Every tool has one nagging seam, and the PivotTable’s is the Refresh button: a pivot is a snapshot, not a formula, and everyone has shipped a report whose pivot was one refresh behind the data. Modern Excel (365) closed that seam with two functions that do a pivot’s job the way everything else on this stage works — as formulas that spill, recalculate instantly, and can be built into other formulas:

=GROUPBY(Orders[Region], Orders[Amount], SUM)

One line: regions down the side, summed amounts beside them, totals included — a summary table that is live. Add a row to the Table and the summary already knows; no cache, no stale copies, nothing to press.

PIVOTBY adds the second dimension — the classic rows-by-columns grid:

=PIVOTBY(Orders[Region], Orders[Month], Orders[Amount], SUM)

Regions down, months across, sums in the middle. The third sibling you already know: the spilled UNIQUE + SUMIFS combo was this idea assembled by hand; GROUPBY is it as a single verb.

the data North · 1,240 South · 860 North · 500 North · 300 ← typed =GROUPBY(Region, Amount, SUM) Region Sum North 1,740 North 2,040 ✓ South 860 Total 2,600 Total 2,900 ✓ no Refresh exists — the summary is a formula, current the instant data changes
Type a row; the summary already has it. The pivot's one seam — the refresh — simply doesn't exist in formula-world.

Where they beat pivots — and where they don’t

Formulas win when the summary is a component: feeding a chart that must never be stale, sitting inside a dashboard’s Calc layer, wrapped in FILTER for a criteria-driven summary, nested inside LET, or — the quiet killer feature — any aggregation as the third argument, including a LAMBDA: GROUPBY(regions, amounts, LAMBDA(x, PERCENTILE.INC(x, 0.5))) gives you median by region, something the classic pivot has refused to do gracefully for thirty years.

Pivots win on interactivity and audience: slicers, drag-to-regroup, Show Values As, drill-down on double-click — exploration by mouse, for people who will never edit a formula. And at data-model scale, pivots-on-the-model remain the tool; GROUPBY eats Tables, not millions of engine rows.

The division falls out neatly: pivots for exploring and for audiences; GROUPBY/PIVOTBY for machinery — the fixed summaries inside reports that must always be current. A lot of workbooks want both, doing different jobs, fed by the same Table.

Two cautions while these functions are young: they need current 365 (a workbook leaning on them excludes older Excels — know your audience), and their argument lists run deeper than shown here (field headers, sort orders, totals control) — the usual advice applies: learn the simple form cold, add arguments when a real report asks.

Thirty years ago the pivot made aggregation a drag-and-drop. This makes it a word — and words compose, refresh themselves, and never go stale. The seam is closed.