When =PY() is the right tool — and when it isn't

When =PY() is the right tool

• Anything that wants a real DataFrame: pivots, groupbys, complex joins.
• Statistical analysis: regressions, t-tests, distributions, percentile bands.
• Any chart Excel can't make cleanly (heatmaps, violin plots, faceted small multiples).
• Date arithmetic that would need ten nested formulas.
• Working with data that doesn't fit on a normal sheet (rare in =PY(), but possible).
• Anything you'd want to re-do programmatically later.

When to stick with native formulas

• Simple lookups, sums, counts, and conditional formatting — XLOOKUP, SUMIFS, COUNTIFS are fast and obvious.
• Interactive what-ifs where every keystroke should ripple through the whole sheet.
• Workbooks shared with people who don't have Microsoft 365 (their cells will show #BLOCKED!).
• Anything that needs to recalc in < 100 ms.

The hybrid pattern

Use native formulas for the live, interactive parts. Use =PY() for the heavier analysis next to them. The two coexist beautifully.

Two performance notes

• =PY() runs in the cloud. Each calc is a round trip. Don't put 500 of them in one workbook.
• One bigger =PY() beats many small ones. Do as much per cell as you can; one cell that produces a 10×10 table is faster than 100 cells each producing one value.