A campaign is a rhythm, not a series of events.
Six phases across 180 days. Each has a defined output, a named owner and a decision that either happens on time or does not happen at all.
Diagnose
Find out what is actually true about the seat before anyone spends money on it.
Work in this phase
- Baseline survey: 2,400–25,000 respondents, stratified and weighted
- Booth classification from three cycles of historical data
- Candidate reputation and search audit
- Opposition mapping, vote-cutter analysis and alliance maths
- Party structure audit — how many listed booth workers actually exist
Output
Seat Diagnostic: win probability, swing booth list, top-3 issues, honest go/no-go.
Build
Stand up the machine — room, people, systems, compliance.
Work in this phase
- War room build-out and staffing
- Booth committees and panna pramukh recruitment and verification
- DLT registration, WhatsApp Business API, sender IDs, ad account verification
- Candidate website, voter app and worker app deployment
- Expenditure register opened and reconciled from day one
Output
Operational war room, verified ground structure, every compliance approval in hand.
Narrate
Set the story while there is still time for it to sink in.
Work in this phase
- Campaign song, anthem and creative production
- Daily content engine: 8–12 assets a day in the local language
- Reputation asset build and search result reshaping
- Grievance capture live — every closed complaint is a vote
- Weekly kill/scale review on every message theme
Output
A message voters repeat back in their own words, measured weekly.
Saturate
Everything on, aimed at the swing booths, tracked daily.
Work in this phase
- Mid-campaign tracker survey and full message re-calibration
- Bulk SMS, IVR and WhatsApp sequences by voter segment
- Vehicle fleet, roadshows and ground events on the swing-booth route
- Printing at scale with 36-hour reactive turnaround
- Rapid response desk at 30-minute SLA, 18 hours a day
Output
Maximum contact frequency where it changes the result, not where it feels good.
Convert
Turn identified support into counted votes.
Work in this phase
- Rolling 3-day tracker to catch late swing
- Voter slip distribution completed and verified
- Turnout call and message sequence to identified supporters
- Polling agent deployment and booth-day briefing
- Silence-period compliance check across every channel
Output
Every identified supporter knows their booth, their serial number and their transport.
Deliver
The day the previous six months either shows up or does not.
Work in this phase
- Hourly booth-wise turnout tracking from 7 AM
- Live vehicle dispatch to booths running below pace
- Complaint escalation line to the returning officer
- Counting agent deployment, round-wise trend capture
- Single spokesperson line and result-day communication plan
Output
Command, not chaos — and a full data handover the week after.
Three things we turn down work over.
Being explicit about this saves everyone a month. If any of these are central to the plan, we are the wrong firm.
Fake accounts and bought engagement
Beyond the ethics, Meta's coordinated-behaviour takedowns and public ad library make it a detectable liability. We build attributable assets only.
Fabricated attacks on opponents
We will run hard, evidenced contrast on record and performance. We will not invent, doctor or plant. It is also the fastest way to lose a defamation case mid-campaign.
Under-invoiced expenditure
Every rupee is documented against the district rate schedule for the expenditure observer. If a campaign needs the books to look different, that is not our engagement.