Stacking is one of the most consequential capabilities in Nexus: it lets you combine multiple data types — or multiple configurations of the same data type — in a single export or subscription, using logical operators. Work that used to require several separate pulls and a manual join in your own tools now happens inside the platform, before the data ever reaches your pipeline.
Where stacking lives
Stacking is an optional step in the Nexus workflow. After you configure and save your first data type on the Select Data screen, click Add another data type. An operator dropdown — OR, AND, or Exclude — appears before you configure the second layer. You can continue adding layers as your query requires.
OR — the consolidation operator
Available on Subscriptions and One-Time Exports. OR returns any voter matching any of your selected criteria. Meaning that a voter only needs to meet one to be included.
Two patterns organizations have used:
- Cross-category consolidation: separate legacy subscriptions for Candidate support, Issue positions, and Engagement data over the same geography become one OR-stacked subscription. Fewer subscriptions to manage, one place to audit, same or better coverage.
- Dual Candidate configuration: a named-candidate selection stacked with an office-level selection. Configure Candidate with your specific target candidates selected; then add a second Candidate configuration with the same Government Level and Office but Candidate Name left at All. The named layer guarantees you capture questions that expressly name your candidates; the office-level layer captures the generic support scales and unnamed questions that make up much of the data for any race. Together, full coverage — neither layer alone provides it.
AND — the intersection operator
One-Time Exports only. AND returns only voters matching all of your criteria. Meaning every voter in the result has records matching every layer.
Example: Candidate support (responses 1 and 2) AND a Vote Plan. Every voter in the result has both a support ID and an indicated plan to vote. Providing a much sharper GOTV list than either data type produces alone.
Exclude — the suppression operator
One-Time Exports only. Exclude removes voters from your results based on a second data type so every voter with a record matching the excluded layer is stripped from the final result.
Example: Candidate support (responses 1 and 2) Exclude strong opposition (responses 6 and 7). The result is a clean support universe with opposition already removed, creating a suppression built into the export with no manual join.
Order matters with Exclude
The first data type you configure is your base universe. Each subsequent operator is applied to that base. Build your positive criteria first, then add Exclude layers. Adding them in the wrong order can produce unexpected results. After activating, use View Details on the Data Exports page to confirm the export is configured the way you intended.
Choosing your delivery method around stacking
Because AND and Exclude are only available on One-Time Exports, the operators you need determine your delivery method. A common pattern: maintain an OR-stacked Subscription for ongoing coverage, and run periodic One-Time Exports with AND or Exclude layers when you need a precisely cut universe.