Campaigns lose before the campaign starts.
Most campaigns run on instinct, outdated surveys, and partisan optimism. The opposition does not. The intelligence that decides elections lives below the constituency average — and it moves daily.
Surveys describe the past
Generic polls average across communities — hiding the shifts that swing seats. By the time results arrive, the moment to act has passed.
Constituency is too broad
Elections are won at the booth and poll level, among specific communities, in specific blocks. Surface analysis leaves the real battlefield unread.
Campaigns move daily
Media narratives and community sentiment shift every day. Weekly briefings cannot keep a campaign competitive in a live environment.
Intelligence stops at the report
Most research is delivered once and goes stale. A campaign needs intelligence that keeps moving with the ground — daily, not quarterly.
Six layers of intelligence. One integrated platform.
Electoral Intelligence
Historical results, voter roll analysis, booth-level mapping, demographic overlays, and swing zone identification. The foundation.
Ground Research
Constituency scans, field interviews, booth and ward surveys, opinion-maker mapping, and community mood tracking.
Modelling & Forecasting
Swing models, MRP seat forecasts, voter cohort segmentation, turnout scenarios, and win-probability bands.
Campaign Intelligence
Message strategy, outreach roadmap, resource allocation, digital calendar, and field target setting.
Real-Time Operations
Daily intelligence briefs, opposition tracking, media monitoring, rapid response, and a live field feedback loop.
Image & Narrative
Candidate perception audit, media sentiment scoring, narrative threat detection, and attack-prevention playbooks.
Decision-ready intelligence, the same day.
Sample data for demonstration. Live engagements run on client-supplied and field-collected intelligence under NDA.
Named. Scoped. Delivered.
We go all the way down to the booth.
Generic research stops at the constituency average. We keep going — segment, booth, community, household cluster — because that is the level at which seats are actually won and lost.
Constituency
The baseline picture — historical results, demographic overlays, and the win-path memo.
Segment
Swing zones, issue heat maps, and the blocks where the contest is genuinely live.
Booth & Ward
Booth-level intelligence below the constituency average — where most firms never look.
Community
Community vote-equation mapping and the local leadership networks that move it.
Punjab 2027 — the Gurdaspur segment.
Our pilot programme began with booth-level historical analysis across the segment, layered with field interviews and opinion-maker mapping. Within three weeks, NEXUS identified fourteen swing booths invisible to constituency-level polling — and an agrarian-distress signal moving faster than any published survey had registered.
Client identities and engagement specifics remain confidential. This note reflects methodology and demonstration findings only.
How we work →Built by strategists, researchers, and data scientists.
Principal Strategist
Owns intelligence synthesis and client advisory across every engagement.
Director, Ground Research
Leads field researchers local to the constituencies they cover.
Head of Data Science
Maintains the forecasting models, dashboards, and access-controlled data layer.
Narrative Intelligence Lead
Runs sentiment scoring, narrative threat detection, and rapid response.
Functional leadership shown. Named profiles available to qualified leadership under NDA.
The brief, updated daily.
Gurdaspur segment model refreshed — 14 swing booths holding; agrarian-distress salience steady at 88.
Hung assembly remains the modal outcome across the latest 117-seat run; three segments shifted to toss-up.
Booth-level intelligence now active across four constituencies; opinion-maker mapping expanded this week.
From the desk.
Why booth-level beats the constituency average
The signal that swings seats is thrown away the moment you average across communities.
Reading the community vote equation
Community and caste equations decide booths. Here is how we map the networks that move them.
Reading sentiment before the polls move
Sentiment turns days before published polling catches it. Here is how we detect it.
Request a confidential intelligence briefing.
First conversations are exploratory — a 30-minute listening session, no commitment. We respond within 24 hours to qualified campaign and party leadership.