Match buying decisions
Choose SKU, site and weekly units that fit ordering. Separate sparse and new items.
RETAIL & COMMERCE / CAPABILITY
Turn forecasts
into buying decisions.
Evaluate forecasts using sales, stockouts, promotions and lead times. Show uncertainty and retain the reasons for buyer adjustments.
Explore the delivery scope ↓OPERATING MODELILLUSTRATIVE
- 01POS, e-commerce & stock
- 02Unify products and orders
- 03Review margin and demand
- 04Decide buying and promotion
Conceptual data-to-operation flow, adapted to the agreed project scope.
What makes
this workflow difficult.
Zero sales can indicate no demand or unavailable stock. Distinguish availability from demand before training on sales counts.
THE SHORT ANSWER
Evaluate forecasts using sales, stockouts, promotions and lead times. Show uncertainty and retain the reasons for buyer adjustments.
DESIGN & DELIVERY
Design in the order the work happens.
Establish baselines
Compare against last week, seasonal patterns or moving averages before choosing complexity.
Connect ordering constraints
Include lead time, pack size, minimum orders and storage limits, with agreed stockout and overstock tradeoffs.
DATA READINESS
Required data and the conditions to check.
A consistent format does not ensure operational usefulness. Confirm meaning, refresh and exceptions with owners, and keep missing information explicit.
| Data | Required fields | Checks |
|---|---|---|
| Sales history | Date, SKU and quantity | Closures, stockouts and discontinued items |
| Promotion | Price, campaigns and events | Whether information was known at forecast time |
| Procurement | Lead times and minimums | Delays and partial deliveries |
EVALUATION
Measure value in the workflow.
These are candidate evaluation measures. Establish a baseline and agree on data, period, targets and exclusions before work starts. They are not claimed performance results.
| Measure | Evaluation method |
|---|---|
| Forecast error | Compare with baseline by category and volume |
| Stockout and excess | Evaluate under equal replenishment assumptions and explicit costs |
| Adjustment burden | Track buyer changes and their reasons |
Begin with recommendations approved by buyers. Seasonal evaluation requires sufficient historical coverage.
Define deliverables
and acceptance.
Review data mappings, workflow, prototype and evaluation evidence before choosing production scope. Estimates distinguish assumptions because permissions and interface terms affect implementation.
Select one category and channel for reporting or weekly forecasting.
QUESTIONS BEFORE YOU START
Questions before starting
Can Shopify and store data be combined?
We assess product identifiers, returns and available fields; CSV exports can be a starting point.
Can sparse items be forecast?
Limited observations make evaluation unstable. Compare category-level forecasts, replenishment rules and buyer judgement.
RELATED FIELD GUIDE
RETAIL & COMMERCE / GUIDE
Before retail demand forecasting: stockouts, returns and promotions
Sales history alone does not make a forecast useful for buying. Distinguish why sales were absent, what was known at prediction time and the constraints of ordering.
Read the guide ↗YOUR OPERATION / OUR STARTING POINT
Start with the work you do today.
Tell us about the workflow, systems and reporting or documents that need attention. We will define a suitable phase, deliverables and assumptions.
Retail and e-commerce — discuss your project How scope and estimates work ↗