The Decision-First Analytics Framework
Most analytics initiatives start with data or technology. This framework shows how to design analytics by starting from the business decision that actually needs to improve.
The Problem
Many analytics initiatives begin with data or technology:
- build a data warehouse
- create dashboards
- run exploratory analysis
The hope is that business value will eventually emerge.
But value rarely appears this way because analytics was never tied to a specific decision.
The Principle
Start with the decision, not the data.
A useful analytics system must answer three questions:
- What decision must improve?
- Who owns the outcome of that decision?
- What information would change the choice being made?
If these questions are unclear, the analytics initiative will likely become an IT project rather than a business capability.
The Framework
The decision-first workflow typically follows five steps:
1. Identify the Decision
Example decisions:
- Which customers are at risk of churn?
- How should prices change next quarter?
- Which suppliers create operational risk?
The decision must be concrete and repeatable.
2. Define the Decision Owner
Every decision has an accountable owner:
- marketing director
- operations manager
- risk officer
They are the real customer of the data product.
3. Map the Information Needed
Identify the signals required to support the decision:
- historical patterns
- predictive indicators
- operational metrics
This stage determines the data requirements.
4. Build the Data Product
Instead of building generic dashboards, build a decision-support system containing:
- metrics
- models
- alerts
- recommended actions
The goal is not information — it is decision support.
5. Measure Decision Quality
Track whether the decision improves outcomes:
- higher revenue
- lower churn
- reduced operational risk
Analytics is valuable only if the decision improves.
Why This Matters
Organizations often invest heavily in data infrastructure but still struggle to produce measurable business value.
The reason is simple:
Data does not create value. Decisions do.
The role of analytics is to improve the quality and speed of those decisions.