How to Use Performance Analyzer in Power BI (and What “Other” Really Means)
A repeatable way to find the slow visual on a report page, read each timing correctly, and decide whether the fix belongs in DAX, the model or the report design.
Core Engineering Takeaways
- •Performance Analyzer splits each visual's load time into DAX query, Direct query, Visual display, Other and (for field parameters) Evaluated parameters.
- •A high DAX query time points at the measure or model. A high Visual display time points at the visual itself.
- •A high Other time usually means the visual was waiting for other visuals — too many visuals on one page.
- •Copy the slow visual's query into DAX Studio to see whether the time goes to the storage engine or the formula engine.
1. What Performance Analyzer measures
Performance Analyzer, on the Optimize ribbon in Power BI Desktop, records how long each visual takes to load and splits the time into categories. Microsoft defines them as:
- DAX query: time from the visual sending its query to the model returning results.
- Direct query: for DirectQuery tables, time for the external source to answer.
- Visual display: time to draw the visual, including fetching web images or geocoding.
- Other: time spent preparing queries, waiting for other visuals to finish, or doing background work.
- Evaluated parameters (preview): time spent evaluating field parameters in the visual.
The total for a visual is the sum. The skill is in knowing which number to act on.
2. Record a clean, repeatable test
Measurements are only useful if they are repeatable. The model caches query results, so the second run of a page is always faster than what a user sees first thing in the morning.
- Open the report and go to a blank page so nothing loads while you start.
- Open Performance Analyzer and select Start recording.
- Switch to the page you want to test, or select Refresh visuals to reload every visual on the current page.
- To test a single visual, use Analyze this visual in its top-right corner while recording.
- Select Stop, then Export if you want to keep or compare the results (it saves a JSON file).
For a true cold-cache figure, connect DAX Studio to the open Desktop model and clear the cache before refreshing visuals — otherwise you are timing cached results.
3. Reading the results: which number to act on
High DAX query time
The query is slow. The cause is in a measure or in the model: iterators over large tables, context transitions inside them, bidirectional relationships, or high-cardinality columns. Take the query to DAX Studio (next section) — and see why this DAX measure is slow for the usual patterns.
High Direct query time
The source database is slow to answer. Look at the SQL it receives, indexes on the source, and whether the table should be Import or Direct Lake instead.
High Visual display time
The visual itself is expensive to draw: too many data points, a heavy custom visual, images loaded from the web, or maps that geocode on every load. Reduce the points or change the visual type.
High Other time
Usually the visual spent most of its time queued behind other visuals. Power BI runs a limited number of queries at once, so on a page with 30 visuals the last ones wait. The fix is report design, not DAX: fewer visuals, cards combined into a single multi-row card or table, decorative shapes removed, detail moved to drill-through pages.
4. Take the slow query to DAX Studio
Expand the slow visual in Performance Analyzer and select Copy query. Paste it into DAX Studio, connected to the same model, and run it with Server Timings enabled.
- Mostly storage engine (SE) time with few queries: the data volume is large. Look at the model — cardinality, relationships, aggregations.
- Mostly formula engine (FE) time, or many SE queries with CallbackDataID: the DAX forces row-by-row work. Rewrite the measure.
Iterate in DAX Studio until the query is fast, then change the measure in the model and re-record the page.
-- Pasted from Performance Analyzer → Copy query, trimmed
EVALUATE
SUMMARIZECOLUMNS (
'Date'[Month],
"Margin %", [Margin %]
)5. Desktop timings vs what users see
Performance Analyzer runs on your machine, against the model loaded in Desktop. Users see the published model on shared or Fabric capacity, with other workloads competing for it, and over their own network. A page that is fast in Desktop can still be slow in the service if the capacity is under pressure — check the Capacity Metrics app before blaming the report.
Paste a slow measure into the free DAX Measure Doctor for a quick check of common anti-patterns before you open DAX Studio.
Found the slow visual — now fix the measure
Scan a measure for common performance anti-patterns in the browser, or get a fixed-scope Health Check of the whole model.
Get the next Engineering Note
Practical DAX, semantic model and Fabric notes. 1–2 a month, unsubscribe any time.