Member guide

How to use Electoral Intel

A practical guide to finding poll questions, comparing firms, tracking movement, downloading briefing packs, and reading numbers against public floors — without treating Census or election returns as a second pollster.

Live product: electoralintel.com. Questions: hello@electoralintel.com.

What Electoral Intel is

Electoral Intel is a membership product for approved organizations. It combines a searchable poll archive with tools to generate tables, compare wording across firms, track aligned answers over time, and package client-ready briefing packs. Alongside the archive, Benchmarks surface public floors — Census composition, last official election results, electorate size, campaign finance, and selected BLS economy series — so you can read a poll in context.

Getting access

1. Apply at Apply for Access. Access is reviewed individually (campaigns, parties, media, advocacy, corporate, research, and related organizations).
2. Once approved, sign in via Member Login. After login you land on your account page, then open the client dashboard.
3. Your account page lists which dashboard tools are included on your plan. Tools that are off for your tier appear locked in the dashboard nav.

Quick start

Signed-in members with the onboarding tour enabled see a four-step strip on Search, Generate, and Benchmarks. It is dismissible (Restart brings it back from the first step). The live tour is guidance — click each step; unlike the public demo, it does not auto-run a Kentucky search for you.

1. Search — run a starter, check hits for a briefing.
2. Generate — pick a poll and break by party (Archive+) or pull toplines.
3. Briefing — return to Search with selected hits and download a pack.
4. Benchmarks — open ACS composition and the last presidential floor for the state.

Prefer a no-login walkthrough first? Open the sample product tour at /demo (sample data, Archive-like layout — not your live warehouse).

Your first briefing pack

1. Go to Search.
2. Click a starter such as Presidential approval or Kentucky ballot. Starters run a search and can auto-select a small set of top matches (capped) for a first pack.
3. Skim the result cards: study name, field date, state, Trust badge, and any election-floor line. Expand clusters to see wording variants.
4. Check the hits you want (or use Select top matches). Optionally name the pack.
5. Use the briefing bar to download. Formats depend on your plan (see Export & briefing packs).

Search

Search is the catalog of poll questions. Use it to find wording, open a study workspace, select hits for a briefing, and hand off to Compare or Trends.

Ways to find questions

• Starters — curated one-click searches (approval, ballot, economy, immigration, abortion, and Kentucky ballot).
• Ask the archive — natural-language box that fills filters and runs a search (details below).
• Popular questions — add common wordings from your access tier; browse more by topic when available.
• Manual slots — up to ten question boxes. Combine with each (separate result sections), or (any match), or and (studies matching every query).

Refine after you search

The refine sidebar (when results are present) filters by topic, state, year or year range, sponsor/pollster, mode, and tracked series. Changing a facet re-runs the search. Public+ plans also get topic and population (party / gender) filters on the main form.

Result cards & handoffs

• Open workspace — toplines for this wording, methodology checklist, and series sparkline when available.
• Compare across polls / View trend — deep-link into Compare or Trends with the question context.
• Track series / Compare series — when a study has a series key and your plan includes series handoff, jump straight to that time series.

When you get zero hits

Empty states are intentional. They explain whether a year filter, state, topic, or long phrase is likely the problem, and offer actions such as shortening the query, clearing filters, or browsing residual facet slices. Use Archive coverage (Public+) to see which states and topics actually have depth in your tier — empty search usually means wording or filters, not an empty product.

Ask the archive

Ask is a rule-based interpreter (not a chatbot). It extracts structured filters from a sentence, then runs Search.

It understands:

• Topics — approval, ballot, economy, immigration, abortion, guns, health, foreign, climate, crime (and common synonyms).
• Geography — state names or two-letter codes; “national” clears a state filter.
• Time — a year (e.g. 2024), ranges, “since 2020”, “this year”, “last year”, “past N years”.
• Population — party (D/R/I) and gender (Public+ apply these filters).
• Pollster / sponsor — e.g. Quinnipiac, YouGov, Morning Consult, Marist, Emerson, NYT/Siena, Pew, and others recognized by name.
• Mode — phone, IVR, online, mixed, text.

