theDATAfirm Humans First RVP · RealVirtualPeople · India Talk to us

Everyone else
sells signals.
We rebuild
the human.

RVP is a working model of India’s whole population — 1.4 billion virtual people, over 5,000 attributes each, every ward and every village. Built from public data and published reports, and collected from nobody.

Test it on a market you know cold

Engagements begin with one question, one market and a fixed fee. Built by seven people with 155 years across research, technology, media and analytics.

What a signal says

A device visited three car websites this week.

What RVP knows

Whether this ward holds that a car reveals who you are, or that a car is only for getting to work — and which two-wheeler households sit ready to move up.

Why we built it

The need not to be watched.
The need to be understood.

Technology now proves itself by how deeply, how quickly and how cheaply it can watch people — and businesses, governments and societies genuinely do need to understand people better in order to build, plan and serve.

But at the heart of being human is the need for privacy. The need not to be watched or judged constantly. To be oneself. To express, to choose, to disagree, to be illogical, to be intimate, to be alone. And just as human is the need to be understood, recognised, valued — to belong.

Everyone wants to be understood.
Nobody wants to be watched.
Until now, you could not have both.

RVP exists to end that conflict rather than pick a side: understanding, without surveillance.

500,000+ hours of work behind the build, and counting

Data scientists. Analysts. Technologists. Statisticians. Business leaders. Marketers. Sales professionals. Students. Partners. Families.

A living testimony of a single belief.
Humans First.

How it’s built

Built person by person.
Read as 25 million+ cohorts.

25m+ distinct cohorts — by who they are, where they live, how they were schooled, what they speak and what drives them. Not segments drawn over a population. A cohort here is what appears when 1.4 billion virtual people, each built one at a time, are read together.

Modelled on the genome. A fixed architecture of positions, identical for everyone, carrying a different value at every position for each person. An inspiration for structure only — no biological, genetic or personal data is used anywhere in RVP.

Constructed, not collected. No personal record was ever ingested, so there is no consent chain to defend and nothing to erase when someone asks.

Built from the person up. Every virtual person is constructed individually; nothing is ever published below a cohort. A cohort is a result here, never a starting assumption — which is why the parts always sum to the whole.

Depth per person
MatrAIx Harvard & MIT · 8.3bn AI personas1,290 dimensions
RVP · carried today5,000+ attributes
RVP · architecturebuilt to carry 1.6 million+ data points per person

MatrAIx is the AI persona dataset Harvard and MIT released in August 2026 — 8.3 billion simulated people, roughly one for everyone alive. It settles whether a population can be rebuilt in software. How deep, and from what, is a different question.
Drawn to scale: we fill only what the published record supports, which is why the third bar is not full. The architecture is not the limit — the evidence is.

How far it reaches

A village of four thousand people,
read as closely as a metro ward.

Nine settlement strata — every one read at the same depth
Mega50 lakh and above
City10–50 lakh
City5–10 lakh
Town1–5 lakh
Town50,000–1 lakh
Townunder 50,000
Ruralabove 10,000
Rural5,000–10,000
Ruralunder 5,000
Most consumer data thins out here
Four levels of resolution
CountryWhere plans are written.
StateWhere data stops being reliable.
DistrictWhere panels run out of sample.
Ward & villageWhere RVP reads.
Geography, benchmarked
MatrAIx Harvard & MIT · 8.3bn AI personasno micro-geography
RVP · The Data Firm759 districts · 81,000+ wards · 6.5 lakh+ villages

Their 8.3 billion personas live nowhere in particular. A market decision is made in a district.

What it answers

One population.
Cut for the question you actually have.

▸ Pick your industry — or watch it cycle
Personal care

“Which towns will pay for the premium variant, and which will only ever buy the sachet?”

    The same people, grouped a different way · illustrative

    Sixteen lenses built, more added as they are asked for. Forty consumer categories underneath — cooking oils, soaps, shampoos, detergents, tea, biscuits, dairy, snacks, chocolates, skincare, baby care, sanitary pads, OTC medicines, water purification and more. Read at category and tier: branded against unbranded, premium against mass.

    Work delivered · minimising wastage, maximising ROI

    Nobody buys a dataset.
    They buy an answer to one question.

    reachablemedia dark
    Uttar Pradesh · PepsiCo

    “Where in this state is nobody reachable by any media at all?”

    Every media-dark micro-geography in the state, found and mapped — delivered as a live dashboard filterable down to a single ward or village.

    • 121,907micro-locations
    • 14,455urban wards
    • 107,452villages
    • 24.6crpopulation
    • 13.3crbeverage base
    • 18.8crsnacks base

    “Our drinker sweats. So where do they sweat?”

    Adults 18–34 who exercise, play, practise sport, run or walk outdoors — intersected with gyms, playgrounds and courts across every ward of Mumbai and Navi Mumbai, to build catchment, distribution, GTM, e-commerce and trade marketing strategy.

