How organisations use data — department by department
What Management, Finance, Sales, Marketing, HR, Product, IT and Operations each need from analysts — and how to work with them.
This lesson builds on: The data analyst's role, and the roles around it.
- List the typical questions and KPIs of each business department
- Explain why the same word (e.g. "revenue", "customer") can mean different things to different departments
- Adapt your communication to the department you serve
Every department is a customer of analytics
Analysts rarely work for "the business" in the abstract. They work with specific people who have specific decisions to make. Understanding what each department is measured on tells you what they will ask, and what will worry them.
| Department | Decisions they make | Typical KPIs | What they need from you |
|---|---|---|---|
| Management / leadership | Strategy, priorities, budgets | Revenue growth, profit, market share, cash | A short, reliable picture and clear trade-offs |
| Finance | Budgets, forecasts, cost control, reporting | Gross margin, EBITDA, opex vs budget, working capital | Numbers that reconcile exactly to the ledger |
| Sales | Targets, territories, incentives, pipeline | Revenue vs plan, conversion rate, deal size, win rate | Rep- and region-level performance, commission accuracy |
| Marketing | Campaigns, channels, budget allocation | Leads, CAC, ROI/ROAS, conversion, retention | Campaign effect measurement, segmentation |
| HR | Hiring, retention, pay, performance | Headcount, attrition, time-to-hire, engagement | Workforce analysis that respects privacy |
| Product | What to build and improve | Activation, usage, funnel conversion, retention | Funnel and experiment analysis |
| IT | Systems, security, data platforms | Uptime, incidents, delivery | Clear data requirements, issue reports |
| Operations | Supply, logistics, service | Stock-outs, lead time, SLA, cost per order | Forecasts, process bottlenecks |
The same word, different meanings
Finance's "revenue" is usually recognised revenue after returns, by accounting period. Sales may count the order value on the day the deal was signed. Marketing may count gross merchandise value including cancelled orders. None of these is wrong — but mixing them in one dashboard is.
Every KPI in a professional dashboard has an agreed definition: formula, filters, source, owner. When two departments disagree, your first job is to surface the definitions, not to choose a side.
Worked example: one dataset, three conversations
Orion Retail's 2025 completed-order net revenue is 1.14M against a sales plan of 1.13M.
- To the CEO: "We grew 12.5% and finished 1% above plan. Istanbul Kadikoy and Ankara (its first full year) drove more than half of the growth. Ankara is still 5% below its plan."
- To Finance: "Net revenue here is order lines × unit price × (1 − discount), completed orders only, by order date. Before this goes into the board pack, I will reconcile it to the ledger — returns booked in a later month are a typical source of difference."
- To Sales: "Store-level plan vs fact by month: four stores finished below plan — Ankara (95%), Tbilisi (97%), Baku Central (97.5%) and Online (98%); Sumqayit led at 111%."
Same numbers. Different emphasis, level of detail and vocabulary.
Working well with each department
- Ask what decision the request supports. "Revenue by region" for a board pack and for a commission calculation require different precision.
- Agree on definitions and sources before building.
- Respect confidentiality. HR and salary data, customer personal data and unreleased financials have restricted audiences.
- Close the loop. After delivering, ask whether the analysis changed the decision; that feedback improves the next request.
Common mistakes
A dashboard that tries to serve the CEO, Finance and every sales rep ends up serving no one.
Practice
Which KPI would you most expect Finance to own?
Sales reports 2025 revenue of 1.21M; Finance reports 1.14M. Sales counts order value on the order date including later-returned orders; Finance counts completed orders net of discount. What should you do?
Summary
- Each department has its own decisions, KPIs and vocabulary.
- The same metric name can have legitimately different definitions; document them.
- Tailor the level of detail and language to the audience, but keep one source of truth.