Service
Decision Modelling & Predictive Analytics
A model should carry the repeatable calculation and leave the judgement visible. We build decision tools that are compared with simple alternatives and honest about what they cannot know.
What you get
- Sales and stock forecasting — predict demand, optimise inventory, protect cashflow
- Net revenue management — optimal pricing, promotions and portfolio decisions
- Stock decision management — consolidate multivariate data into actions your team can take today
- Commission and incentive scheme modelling, simulated against your real transaction history
- Causal impact measurement — know what an intervention actually changed
- Python
- pandas
- SQL
- Statistical modelling
Forecasting and stock decisions
We build forecasts at the level people buy and plan at, then compare them with simple rules before claiming an improvement. When the signal is useful, the team holds less unnecessary stock, misses fewer sales and can put cash to better use.
Pricing and net revenue management
Price, promotion and range decisions interact across a product’s life. We make those trade-offs easier to inspect so the people responsible can decide deliberately rather than reconstruct the same argument each time.
Incentives that pay for the behaviour you want
Commission schemes drift. We test candidate designs against the transactions the business actually made, so sales and finance can see the cost, the winners, the losers and the behaviour each design rewards before committing to it.
Measuring what actually changed
When you open a store, change a price, or launch a channel, the honest question is “what would have happened anyway?” We estimate the effect by comparing what happened against a matched group that didn’t get the change — and report the assumptions it rests on, the uncertainty around it, and what the comparison genuinely cannot tell apart. An estimate with its limits stated is worth more than a confident number that quietly depends on them.