Custom breakdowns

A custom breakdown lets a user define the rules for how imported data is grouped, on top of the breakdowns the data source itself provides. Use custom breakdowns to organize campaigns, ad sets, or ads into groupings that reflect the business — brand vs performance, product lines, cost centers, geography — and then use those groupings in table widgets, chart widgets, and the List filter widget.

​

Where to create a custom breakdown

Custom breakdowns are created from a widget’s settings panel:

​

Labels — the core of a custom breakdown

A custom breakdown is a collection of labels. Each label is a rule that says which rows of imported data belong to that label. Once a custom breakdown is used in a widget:

• In a table widget, each label becomes a row.
• In a chart widget, each label becomes one bar, slice, or line — one “group” in the Group by.

Every row in the underlying data whose values match a label’s rule is grouped under that label. For example, a custom breakdown called Products could have one label per product type — each label extracting the product name from the ad name, so that every ad promoting a product is grouped under that product’s label automatically.

​

Manual by Custom Filter

In Manual by Custom Filter mode, each label is created by hand and given a set of filter conditions. Any row that matches the conditions is grouped under that label. Labels persist as-is until edited or deleted — the underlying data doesn’t change them automatically. Use this mode when the groupings are fixed and don’t need to reflect new data as it comes in.

​

Auto by Segment

In Auto by Segment mode, labels are extracted automatically from names at a selected level — the extraction can be set to run against Campaign names, Ad set names, or Ad names. As new data is imported that matches the segmentation rule, new labels are added to the breakdown automatically. Two segmentation methods are available.

​

By Separator

Splits the selected level’s name by a chosen character and takes a specific segment as the label. Fields:

• Segmentation method — set to By Separator.
• Separator — the character to split on. Options:
  • Hyphen (-)
  • Underscore (_)
  • Bar (|)
  • Caret (^)
  • Slash (/)
  • Double slash (//)
  • Custom — enter any character or short string (1-100 characters).
• Ad level — the level to extract from: Campaign, Ad set, or Ad. When Campaign is set, extraction runs on each campaign’s name; when Ad set is set, on each ad set’s name; when Ad is set, on each ad’s name.
• Nth group — which segment position (1-indexed) becomes the label. For example, splitting [BRAND]_Summer2025_Meta by _ at position 1 yields [BRAND] as the label.

Rows that don’t contain the separator, or whose split produces an empty segment at the chosen position, are grouped under others.

​

By Pattern (regex)

Extracts labels from the selected level’s name using a regular expression. Fields:

• Segmentation method — set to By Pattern (Regex).
• Regex — the regular expression pattern. The first capturing group in the pattern becomes the label. Example: \[([^\]]+)\] extracts anything between square brackets — so [BRAND] Summer Campaign yields BRAND.
• Ad level — the level to extract from: Campaign, Ad set, or Ad. Extraction runs against the selected level’s own name.

Rows that don’t match the pattern are excluded from the breakdown result (not grouped under “others”). This is the key difference vs the Separator method.

​

Label limit

Auto by Segment caps at 10,000 labels per breakdown. If extraction produces more, the top 10,000 by matching row count are kept.

​

Scope

Custom breakdowns can be saved at two scopes:

• Workspace-scoped — available across every teamspace and dashboard in the workspace.
• Teamspace-scoped — available only inside the teamspace where it was created.

Pick the scope when saving the custom breakdown. A teamspace-scoped custom breakdown doesn’t appear in a different teamspace’s breakdown picker.

​

Troubleshooting

For common problems — labels missing campaigns, totals not matching the ad-account total, auto-generated labels hitting the label cap, and other configuration pitfalls — see the Breakdowns & metrics FAQs.

​

Related