Why Most Analytics Projects Fail to Change Decisions

## The gap between insight and action

A large proportion of analytics investment produces dashboards that are looked at once, reports that are filed unread, and models whose outputs are never incorporated into the decisions they were meant to inform.

This is not primarily a technical problem. It is a design problem. The relationship between the analytical output and the decision it is meant to serve has not been designed — it has been assumed.

## The three common failure modes

**Output-first design**: The analyst builds the analysis and then asks who might use it. This produces outputs that are technically correct but functionally disconnected from any real decision-making process.

**Metric proliferation**: The organisation measures everything and therefore acts on nothing. When every metric is on the dashboard, no metric has primacy. Decision-makers faced with 40 numbers default to the three they already understood before the dashboard existed.

**The interpretation gap**: Analysis is handed over without a structured interpretation. The decision-maker reads the chart, draws their own conclusion, and acts on it — which may or may not be what the data actually supports.

## The data-to-decision design principle

Effective decision intelligence starts with the decision, not the data. The sequence is:

1. Define the specific decision that needs to be made
2. Identify what information would change that decision if it were different from what you currently believe
3. Identify the data required to generate that information
4. Build the analysis to produce only that information
5. Structure the output as a decision recommendation, not a data summary

This sounds simple. It is not. It requires the analyst and the decision-maker to have a genuine conversation about the decision — which organisations often avoid because it forces clarity about who is responsible for what.

## What this means in practice

For African market contexts, where data is often thinner and less reliable than in mature markets, this design principle is even more important. When data is limited, every analytical choice matters more. Starting with the decision helps prioritise which data gaps matter and which can be managed with explicit assumptions.

The Herufi Data-to-Decision Workflow framework (available in the Frameworks section) provides a structured six-step process for applying these principles to real decisions.