Examples: “Quinnipiac Pennsylvania approval this year”; “live phone KY ballot”; “immigration favor oppose past 2 years”.

Generate

Generate builds a table from one poll or a tracked series— not a cross-firm compare.

• Toplines — available on Free+. The lead question is charted with a tracked-answer takeaway; the full table lists published choices.
• Break by demographic — Archive+. Dimensions: race, gender, education, age, party. Bars follow the tracked answer across breaks when possible. Some polls only support sample composition (share of respondents) when no ballot/vote column exists.
• Series over years — Public+. Pick a tracked series for a year-over-year sparkline and table.

Compare

Compare lines up the same kind of question across pollsters and returns one tracked answer per poll (for example Approve, Favor, Yes, or a named-candidate share when alignment can resolve it). Public+ only.

• Use starters (approval, generic ballot, issue topics) or paste a short phrase.
• Optional topic and state filters narrow the set. Prefer short keywords (“approve”) over full questionnaire text.
• The bar chart and table show the tracked answer; open a poll name for methodology. A spread line summarizes the range across firms when percentages exist.
• If rows appear but nothing charts, you are often looking at salience / “most important issue” lists without a single aligned share — try an approval or ballot starter, or open Trends.

Trends

Trends answers: where does this answer sit, and is it moving? The default chart is a poll cloud with a monthly median across firms— not a rainbow of one-shot pollster lines. Toggle cloud + median, median only, range band, or latest bars as needed.

• Start from Trends starters, Search handoff, or a series key.
• Optional quality filters: minimum sample size and maximum margin of error.
• Hide individual studies in the legend when one firm dominates the eye.
• Export the latest-point table as CSV or Excel (plan-dependent).

Race Pages

Race pages are decision views for a specific contest: recent ballot polls with disclosure grades, public floors (last presidential and Senate certificates, CVAP), FEC receipts, prediction markets, Poll Desk coverage, and one-click briefing export — so you do not assemble that stack from five tools before a meeting.

• Open Races in the dashboard (Search-capable plans).
• The live templates are Kentucky and Pennsylvania U.S. Senate — same layout clones across states and offices.
• Floors and markets are calibration / context, not a substitute for a poll. Prefer graded ballots over a single leaky topline.
• Use Download race one-pager for a single PDF of floors + selected polls (Archive+). The briefing bar still builds fuller packs.

Export & briefing packs

The Export page is a hub: packaging work you already did in Search, Generate, or Compare. Briefing packs are built from checked Search hits.

Formats by depth

• Public — CSV tables and PNG charts from tools that support export; briefing CSV and citation copy; share links on signed-in accounts.
• Archive / Analyst — Excel and PDF briefing packs, plus richer chart exports from Generate / Compare where available.
• Free — no export depth; use the dashboard to explore, then upgrade for packs.

What is inside a briefing pack

• Cover summary: question count, studies, field dates, geographies, topics.
• Public floor lines when calibration is on (ACS highlights, last presidential result, electorate / CVAP notes, FEC receipts, BLS series with retrieval date and disclaimer, recent midterm floors where present).
• Per-question toplines, methodology fields, and disclosure grades.
• Optional aligned sparklines when related selected hits share a series or near-duplicate wording.
• Memo cover (Archive+ when the memo add-on is on): tracked-answer spreads across firms (“Approve 42–48% across 6 polls”) and disclosure callouts before you cite.

Benchmarks & public floors

Benchmarks collect public context for a geography. Open the full page from the dashboard, or read the compact Census context strip that appears on Search / Compare / Trends / Generate when a state is in play.

