Chapter 1
From data to impact
Data is not the goal. Each step up the pyramid adds value to the one below, and answers a bigger question. This chapter shows the steps, and the four levels of analytics maturity that grow with them.
The DI-KIWI model
Six steps from raw data to impact
The first half explains the past of the business. The second half shapes its future.
- D
Data
The raw material: facts and events as they are captured.
- To do
- Capture and store it reliably.
- Goal
- A trustworthy base to build on.
- I
Information
What?
Given context, data becomes information.
- To do
- Contextualise data and make it useful: dashboarding and reporting.
- Goal
- Build a source that answers questions.
- K
Knowledge
How?
Given meaning, information becomes knowledge.
- To do
- Understand and compare the information with expected or reference values.
- Goal
- Find the gaps, deltas and gains, and develop a hypothesis.
- I
Insights
Why?
After synthesis, knowledge becomes insight.
- To do
- Compare and understand the information, deep-dive and synthesise.
- Goal
- Analyse and test hypotheses to support claims and decisions.
- W
Wisdom
With explanations, insight becomes wisdom.
- To do
- Discuss with teams and explain with an estimated impact.
- Goal
- Reveal the direction: what is best.
- I
Impact & observation
Given purpose, wisdom becomes a decision, and the decision has an impact to observe.
- To do
- Act, then measure what changed.
- Goal
- Close the loop and feed the next round.
Analytics maturity
Four levels, each with a bigger payoff
As analytics moves up, the degree of transformation and the business impact both grow. Start at the bottom, and climb deliberately.
WHAT IF
Discovery / wisdom analytics
What-if analyses that find new possibilities.
| Method | Question it answers |
|---|---|
| New product, service or experience innovation | How can we create or discover new products and services, or make existing ones more efficient? |
| Meta knowledge | How can we apply knowledge about knowledge? |
SHOULD
Prescriptive analytics
What should the optimal outcome be?
| Method | Question it answers |
|---|---|
| Optimisation | How can we achieve the best outcome? |
| Decision-making under uncertainty | How can we decide with incomplete information? |
| Impact analysis | What action should be taken, and what would its impact be? |
COULD
Predictive analytics
Make an informed forecast.
| Method | Question it answers |
|---|---|
| Predictive analysis | What is likely to happen? |
| Forecasting | What trends are foreseen? |
| Simulation | What other scenarios or alternatives are there? |
IS
Descriptive analytics
Understand past and current business.
| Method | Question it answers |
|---|---|
| Query / drill-down | Where exactly is the problem? |
| Routine or ad hoc reports | What happened, how many, when, how often, where? |
| Dashboards | What alerts can be identified? |
| Visualisation / charts | How can we present the data? |
The original slides
See the two diagrams
Building or scaling an analytics function?
Let’s talk about building your team, making business metrics reliable or improving access to analytics.