- The visual shows "Add an ID field (or a pre-aggregated Count)" instead of a chart.
- The visual counts distinct IDs, so it needs to know what one person or thing is. Add the respondent, customer or user ID column to ID (respondent, customer…). If your table is already a summary of combinations and counts, put the number in Count (pre-aggregated) instead.
- Everyone appears to be a member of every set, including people who answered No.
- A text column dropped into Set flags (one per set) is aggregated with Count by default, and Count counts "No" as a value. The chart says so in its notes. In the Visualizations pane, open the field's menu under Set flags and change the aggregation to First or Max. 1, TRUE, Yes and other non-blank text then mean member; 0, FALSE, No and blanks do not.
- Shares look too high, or "None of these" is missing.
- The base is every ID that reaches the visual. If your table only holds rows for people who selected something, people who selected nothing are not in it and cannot be counted. Keep a row for them, either with a blank answer or with an explicit answer such as "None of these". Answers listed in Format > Data > "None" answers (by default None of these; None; None of the above; Nothing) count in the base but in no set; add your own wording there, separated by semicolons. View > Show "none of these" controls whether the bar or label is drawn. Rows with a blank ID are left out (the notes count them), because a blank is nobody in particular.
- Filtering the report to a few sets changes every percentage.
- A slicer or filter on the Sets field removes rows, so IDs that only belonged to the filtered-out sets leave the base. To hide sets without changing the base, leave the filter off and switch sets off in Format > Sets shown, or use Max sets (largest first). IDs whose only sets are hidden are kept in the base and shown as "Only other sets", never mixed into "None of these".
- The UpSet bars add up to more than the base, or a combination is larger than I expected.
- Check Format > View > Intersections. Exclusive (exactly these sets) puts each ID in one bar, so the bars plus "None of these" add up to the base. Inclusive (at least these sets) counts an ID in every combination it contains, so bars overlap by design. The base note under the chart always says which one is shown. Venn regions are always exclusive.
- Some combinations or sets are missing from the UpSet.
- The notes say why. "+N more combinations" means Max intersections (20 by default) or the visual's size cut the list; raise it in Format > UpSet or enlarge the visual. Min size, Min degree, Max degree and Hide empty in the same card filter combinations out. "The N largest sets fit this size" means the visual is too short for every set row: make it taller or pick sets in Format > Sets shown. More than 30 distinct sets are folded to the 30 most frequent, and the notes count the rest.
- Weights, letters, TURF or the duplication matrix show in Desktop but not for my colleagues.
- These are the premium Overlap intelligence features. They preview in full in edit mode, so the author sees them. A reader without a licence gets the free chart instead, never a blocked visual: weighted numbers become unweighted counts labelled "unweighted", Duplication matrix and TURF reach fall back to the Venn or UpSet view, and Compare by shows one pooled chart. A host banner names what a licence would add. To give readers premium, buy the plan in Microsoft Marketplace and assign a licence to each reader in Microsoft 365 admin center > Billing > Licenses.
- Can I use the chart with a keyboard or in high contrast?
- Yes. Tab into the visual, move between bars, regions and sets with the arrow keys (Home and End jump to the ends), select with Enter, add to the selection with Ctrl+Enter and clear with Escape; each element is announced with its count and share. In Windows high contrast the visual uses the system colours and drops the fills.
