Measure Killer Measure Killer

How do I see which semantic models use Import, DirectQuery or Direct Lake?

Run Tenant Analysis on the models and filter the Storage mode column on the Semantic models tab. Read it next to size, CU (s) and the load times of the reports, because a slow report on a DirectQuery model often points to the source, not the model.

Last updated · November 9, 2026

Steps

  1. Run Tenant Analysis or Limited Tenant Analysis, tick the models and click Analyze (details).
  2. On the Semantic models tab, filter or sort the Storage mode column: Import, DirectQuery, Dual or Direct Lake (details).
  3. Read it next to the other columns:
    • Import with a large size: remove unused columns and measures first (details).
    • DirectQuery or Dual with a high CU (s): every visual sends queries, so the source and the number of visuals matter most.
    • Direct Lake: check for implicit measures, which Direct Lake does not support (details).
  4. Compare with the Average load time of the reports on each model (details).

What you see

Semantic models tab with the Storage mode column showing Import, Direct Lake and DirectQuery per model, next to used size, size, average refresh duration and best-practices score

Go deeper

Something wrong or out of date on this page?

Report an issue with this page

Tell us what is wrong or out of date. No email needed.