    PepsiCo · GTM, distribution & media

    “How many older Indians are there really?”

    Not a share of a national total. Every one located to their ward or village, and profiled by household, means and outlook.

    GenSxtyTribe · markets counted in people

    “Which micro-markets hold enough of our buyer?”

    An ideal-customer profile translated into micro-geography, then every place ranked by how many matching households actually exist in it.

    Aranyaka Estates · micro-geography

    GTM, distribution, media & catchment PepsiMountain Dew StingGatorade
    Markets counted in people Shriram Life InsuranceESAF Small Finance Bank GenSxtyTribeCrisp Insights
    Micro-geography Aranyaka Estates
    Maruti Suzuki Mobility Challenge
    Second runner-up, 2026     Maruti Suzuki Mobility Challenge
    India partner One & One Consulting
    Technology partner DOMO

    Answered from RVP alone — no client data used, and none needed. Crisp Insights resell the same work to their own clients, under their own name. Second runner-up, Maruti Suzuki Mobility Challenge 5.0, run with T-Hub, March 2026. Diagrams are schematic. All logos are the property of their respective brand owners.

    See it working

    Don’t imagine it.
    Open it and click around.

    Two things you can look at right now, without talking to anyone. Both are live.

    Live dashboard Cut a population and watch it reconcile.

    A working BFSI dashboard built on RVP and served through DOMO. Move between geographies, switch the population base, and see what an answer actually looks like before you ask us for one.

    • DataIndicative
    • MachineryReal
    • People in itNone
    Open the dashboard  →     Public archive · April 2020 The estimate we published for the country, and gave away.

    National COVID-19 reference estimates resolved to district, sub-district, town and ward across all 37 states and union territories, cut by age and gender. Released openly, as ranges, marked indicative rather than absolute — because that is what the evidence supported.

    • PublishedApril 2020
    • ResolutionTo the ward
    • PriceFree
    Read the archive  →

    Both open on thedatafirm.in, our public archive. The dashboard shows indicative data to demonstrate the shape of an output — it is not a live read on any market, and there is no audience in it to buy.

    Check the work

    You don’t have to trust us.
    You can check us.

    1
    The parts sum to the whole.Every figure reconciles to the population it is drawn from.
    2
    A market cannot exceed the people in it.Bounded by the humans who exist. Reach figures elsewhere routinely are not.
    3
    However finely you cut it, you never reach a person.Every answer describes a group big enough that no household inside it can be picked out.
    4
    Every number carries its confidence.Marked certified or directional where it appears, not buried in an appendix.
    5
    Reference sources named, not counted.Census, NSS, NFHS, RBI, TRAI, multiple ministries, the World Bank, NCAER, the Lancet…
    —
    And where it does not apply.No brand-level data. No audience to buy. No real-time measurement. We stop above the street corner.
    Provenance, benchmarked
    MatrAIx Harvard & MIT · 8.3bn AI personas600,000 records grounded in real human data, de-identified
    RVP · The Data Firmnothing about any person collected, held or inferred

    De-identified is not the same as never collected. A device graph can tell you it reached four hundred million Indians — not how many people that is, or one true thing about any of them.

    RVP will never tell you which person. It will tell you how many there are, how they live, what they believe, and what would move them.

    Start here

    Test it on a market
    you already know cold.

    Pick a district, a category or a segment you understand better than anyone. Write your own view down first, sealed, before we show you ours. Then judge the fit rather than the novelty.

    Most engagements begin with one question, one market and a fixed fee.

    or write to

    The Data Firm · Dubai International Financial Centre, Sheikh Zayed Road, Dubai, UAE

    Entering IndiaHow big is this category in every district?
    Sizing a targetIs this market the size they say it is?
    New statesWhich micro-markets carry real demand?
    Marketing & mediaWhich language and disposition, ward by ward?
    Risk & inclusionWhere are the uninsured and unbanked?
    Policy & healthWho is not reaching care, and why?
    Who we are

    Seven people who spent thirty years
    on the other side of this problem.

    Dr. Amarnath30 yrs+IBM, TCS
    Habeeb25 yrs+IPG, WPP, The Times of India
    Arvindh20 yrs+Infosys, Cisco
    Love20 yrs+Publicis, IPG, WPP
    Pradeep20 yrs+Crane Software
    Sriraj20 yrs+Accenture, Tiger Analytics
    Vijay20 yrs+BARC, WPP, IPG
    155yrs of combined experience across research, technology, media and analytics

    Business leaders, innovators and domain experts who have worked across multiple geographies and verticals — and who, in those prior roles, were responsible for client portfolios worth more than $100 billion for companies including Amazon, Microsoft, L’Oréal, ING, Intel, Samsung, ICICI, Tata, Mahindra, Nestlé and PepsiCo.

    Those engagements were at the firms named above, not at The Data Firm. We mention them because they are where this problem was learned.