🐢 ⚙️ 🧮 📦

Power BI Literacy

Slow Report Causes: Find the Bottleneck Fast

💡 Tap the 💡 icons to see real examples of what slows reports down and how to fix it
🐢

Heavy DAX Measures

  • Iterators (SUMX, FILTER) over large tables can be expensive
  • Complex logic in visuals can run multiple times per interaction
  • Avoid recalculating the same logic in many measures
  • Use variables and pre-aggregation where it improves performance
📦

Model Size and Cardinality

  • High-cardinality columns reduce compression and increase memory
  • Too many columns and unused tables slow down everything
  • Text columns with many unique values are common culprits
  • Prefer keys, remove detail columns, and optimize data types
🔁

Poor Relationships and Filters

  • Many-to-many and bidirectional filtering can add complexity
  • Ambiguous paths can create extra work for the engine
  • Unnecessary relationship activation increases query cost
  • Aim for a clean star schema and predictable filter flow
🖼️

Visual Rendering and Layout

  • Too many visuals on a page increases render time
  • Custom visuals can be slower than native visuals
  • High-density tables and matrices can be expensive
  • Reduce visuals, simplify formatting, and limit data points
🌐

DirectQuery and Source Latency

  • DirectQuery depends on network, gateway, and source performance
  • Complex visuals may generate many backend queries
  • Lack of proper indexes can make queries slow
  • Consider aggregation tables, composite models, or Import mode
🧪

Not Using Performance Tools

  • Guessing wastes time, measure the bottleneck instead
  • Use Performance Analyzer to isolate slow visuals
  • Use DAX Studio to inspect server timings and query plans
  • Test after each change and keep a baseline

Best Practices for Diagnosing Slow Reports

🔍
Start With Performance Analyzer
Capture a refresh of the page and identify which visual or step is slow.
🧮
Rewrite Heavy Measures
Use variables, reduce iterators, and avoid scanning large tables repeatedly.
📉
Reduce High Cardinality Columns
Remove unnecessary text columns, split columns, and use integer keys where possible.
🧱
Keep a Star Schema
Simplify relationships and avoid unnecessary bidirectional filtering.
🧼
Simplify the Page
Reduce visual count, avoid dense tables, and prefer native visuals.
🧪
Validate Each Change
Change one thing, retest, and compare against a known baseline.