• ACS / Census — composition and income (live Census API when configured; otherwise cited QuickFacts fallbacks for selected geos).
• 2024 president — statewide certificate lines. Sub-state ACS views still use the statewide presidential certificate unless otherwise noted.
• CVAP & turnout proxy — citizen voting-age population and votes/CVAP where available.
• FEC — top candidate receipts for the relevant cycle/office when the key is configured.
• BLS — unemployment and CPI-U year-over-year with retrieval date. Exact disclaimer shown in product: BLS.gov cannot vouch for the data or analyses derived from these data after retrieval. No BLS logo is used.
• 2022 / 2018 floors — statewide races where certificates are on file.
• Prediction markets — related Polymarket contracts (prices and liquidity) when a matching market is found. Public betting odds, not surveys; always compare to poll floors separately.

Live Benchmarks can load a county / congressional-district ladder for ACS; if that ladder is unavailable, the page says so. Weighting-target CSV downloads are available from the Benchmarks panel when offered.

Saved searches & alerts

Public+ members can save the current Search queries and filters, rerun them later, and optionally email when something new or moved appears.

• Email alert — weekly by default. Daily cadence is available when the daily-alerts add-on is on for your plan.
• Min tracked % — only alert when the tracked answer is at or above a threshold.
• Move ≥ N pts — default is 3 points; adjustable when daily-alert controls are shown.
• Digests distinguish new matches from moved tracked percentages, and link back to Search and Trends.

Poll Desk

Poll Desk is a curated poll-release radar — not a general news feed. When it is on for your membership (Public+ by default; admins can grant or revoke per client), open Poll Desk from the dashboard nav. Search also shows a short recent-releases strip when the radar is on for you.

• Radar (A) — headlines that look like poll writeups from a fixed outlet list, with Open in Search links.
• Alert explainers (B) — when a saved search moves ≥ your threshold, digests can attach 1–3 related headlines from the same corpus.
• Your controls — turn Poll Desk off for yourself, toggle A/B independently, and filter outlets, states, topics, and lookback. You cannot unlock Poll Desk if your account is not entitled.

Trust & methodology

Every result can carry a disclosure badge. This is not a pollster quality score. It measures how completely the study reports methodology fields you can verify on file.

• Strong — score ≥ 85
• Solid — ≥ 65
• Limited — ≥ 40
• Thin — below 40

Checks include sample size, margin of error, mode, sample frame, weighting note, sponsor, and field date. Open the methodology drawer or workspace checklist for the field-by-field view. Briefing CSV exports include disclosure score, grade, and missing fields.

What each membership includes

Exact packaging can be adjusted per organization. In product terms, depth works like this:

Analyst also includes selected Electoral Intel DB access, executive briefing cadence, and customized cuts as contracted. Your account page is the source of truth for what is switched on for you.

Product demo

/demo is a sample-data walkthrough with an Archive-like layout. Use it to learn the flow before or alongside a live account. Differences that matter:

• Data is static sample content, not your live warehouse.
• The KY tour auto-runs the ballot starter; live accounts do not.
• Saved searches stay in the browser; no email alerts.
• Briefing share links are disabled.
• Benchmarks use the static demo bundle (no live county/CD ladder).

Tips & troubleshooting

• Prefer short search phrases. “Approve” beats a pasted 40-word question. Use Ask or facets for pollster and state.
• Zero hits ≠ empty archive. Clear year or state first; open Archive coverage; try a starter.
• Tracked answer, not every choice. Compare and Trends intentionally align one answer (Approve, named candidate, etc.). Open the study for the full distribution.
• Cite only what disclosure supports. Thin or limited grades and missing MoE/weighting are warnings, not decorations.
• “Search is temporarily unavailable” means the live warehouse connection is down — retry shortly; this is not a filter problem.
• Need a change? Email hello@electoralintel.com or your Electoral Intel contact. Capability add-ons (memo cover, daily alerts, coverage map, series handoff, Poll Desk, and others) can be switched per membership without rebuilding the product.