The Select Data screen in Nexus (Step 3 of the workflow) provides a powerful set of filters to precisely define the data you receive. Used well, filters are how you turn the breadth of the Exchange into a precise, program-ready dataset.
The most important principle: start broad, then narrow
Every filter in Nexus defaults to All — and in most cases, that’s where it should stay until you’ve seen what’s available. Adding too many filters at the outset what the Exchange has for your area is the most common reason users miss data they would have wanted. You can always refine a query; you can’t recover data you filtered out before you knew it existed.
Filters by Data Type
- Survey Questions: Each of the six data categories (Candidate, Issue, Ballot Initiative, Voting, Demographic, Party ID & Engagement) have their own filters
Because subscriptions are attribute-based, your filter configuration defines the category of data you receive, including questions that don’t exist yet.
- A special caution on Candidate Name: Much of the Support ID data in the Exchange never names a specific candidate. Filtering by name means seeing only questions that expressly name that individual, and missing the majority of relevant data. Leaving it at All allows you to capture the entirety of responses that match your selections, including questions that name your candidate(s) as well as generic and party level responses, which make up a majority of Support ID records.
- Contact Attempts: There are three dimensions to Contact Attempts — Contact Method, Contact Channel, and Contact Response. Contact Method is also how you acquire data by mode. Canvass records for voters who talk at doors, Phone records for voters reachable by phone.
- Universes: A table of all available Contact Mode and Purpose combinations in your geography, with estimated record counts. All combinations are selected by default; uncheck rows to narrow.
One-time-export-only filters
Four capabilities are available only on One-Time Exports - AND stacking, Exclude stacking, Saturation, and Most Recent Response.
Most Recent Response deserves special attention, because members often miss it or assume it’s the default. It limits results to each voter’s latest response for a question type, rather than their full response history. If you want a universe reflecting where voters stand now — not every answer they’ve ever given — select Most Recent Response. It is not on by default.
Saturation (Contact Attempts only) filters by the total number of times a voter has been attempted across the entire ecosystem: Greater Than, Less Than, or Equal To 0. See “Using Saturation Data.”
Layering data types with stacking
Stacking combines multiple data types in a single export or subscription using logical operators — replacing what used to take multiple pulls and a manual join in your own tools. After saving your first data type, click Add another data type and choose an operator:
- OR (Subscriptions and One-Time Exports) — returns any voter matching any selected data type. The consolidation operator: one OR-stacked subscription can replace several single-purpose ones covering the same geography. The same operator powers the dual Candidate configuration — an office-level layer plus a named-candidate layer — described in “Layering Your Data with Stacking.”
- AND (One-Time Exports only) — returns only voters matching all criteria. The intersection operator: supporters who also have a Vote Plan is a sharper GOTV list than either data type produces alone.
- Exclude (One-Time Exports only) — removes voters matching a second data type. The suppression operator: supporters minus confirmed opposition, cleaned before the data ever reaches your pipeline.
Order matters with Exclude: the first data type is your base universe, and each subsequent operator applies to that base. Build your base criteria first, then add Exclude layers. After activating, use View Details on the Data Exports page to confirm the export is configured the way you intended.
Exploring before you commit: Select Responses
For Survey Questions, the Select Responses function shows you the actual question text and response volumes matching your current filters — a discovery tool to use before deciding whether to narrow further. Note that pinning to specific question text stops future matching questions from being automatically included.
Checking scale before you spend
Your estimated Credits, Unique IDs, and Records appear in the Delivery Information bar on the Summary & Confirm screen before you activate. These are estimates, but they give you a meaningful sense of the query’s scale before credits are spent.