What does The Data Firm do?
The Data Firm answers market questions for businesses operating in India — how big a market is, who exactly is in it, and where those people are. It does this from RVP, its own working model of India's entire population. Companies use it to size a market, find where their customers actually are, plan distribution and media, and decide where to enter next.
What is RVP?
RVP stands for RealVirtualPeople. It is a working model of India's whole population: 1.4 billion virtual people, each carrying over 5,000 attributes, resolved to every ward and every village. It was built from public data and published reports, and collected from nobody.
Does RVP contain personal data?
No. Nothing about any person is collected, held or inferred. No personal record was ever ingested, so there is no consent chain to defend and nothing to erase when someone asks. RVP will never tell you which person — it tells you how many there are, how they live, what they believe and what would move them.
How can a population be modelled without collecting personal data?
Every virtual person is constructed individually rather than drawn from a record. The structure is modelled on the genome: a fixed architecture of positions, identical for everyone, carrying a different value at every position for each person. The architecture is an inspiration for structure only — no biological, genetic or personal data is used anywhere in RVP. Values are built from public data and published reports, including the Census, NSS, NFHS, RBI, TRAI, multiple ministries, the World Bank, NCAER and peer-reviewed literature.
How granular is RVP?
It reads at ward and village level: 759 districts, 81,000 or more urban wards, and 6.5 lakh or more villages. Nine settlement strata, from mega cities of 50 lakh and above down to rural settlements under 5,000 people, are all read at the same depth — which is where most consumer data thins out. Read together, the population resolves into more than 25 million distinct cohorts.
Can an individual household be identified in RVP?
No. Nothing is ever published below a cohort. However finely a question is cut, every answer describes a group large enough that no household inside it can be picked out.
How does RVP compare to MatrAIx from Harvard and MIT?
MatrAIx is the AI persona dataset Harvard and MIT released in August 2026: 8.3 billion simulated people, roughly one for everyone alive, carrying 1,290 dimensions each and grounded in 600,000 de-identified human records. It settles whether a population can be rebuilt in software. RVP differs on three counts: depth, with 5,000 or more attributes carried per person today against 1,290; geography, with resolution to district, ward and village where MatrAIx has no micro-geography; and provenance, where RVP collected nothing about any person at all, rather than de-identifying collected records.
What can RVP not do?
There is no brand-level data, only category and tier — branded against unbranded, premium against mass. There is no audience to buy and no real-time measurement. RVP stops above the street corner. Every number carries its confidence, marked certified or directional where it appears.
How does an engagement start, and what does it cost?
Most engagements begin with one question, one market and a fixed fee. The recommended first step is to test RVP on a district, category or segment the buyer already knows better than anyone: write their own view down first, sealed, then judge the fit rather than the novelty. Enquiries go to hello@thedatafirm.co.
Can I see RVP working before talking to anyone?
Yes. Two things are live on thedatafirm.in, the public archive. A working BFSI dashboard built on RVP and served through DOMO shows the shape of an output — the machinery is real, the data is indicative, and there is no audience in it to buy. The second is the national COVID-19 reference estimate published in April 2020, resolved to district, sub-district, town and ward across all 37 states and union territories, released openly and free as ranges marked indicative.
Who built RVP?
Seven people with 155 years of combined experience across research, technology, media and analytics, from IBM, TCS, Infosys, Cisco, Accenture, Tiger Analytics, BARC, IPG, WPP, Publicis and The Times of India. In those prior roles they were responsible for client portfolios worth more than $100 billion. More than 500,000 hours of work sit behind the build. The Data Firm is registered in the Dubai International Financial Centre and operates in India.
What work has The Data Firm delivered?
For PepsiCo in Uttar Pradesh, every media-dark micro-geography in the state was found and mapped across 121,907 micro-locations, delivered as a live dashboard filterable to a single ward or village. In Mumbai and Navi Mumbai, adults aged 18 to 34 who exercise or play sport outdoors were intersected with gyms, playgrounds and courts to build catchment, distribution, go-to-market, e-commerce and trade marketing strategy. For GenSxtyTribe, older Indians were counted and located to ward or village rather than taken as a share of a national total. For Aranyaka Estates, an ideal-customer profile was translated into micro-geography and every place ranked by how many matching households exist in it. Crisp Insights resell the same work to their own clients under their own name. The Data Firm was second runner-up in the Maruti Suzuki Mobility Challenge 5.0, run with T-Hub in March 2026.
