Brand strategy · Cultural analysis · September 2024
Strategy for a generational jeweller
A four-generation family house was losing ground to organised chains spending 4 to 5% of revenue on marketing alone. We read what jewellery actually means in India, from the Samudra Manthan to the shringara that brings a murti to life, and set it against the codes the category advertises with. The client's cultural capital turned out to be innate, resilient and essentially free to leverage.
7Ways of relating to jewellery
8Devices in category communication
4Costed strategic routes
↓ Download the case study
Brand strategy · Socio-cultural analysis · 2024
Reading money in India for a fintech brand
A digital-first buy-now-pay-later product had mass awareness and no depth of use. Every competitor was pulling the same levers: more rewards, higher limits, longer windows. We went underneath, decoding five eras of money and four registers of borrowing, and found a category speaking entirely in settlement mechanics. Nobody had an emotional moat because nobody had language outside the repayment schedule.
5Eras of money decoded
4Registers of borrowing
2Brand territories built out
↓ Download the case study
Real estate · Pricing science · February to March 2026
Premium residential pricing and positioning
A premium gated high-rise in central Hyderabad with no local comparable, so every decision needed evidence. Ten in-home immersions found what families actually argue about, and only those variables were tested against a hard price tag with 614 budget-vetted buyers. Price came third. The gated format itself drove 6 to 12%, because buyers at this level treat the gate as a baseline right.
614Budget-vetted buyers
10In-home immersions
4Buyer playbooks
↓ Download the case study
Consumer intelligence · Qualitative research · May to June 2026
The Telugu NRI property buyer
Immigration policy, AI pressure on IT careers and a surging India growth narrative were hitting the Telugu diaspora at once. Fifteen conversations across the United States surfaced the divergence that mattered: the people making the most noise were not the people buying. Screened rigorously for in-market buyers, a different demographic appeared. Settled, unhurried, driven by planned life phases rather than visa anxiety.
15Depth interviews across the US
2Phases, different sampling logic
4Obstacles between interest and purchase
↓ Download the case study
Political strategy · Socio-cultural research · May to July 2025
The first-time voter study
Roughly a hundred million Indians cast a first national vote in 2029, and they have no electoral record to model. Run a full year before Gen Z protests dominated the news cycle, this mapped the cohort whole across 33 cities, from Jhansi and Cuttack to Mumbai. The finding that reframed the media plan was a three-step trust pipeline, where Instagram carries enormous affinity and almost no credibility.
426Respondents aged 15 to 20
33Cities, metro to Tier 2
3Step media trust pipeline
↓ Download the case study
Political strategy · Community intelligence · January to February 2024
Scheduled Caste perception study
Indian strategy has long treated the Scheduled Caste vote as a single bloc, read through aggregate polling. We went to the margins instead: shanties along railway tracks, settlements on the edges of villages, sixteen locations across nine states. Each of the thirteen communities was a universe with its own grievances and its own local heroes. Scheduled Caste is an administrative category, not a voting bloc.
320Respondents, focus-group led
13Distinct communities
9States, 16 locations
↓ Download the case study
Electoral research · AI-led data collection · April 2026
Voice agents at state scale
Reaching all 294 West Bengal constituencies inside a campaign window normally forces researchers onto modelled estimates for most of the state. We removed the constraint. AI voice agents ran up to 2,000 simultaneous lines for sixteen days, with a human-interviewed subset running alongside throughout to calibrate selection effects. Every seat measured directly, none extrapolated from its neighbours.
2.22LValid responses collected
294Constituencies in 16 days
84.35%Of seats called correctly
↓ Download the case study
Every download is the full five-page method write-up. The framework behind the study, the analytical moves that produced the finding, and what the client did next.
We ask for a name and a work email. Nothing else, and no mailing list.