the demand is.
Three things separate this from a forecasting tool, and none of them is about better mathematics.
A sales record can only show what was available to buy. We rebuild the demand that was never recorded, because by then the size had gone.
You can forecast a style perfectly and still send the wrong sizes to the wrong stores. Every number we produce is for one SKU, in one size, in one store.
Instead of a single prediction to hit or miss, you get the range of demand that could occur, and you decide how much of it your stock should satisfy.
Four steps, and only two of them are yours. There is no system to replace, no integration to build and no new team to hire. What arrives is a list, this store, this SKU, these sizes, these quantities. After the pilot the same four steps run automatically, into your ERP, WMS or any other open system.
SKU, store, date and units sold, at the most granular level you hold, with size as a separate field if your SKU code does not carry it. One file. No system access, no integration, no personal data, no inventory file.
We reconcile the file against totals you already know, then work out, for every SKU, store and size, what would have sold if the size had been on the shelf, not only what did sell while it was there.
How many of each SKU size every store needs and when, as a plain list ready to hand to your allocation team, with the potential upside measured on your own data. If the data will not support a commitment, we say so before you spend anything.
You supply the sizes from wherever you hold them. We keep measuring units sold against the baseline, season after season.
This runs alongside whatever you already use. If your team already produces a forecast, we can build on theirs rather than replace it.
Every recommendation shows the demand estimate behind it, including the stores and sizes where we found nothing wrong.
Your planners keep the decisions. What changes is that they are working from demand rather than from a sales record that never saw it.
Every term is agreed in writing before the pilot begins, so there is nothing left to negotiate afterwards. One number decides it, and you owe nothing until it is met.
Two business days on your own sales extract. What comes back is the potential upside and the plan, how many of each SKU size every store needs and when.
No charge, and no obligation to go further.We agree the SKUs and stores, the baseline, and how the result will be judged. Absolute units for each SKU, store and size, against comparable SKUs in the same stores. Any size you cannot supply is excluded, recorded at the time.
Agreed in writing before anything starts.The trigger, the commercial terms and the term go into one document, signed before you move a single unit. Converting on at least 10% more units sold on the pilot SKUs, set well below what we expect.
Signed, not invoiced.You reallocate, moving sizes from the stores where they sit unsold into the stores that are short of them. We measure and report the actual increase in units sold, and its financial impact.
No new inventory. Out-of-stock events reported alongside.The pilot runs on a few SKUs in a few stores, from one extract, without touching your systems. Converting it turns that into an operating process.
Units on the pilot combinations rise by at least 10% and the conditional purchase order converts automatically. The terms were agreed before the pilot began, so there is nothing left to negotiate.
What the pilot did for a handful of SKUs runs across every store and category you choose, and TDI Markdown and TDI Order run alongside it whenever you want them, so allocation, discounting and the season order all work from the same estimates.
Point-of-sale data reaches TDI on a schedule, and the directives return automatically into your ERP, WMS or any other open system, ready to action. No file handling, no manual step, and no new system for your planners to learn.
