1. The data model is wrong
Nine times out of ten, the root cause of "Power BI is slow / numbers don't match / totals are wrong" is a flat, one-big-table model. A proper star schema — clean fact and dimension tables, explicit relationships, a single date table — fixes the performance, the DAX and the accuracy in one pass.
2. The DAX doesn't match the business
Measures written on assumptions nobody confirmed: "sales" that include cancelled orders, "customers" that include duplicates, "margin" that forgets rebates. We rewrite measures alongside finance and ops so the dashboard matches the P&L.
3. Refresh is flaky
Gateway timeouts, source API throttling, dataflows silently failing. A Power BI consultant fixes it by moving transformation upstream into a proper warehouse — see ETL pipelines & data warehousing — so Power BI imports clean, modelled data instead of doing heavy lifting at refresh time.
4. Row-level security is missing or wrong
Every client team you show a dashboard should only see their numbers. We set up RLS properly, with tested role coverage — not "we'll add it later".
5. Nobody uses the reports
Often because the reports answer the wrong questions. Good Power BI work starts with "what decision does this help someone make on Monday morning" — not "which visuals do we have".
What changes after a Power BI consulting engagement
- Dashboards open in seconds, not minutes.
- Totals reconcile to the finance ledger.
- Refresh runs green every day.
- Business users trust the numbers enough to act on them.
For a worked example of the full stack — warehouse, ETL and Power BI — see the BI Sense case study.
PBIX file giving you grief?
Send it over. We'll tell you honestly whether it needs tuning or a proper model rebuild.