Where will demand move?
Forecast at store, category, product or regional level so stock, staffing and planning decisions happen where the business operates. See what happens after the forecast.
← All Decision Packs
GN Retail Decision Pack · Retail Decision Intelligence
Retail decisions, packaged around the actions teams make every week.
For merchandising and inventory teams deciding what each store should carry and where stock should go. Use first-party transaction, customer, product and store data to decide where demand is moving, which customers matter most, and where stock, targeting or store-level action should change.
Decisions this pack supports
Start with the choice, not the model.
Which customers are worth prioritising?
Use CLV, churn or repeat-purchase signals to focus retention and acquisition effort where expected value is highest.
Which SKUs should each store or channel carry?
Balance local demand, substitution, cannibalisation, margin and operating constraints instead of ranking products on sales alone. See assortment optimisation.
Who or what should receive the next action?
Combine propensity, value and operational constraints to prioritise targeting, offers, inventory or store interventions.
From today's process to the next decision
Choose the range and allocation each store needs.
How teams make the decision
Teams combine sales reports, demand forecasts and spreadsheet rules to choose ranges and allocate stock across stores.
What this leaves unresolved
A sales ranking or forecast alone does not resolve substitution, margin, shelf space or limited stock. The resulting plan can leave excess inventory in one store and missed sales in another.
What the GN Pack adds
Compare assortment and allocation choices using local demand, product economics, substitution and your operating constraints, then recommend a feasible store-level move.
What your team receives
A recommended SKU range or stock allocation by store or channel, with the expected outcome, supporting evidence and constraints used.
See the pack in action
From your business data to a decision you can act on.
Explore what a director or manager could receive. Choose a model, see its output, then compare practical actions with costs and constraints.
Illustrative interactive demoAll figures and businesses are fictional. These are examples of potential deliverables, not actual customer results, guarantees or live predictions.
RETAIL / EXAMPLE DECISION WORKSPACE
What should your stores do next?
See how forecasts, customer value and commercial rules turn into specific store-level moves.
01 / Explore a model output
EXAMPLE OUTPUT 01 / 05
27 stores
At risk of stockouts
What we found
Fourteen high-demand products may run short within seven days. Estimated unmet demand: 1,850 units.
Recommended next move →
Review replenishment and available stock by location.
Decision-relevant output1,850 units of demand at risk
02 / Compare possible decisions
How should we allocate 420 available units?
Switch between the options and compare expected contribution after intervention costs. Numbers use fixed fictional assumptions.
Gross incremental contribution+€13,900
Intervention cost€1,500
Net incremental contribution+€12,400
Selected action
Recommended redistribution
Move available units to 11 stores with the highest expected unmet demand.
Expected recovered contribution €13,900 less €1,500 transfer and handling costs.
Scope: 420 unitsFeasibility: Within stated scenario rules
Operating constraints considered
See what this pack could recommend for your business →- 420 units available
- Transfer, shelf-space and minimum display-stock limits
- No transfer where logistics costs exceed expected recovered margin
Read every example output and decision alternative
Demand Forecasting: 27 stores — At risk of stockouts
Fourteen high-demand products may run short within seven days. Estimated unmet demand: 1,850 units. Review replenishment and available stock by location. 1,850 units of demand at risk.
Customer Value + Propensity: 310 customers — High-value customers with declining visits
These customers have historically contributed more, but return likelihood has fallen. Prioritise investigation and test a proportionate retention treatment. 310 customers to assess.
Assortment Optimisation: 12 SKUs — Candidates for replacement
Twelve slow-moving items consume space that could support eight locally relevant products. Test a revised assortment in matched stores before wider rollout. +4.2% modelled category contribution.
Promotion + Price Effect: 12 stores — Where a smaller promotion may work
An earlier broad discount drove sales but did not cover the estimated margin given away. Test a 10% promotion only in qualifying stores. Avoid unprofitable blanket discounts.
Decision Optimisation: 420 units — Available for reallocation
Eleven stores face expensive stockouts while other locations have surplus stock. Transfer stock within store, logistics and display-stock limits. +€12,400 modelled net contribution.
Decision options
Keep inventory in its current stores. Store-level demand gaps remain; no transfer cost. Scope: 0 units.
Move available units to 11 stores with the highest expected unmet demand. Expected recovered contribution €13,900 less €1,500 transfer and handling costs. Scope: 420 units.
