Client · 2021–2025 · Self-service BI
Three dashboards. Three different revenue figures. Every single month.
Finance couldn't explain which number was right. Board reporting got delayed. Strategic decisions made on figures nobody was sure of. 20,000 dashboards. Not a single source of truth.
How a data landscape works, and what went wrong
Hover over each step to see what it contains, what it delivers and what went wrong at that point.
01
Sources
→
02
Ingestion
→
03
Data Lake
→
04
Processing
→
05
Modelling
→
06
Data Warehouse
→
08
Delivery
→
09
Users
Freedom for business, control for IT
Hover over a step
How a data landscape works
Move your mouse over an icon above to see what happens in that step, what it delivers and, more importantly, where it went wrong before the governance layer was introduced.
The key: business and IT on the same foundation
Before
Parallel worlds
IT builds the pipeline
↓
Business builds around it
↓
Diverging definitions
↓
Argument over the figures
↓
Board reporting delayed
No owner of the truth. Same conversation every month.
After DTS role
One environment, one review
150 business users + 6 IT engineers working in the same staging environment
↓
Business proposes a change
↓
Automated code review
↓
Approval by IT
↓
Promoted to production
One definition. Always. No parallel truth possible anymore.
20,000 → 200
Certified dashboards
5,000
Users on one source of truth
150 + 6
Business and IT, same environment
Stack
Power BI
Snowflake
dbt
GitHub
Approach
Data Mesh
Medallion architecture
Self-service governance
Code review for business
OTAP deployment
Data Contract
Row Level Security
Column Level Security