05Predict
Most businesses plan from last year and react to this month.
We start from the decision, not the model. A forecast nobody acts on is reporting with extra steps.
Predict in practice
Most businesses plan from last year and react to this month. The sales forecast is a number someone defends rather than a number anyone trusts, and the cash position is known accurately once a month, usually too late to do much about it.
So we start from the decision, not the model. Which decisions are being made badly or late, what would have to be known to make them better, and whether the data behind that is trustworthy enough to rely on. A forecast nobody acts on is reporting with extra steps.
We have built and governed decision systems where being wrong carries a real cost: trading and market risk, credit and portfolio exposure, demand and capacity in asset-heavy operations. That is also where we learned to distrust a model that cannot show its working.
What we do as owner
We start from the decision, not the model. Which decisions are being made late or badly, and what would have to be known to make them better.
The British Business Bank reports that around half of smaller businesses used or sought external finance in 2025, with increased use of flexible facilities to support cash flow. Cash is the decision most often made with the worst information. British Business Bank, Small Business Finance Markets Report 2026, 17 March 2026
The decisions we improve first7
- Sales and demand forecasting, built from the pipeline rather than last year plus a percentage
- Cash forecasting and liquidity planning, weekly rather than monthly
- Customer renewal, churn and retention signals
- Inventory, capacity and supplier lead times
- Maintenance and asset intervention
- Pricing and margin, where it is holding and where it is eroding
- Labour and capacity planning against demand
What makes a forecast reliable6
- Data quality, lineage and ownership
- One agreed definition of each measure, so the numbers stop being argued about
- Early-warning indicators and exception reporting
- Scenarios and sensitivity for the board
- Back-testing: does the forecast turn out to be right
- Decision records, so the forecast improves from what happened
Putting the forecast in front of management4
- Reporting built around decisions rather than around available data
- Dashboards the management team uses without asking for help
- The forecast in the management meeting, not in a separate pack
- Training so the numbers are understood as well as received
Where we have built forecasting4
- Trading, market and credit exposure, where being wrong has a price
- Demand and capacity in asset-heavy operations
- Portfolio and concentration risk across a group
- Model governance in regulated environments
Not selling? The same capability is available as an executive mandate. See Advisory