Starting Point
A multinational beverage and FMCG corporation with sales operations across more than 10 countries faced a reporting problem familiar to every fast-growing enterprise: each country organization, each sales region, and each product category had developed its own Excel reports — with its own definitions, its own formats, and its own data sources.
The symptoms:
- 150+ Excel reports in circulation, at least 30 with contradictory numbers
- Management meetings regularly started with: “Which number is correct?”
- Monthly reports took 5 working days to produce — 3 of those for manual data preparation
- No ability to compare sales data across countries in real time
- Regional managers had no self-service access to their own data
Challenge
Requirements for the new reporting went beyond “prettier dashboards”:
- Single source of truth: One binding definition for every KPI — across all countries
- Near-realtime: Data no older than 24 hours (instead of monthly batches)
- Self-service: Regional managers should create their own analyses without an IT ticket
- Mobile: Field sales reps need dashboards on tablets
- Compliance: Audit trail for all data sources and calculations
Solution
Phase 1: Data Model and Definitions (4 Weeks)
Before a single dashboard was built, we defined and documented 15 core KPIs with the business teams:
- Net revenue: exact calculation formula including returns, bonuses, and currency conversion
- Sales volume: units vs. hectoliters vs. revenue equivalents — one clear hierarchy
- Distribution rate: numeric vs. weighted, with unified calculation per market
- Every KPI received a data owner — one person responsible for definition and quality
Phase 2: Data Integration (6 Weeks)
- ETL pipeline from SAP BW to Azure SQL Database
- Daily refresh with incremental loading (delta load)
- Data quality checks before loading: completeness, plausibility, consistency
- Historical data: 3 years migrated for trend analysis
Phase 3: Dashboard Development (8 Weeks)
Five central dashboards built with Power BI:
| Dashboard | Audience | Refresh |
|---|---|---|
| Executive Summary | C-Level | Daily |
| Regional Performance | Regional Directors | Daily |
| Product Category Analysis | Category Managers | Daily |
| Field Sales Cockpit | Sales Reps | Daily |
| Ad-hoc Analysis | Power Users | Self-service |
Phase 4: Rollout and Training (6 Weeks)
- Train-the-trainer program: 20 power users as multipliers
- Video tutorials for standard functions
- Weekly Q&A sessions for the first 4 weeks
- Parallel operation with Excel for 8 weeks — then mandatory switchover
Results
After 6 months in production:
- Report creation: from 5 working days to 0 — dashboards update automatically
- Data freshness: from monthly to daily (T+1)
- Contradictory reports: from 30+ to 0 — one single truth
- Self-service usage: 85% of regional managers create analyses independently
- Adoption: 420 active users (target was 300)
- ROI: Estimated savings of 1,200 person-hours per month in manual report creation
Lessons Learned
1. Definitions Matter More Than Dashboards
The 4 weeks of KPI definition were the most valuable phase of the project. Without unified definitions, you’re just building prettier versions of the same problem.
2. Don’t Ban Excel — Make It Obsolete
We never banned Excel. Instead, we made Power BI so convenient and fast that Excel became unnecessary. The mandatory switchover after 8 weeks of parallel operation was then just a formality.
3. Data Governance from Day One
Every KPI has an owner, every data source a documented origin, every calculation a traceable formula. Without this governance, Power BI would be the same chaos as Excel after 12 months.