Great analysts follow a repeatable process rather than diving straight into charts. This framework (used by Google's analytics programme) keeps your work rigorous, useful and defensible.
1. Ask — define the question
Start with the business question, not the data. What decision needs to be made? Who's the stakeholder? What would a useful answer look like? A vague brief ("look at sales") produces vague analysis. Turn it into a sharp question: "Which regions and products drove the drop in Q3 revenue?" Agree success criteria up front.
2. Prepare — get and clean the data
Identify the data you need and its source. Then do the unglamorous but vital work: clean and validate it — handle duplicates, missing values, inconsistent formats and outliers (see Data Fundamentals for quality dimensions). Document every assumption and transformation so your work is reproducible.
3. Analyse — find the story
Explore the data to answer the question:
- Aggregate (totals, averages, counts by group) — often with SQL or PivotTables.
- Compare across time, segments or categories.
- Look for patterns, trends and outliers — and why they occur. Be curious but disciplined: test whether what you see is real or noise, and beware correlation vs causation.
4. Share — communicate clearly
An insight nobody understands has zero value. Lead with the answer (the headline), support it with the right visual (see Visualisation Principles), and tailor depth to the audience — an executive wants the "so what", an analyst wants the detail. Tell a story, not a data dump.
5. Act — drive the decision
Finish with clear, specific recommendations tied to the original question, plus the confidence and caveats. The best analysts don't just report what happened — they suggest what to do about it and follow up on the outcome.
Put it to work
Take a real question at work or study, run it through all five steps, and practise turning findings into recommendations in the Sales Analysis lab.