Publish Date :
What Data You Need to Start a Planning POC
Start a planning POC with an agreed merchandise hierarchy, dated sales and stock, open commitments, fiscal calendar and selected budget assumptions. Core can support the evaluation only when those inputs mean the same thing to the retailer and the setup team. A larger export does not compensate for unclear definitions.

THE MERCHMIX DECISION / Core
Start with identifiers the files can share
A planning POC needs a small, defined data pack whose product keys, periods, stock states, commitments and budget basis can be reconciled.
Independent retail planning example. Financial amounts are US dollars. Forecasts, quantities and proposals are assumptions for this decision, not customer results.
Input pack / hierarchy, trading and stock
| Input family | Required meaning |
|---|---|
| Hierarchy | SKU, style, size, store and ownership |
| Trading | Period and net-sales basis |
| Stock and orders | Snapshot and commitment status |
The counts describe an example input pack, not a customer dataset.
An independent input pack has one hundred SKU records, with seven still unmapped.
THE KEY IDEA
Agree the keys, definitions, timestamps and exceptions before judging the planning result produced from the data.
Agree the decision scope before requesting every file
Choose the categories, stores and planning window required for the first test. That scope determines which history and stock records matter. Collecting all available tables before agreeing a decision can create a long reconciliation exercise without a clear evaluation objective.
Merchmix’s published POC accepts anonymised style, sales and stock information, including spreadsheet exports or a read-only view. The required fields and setup depend on the agreed test. Core does not need an invented universal file template imposed on every retailer regardless of its merchandise model.
Make the merchandise hierarchy a shared reference
Define the SKU, style, colour, size, category and store identifiers, including how discontinued or inactive records are represented. The same keys must join across sales, stock and commitments. A product description alone is an unreliable join when names change or similar options share wording.
In an independent input pack, 93 of one hundred SKU records map and seven require review. Keep those seven explicit. Resolve them or agree an exclusion with the owner before evaluating totals, rather than quietly dropping the rows and presenting the smaller result as the whole business.
Give the sales file a period and return convention
The example trading period records 125 gross sold units and five returned units, giving 120 net units. Confirm whether the planning measure is gross or net and how return dates are attributed. Comparable history needs consistent fiscal weeks, timezone and product scope.
Sales value also needs a defined basis for tax, discounts and currency. Keep the underlying distinction available to the Core evaluation so the team can explain an apparent demand change. A file that imports successfully can still compare different commercial measures if those conventions were never agreed.
THE FIGURES BEHIND THE DECISION
| Owner | Decision required |
|---|---|
| Merchandise owner | Resolve or explicitly exclude the seven keys |
| Trading owner | Confirm net units and periods |
| Stock owner | Confirm thirty usable and twenty-four incoming |
Sales check: 125 less 5 = 120 net units. Separate stock check: 40 physical less 8 reserved less 2 quality held = 30 usable. Twenty-four incoming units remain future supply.
Reconcile physical, usable and future stock separately
A separate stock snapshot contains forty physical units, including eight reserved and two quality held. With those restricted groups distinct, thirty are usable at the snapshot. Twenty-four incoming units belong to a future supply record and are not added to today’s usable amount.
The stock snapshot is not derived from the sales example by adding or subtracting the same period again. Agree its timestamp and which events are already included. Core’s forward comparison needs a credible opening position and order state, rather than a mixture of physical stock and expected arrivals.
Bring the actual commitments and selected budget version
For incoming lines, retain the supplier/order reference, SKU, open quantity, status and date meaning. Distinguish an original order from the amount still open after receipts. A proposed cancellation is not a confirmed reduction, and an acknowledgement does not establish physical receipt.
Choose the budget version and fiscal calendar that the test should use, including the cost and retail-value bases. Core supports budget and MFP/WSSI comparisons, but the setup team needs to know which accepted assumptions the retailer expects to reproduce. A target with no version or scope is difficult to reconcile.
Make exceptions part of the readiness decision
The seven unmapped SKUs, the return convention and the split between thirty usable and twenty-four incoming are concrete owner questions. Record who can resolve each one and whether it blocks the chosen test. Some exceptions can sit outside a narrow POC if that boundary is agreed.
Use anonymisation and access arrangements appropriate to the evaluation. Avoid sending customer-identifying information that the selected planning test does not require. The retailer and setup team should agree the actual data-handling process rather than infer security or retention terms from a generic article.
| Signal | Merchmix workspace | Decision to review |
|---|---|---|
| Merchandise keys | Core / evidence | An independent input pack has one hundred SKU records, with seven still unmapped. |
| Sales basis | Core / comparison | 125 gross sold units less five returned units gives 120 net units. |
| Readiness gate | Core / decision review | Core needs a defined input scope; a clean import cannot settle every business definition. |
Start the POC when the pack can be explained
The first baseline should reconcile to the agreed source pack, including its exclusions. Retain a short data dictionary and the reason for each accepted adjustment. That makes later changes in the Core comparison understandable to the people evaluating it.
The goal is a small, relevant and reproducible starting point. Once the input basis is accepted, the team can spend the POC testing planning decisions rather than discovering that different files describe different products, periods or stock states.
BEFORE THE NEXT DECISION
- Planning owner
This is a sales-period reconciliation, separate from the stock snapshot below.
- Commercial owner
The incoming quantity must retain its own status and expected date.
- Decision owner
Core needs a defined input scope; a clean import cannot settle every business definition.
Common questions
Do we need to export every system before starting?
No. Agree the first decision and merchandise scope, then assemble the relevant reconciled inputs.
What happens to the seven unmapped SKUs?
They need resolution or an explicitly agreed exclusion. They should not disappear silently from the evaluation totals.
Why are net sales 120 units?
The example has 125 gross sold units and five returned units on the selected period basis.
Why is usable stock thirty rather than forty?
Eight reserved and two quality-held units are excluded from forty physical units, with no overlap between those restrictions.
Are the twenty-four incoming units available now?
No. They retain their own commitment status and date and remain separate from current usable stock.
FROM THE ARTICLE TO YOUR OWN RETAIL PLAN
Review the decision.
See the workflow.
Bring your planning questions to a demo, or explore a proof of concept on your own data.