Split the available units across 21 stores. Simple allocation, but less stock reaches the locations with the highest opportunity. Scope: 420 units.
Models inside this pack
One product bundle around a recurring business decision.
GN Retail combines the models, business logic and constraints needed to move from your operating data to a decision-ready recommendation.
DECISION PACKGN Retail
Built aroundTransactions · customers · products · stores
Demand ForecastingForecasts demand at store, category, product or regional level with uncertainty where useful.
Customer Value + PropensityEstimates customer value, repeat-purchase likelihood, churn or response propensity.
Assortment OptimisationCombines local demand, product value, substitution, cannibalisation and operating constraints.
Promotion + Price EffectSeparates observed lift from credible incremental effect and models response to commercial interventions.
Decision OptimisationCombines predictions, expected effects and real constraints into a feasible ranked action.
OUTPUTDecision support your team can use
Models are combined only where the decision needs them. The pack returns a business-ready action with supporting evidence and expected impact.
- Stock
- assortment
- targeting
- store action
- Expected impact and supporting evidence
- Configured around your real operating constraints
Demand Forecasting
Forecasts demand at store, category, product or regional level with uncertainty where useful.
What to stock, where to allocate it and when plans should change.
Customer Value + Propensity
Estimates customer value, repeat-purchase likelihood, churn or response propensity.
Which customers deserve retention, acquisition or campaign priority.
Assortment Optimisation
Combines local demand, product value, substitution, cannibalisation and operating constraints.
Which SKUs each store, channel or customer set should carry.
Promotion + Price Effect
Separates observed lift from credible incremental effect and models response to commercial interventions.
Which promotion or pricing action is worth repeating, changing or stopping.
Decision Optimisation
Combines predictions, expected effects and real constraints into a feasible ranked action.
What stock, targeting, assortment or store-level move should happen next.
Pack anatomy
One repeatable path from retail data to action.
Prebuilt where the decision pattern repeats. Configured where your business is different.
Your operating data
Transactions, customer history, products, stores, campaigns, inventory, pricing and relevant external drivers.
Predictive, causal and optimisation models
Use the modelling layer needed for the question rather than forcing every use case through the same algorithm.
Your objectives and constraints
Encode the business objective, available interventions, budgets, capacity or policy rules that make the recommendation feasible.
A ranked next move
Return a forecast, prioritised audience, store action, intervention or recommended allocation with evidence and expected impact.
Prebuilt + configured
Reusable decision IP without pretending every retailer is identical.
Graphite Note reuses decision workflows, modelling patterns, evaluation methods and output structures. We configure the objective, constraints, data mapping and intervention set around how your retail business actually works.
Decision workflow
Model patterns, evaluation, governance and output structure.
Business logic
Objectives, interventions, constraints, thresholds and data mapping.
First-party context
Your customers, products, stores, economics and operating knowledge.
Direct answers
Questions about the GN Retail Decision Pack.
What is the GN Retail Decision Pack?
It is a packaged Decision Intelligence solution for recurring retail decisions across demand, customer value, targeting and store-level planning. It combines reusable decision workflows with the retailer's own first-party data, objectives and constraints.
What data can the pack use?
Typical inputs include transactions, customer history, products, stores, campaigns, inventory, pricing and relevant external drivers. The exact dataset depends on the decision being made.
What models sit underneath?
Depending on the problem, Graphite Note can use predictive machine learning, AutoML, causal methods and optimisation. The pack is organised around the decision, not around a single model type.
What comes out of the pack?
The target output is a business-ready decision such as a forecast, prioritised customer or store list, recommended intervention or allocation, together with the evidence and expected impact needed to act.
Decision IntelligenceFrom signal and prediction to a governed business action.AutoMLAutomate repetitive model-building work without losing business context.Causal Machine LearningSeparate drivers from correlations and estimate intervention effects.Predictive AnalyticsForecast demand, score customers and anticipate business outcomes.Decision OptimisationCompare feasible actions and rank them by expected business impact.Forecasting vs OptimisationSee where prediction ends and the business decision begins.Decision APIDeliver recommendations, expected impact and evidence into workflows.Embedded Decision IntelligenceAdd a decision layer to ERP, CRM, BI, workflow and vertical software.Build vs BuyA practical framework for deciding whether to build, buy or embed.
Bring one recurring retail decision.
We will scope the data, models, constraints and action around the outcome you need to move.