Floemdata
Data & ETL

What is data analytics? Four levels

Descriptive, diagnostic, predictive and prescriptive analytics explained — and the order that actually works for mid-sized companies starting out.

Floemdata · · 6 min read

Data analytics is the work of turning records into decisions. Reporting answers 'what happened'; analytics continues into 'why' and 'what should we do'.

Four levels

  • Descriptive: what did we sell last month?
  • Diagnostic: why did margin drop three points?
  • Predictive: when will this customer pay?
  • Prescriptive: how much of this product should we build?

The order matters

Prediction built on unreconciled data hides its own errors. In practice the first three steps are: bring the data together, fix each metric to a single definition, and reconcile the result against the source system.

Questions that pay off first

  • Which product group earns margin rather than just revenue?
  • How much receivable is overdue, and with which customers?
  • Which stock has not moved for a year, and how much cash is tied up?
  • How late do supplier deliveries arrive, and what does that delay?

All of these are answerable from data that already exists in the ERP. What is usually missing is not data but the layer that brings it together.

← All guides