Data and forecasting

Give decision-makers information they can use in time.

A dashboard can explain the past. An alert can highlight a threshold. A forecast can estimate what may happen next. We choose the right approach for the decision and the quality of available data.

Typical flow from data to decision
  1. QuestionWhat needs deciding
  2. DataOperational data
  3. CheckSource and update time
  4. SignalDashboard, alert or estimate
  5. DecisionManagement
  6. OutcomeAction taken early

On this page

What it solves

Three tools for three situations.

We choose the right approach for the decision and the quality of available data.

  1. DashboardExplains what happened.
  2. AlertHighlights a crossed threshold.
  3. ForecastEstimates what may happen next. Only where the data can support it.

How it works

Begin with the decision.

First we clarify which decision is to be supported. That determines the data, indicators and update frequency.

We separate facts from estimates.

We do not label every chart a prediction.

Fact

Current figures show their source and update time.

  • Source
  • Update time

Estimate

Forecasts show assumptions, prediction period and uncertainty.

  • Assumptions
  • Prediction period
  • Uncertainty
Management team reviewing a process map

Management decides

Management decides, not the model.

Every estimate shows its inputs and limitations, so it is clear what it rests on.

  • Alerts arrive before a decision becomes urgent.
  • Each estimate shows what it is based on.
  • Management decides how to respond.
  1. Operational data
  2. Risk alert
  3. Management decides
  4. Action taken early

systemautomationa person decidesoutcome

How we validate

We check whether the estimate is useful.

We compare estimates with what actually happened. We use a complex model only where it is reliable enough for the decision.