The Man Who Wrote to the Chief Statistician of India — And Was Ignored
Between 2013 and 2015,
a Mumbai entrepreneur wrote to a student, a tech startup, and the Chief
Statistician of
the Government of India — proposing what would
become a billion-dollar global industry.
None of them acted on it. The world did.
Three
Letters. Three Recipients. Zero Response.
Most visionaries get ignored once. Hemen Parekh got
ignored three times — by three very different people —
while saying essentially the same thing.
In May 2013, he wrote to Rohini
Damahe, a lecturer at L&T Institute of Technology, asking
her to build a data-mining engine on top of his database of five million Indian
job advertisements.
In February 2016, he forwarded the same proposal
to Deepak, co-founder of Julia Computing — a high-
performance scientific computing startup — framing
it as "a service to the nation."
But it is the letter in between — dated 27 October
2015 — that deserves the most attention.
That one
was addressed to Shri TCA Anant, Chief Statistician of India, Ministry
of Statistics and
Programme Implementation, Government of India.
Its subject line: "A
faster / cheaper / more accurate alternative to the Annual Employment Survey?"
What He Was
Proposing to the Government
India's Annual Employment Survey is a large,
expensive, time-consuming exercise. Surveyors go door to door.
Companies
fill out forms. The data arrives months late,
covering a world that has already moved on.
Parekh's proposal was simple and radical:
you already have the data. It's sitting on Naukri,
TimesJobs, and MonsterIndia. Let me show you what it can tell
you
He had, at that point, a database of over five
million job advertisements downloaded over six to seven years
from Indian job portals.
Each record contained the company name,
designation, job description, desired profile,
compensation, experience required, industry
type, education qualification, city, keywords, posting date, and
expiry date.
From this, he argued, you could answer questions
that no Annual Employment Survey had ever attempted:
Which companies are hiring, and when?
Which cities have the most and fewest new jobs?
Which industries are growing and which are
contracting?
What education qualifications are in maximum
demand?
What is the correlation between a job title and
years of experience required?
Can the data predict, six months in advance, how
much the job market will shrink or grow?
He even proposed cross-referencing job-ad data with
IIP figures, the RBI's primary bank rates, the stock market
index, the
rupee exchange rate, GDP growth, and the current account deficit — to build a "Predictive Model of
Job Market" that would tell economists not just where hiring
stood today, but where it was heading.
"Such a model," he wrote, "would be of immense interest to not only the
economists but also to the HRD Ministry,
Planning
Commission, Educational Institutions, and of course the students
themselves."
The Phrase
That Captures Everything
Buried in the 2015 letter is
a phrase Parekh had borrowed from author John Battelle's book about Google. He
used it to describe what five million job
advertisements actually are:
"A database of intentions."
Every job
ad, he argued, is an intention.
A company intending to grow a team, enter a market,
acquire a skill, or replace a departing employee. Multiply
that by five
million records across seven years, and you have the
most detailed, continuous, real-time record of
what the Indian economy wants from its people — updated daily,
free of charge, requiring no surveyor, no form,
no delay.
The Annual Employment Survey costs crores of rupees
and delivers data that is already old when it arrives. The
job-ad
database costs nothing to collect and is updated by one thousand new
advertisements every single day.
He was not proposing to supplement the government's
employment data. He was proposing to replace it.
Then &
Now — The Full Scorecard
Here is how every idea across those three letters —
to a student, to a tech startup, and to the Chief Statistician of
India — maps
to reality in 2026. Ideas pitched to the government are marked ★.
|
His Idea (2013–2015) |
Written |
Proposed To |
Who Actually Built It |
Status |
|
★ Replace Annual Employment Survey with job-ad mining |
Oct 2015 |
Chief Statistician of India, MoSPI |
BLS (USA), ONS (UK), OECD — all now supplement
official surveys with job-ad data |
✓ Materialized — standard
methodology in 30+ countries; India followed ~8
years later |
|
★ Job-ad volume as real-time economic indicator |
Oct 2015 |
Chief Statistician of India |
Indeed Hiring Lab, LinkedIn Economic Graph,
Naukri JobSpeak Index |
✓ Materialized — monthly
reports cited by IMF, RBI, Federal Reserve |
|
★ Cross-correlate job data with IIP / RBI / GDP / CAD |
Oct 2015 |
Chief Statistician of India |
Federal Reserve (USA), OECD, World Bank |
✓ Materialized — standard
macro-economic tool in every G20 country |
|
Industry / city / skills visual dashboards |
May 2013 |
Rohini Damahe, LTIT |
Lightcast (ex-Burning Glass), LinkedIn Talent
Insights |
✓ Materialized — billions
of data points; real-time skill demand maps across 180+ countries |
|
Auto-fill JD from just a designation |
May 2013 |
Rohini Damahe, LTIT |
LinkedIn Recruiter AI, Workday, Eightfold.ai,
ChatGPT HR plugins |
✓ Materialized — standard
feature in every major ATS platform by 2024 |
|
Keyword frequency analysis across
500 million words |
2013–16 |
Rohini / Julia Computing |
Lightcast / Burning Glass Technologies |
✓ Materialized — $500M+
valuation; acquired by EAB in 2022 |
|
Designation ↔ education correlation rules |
Aug 2013 |
Rohini Damahe, LTIT |
LinkedIn Skills Graph, Coursera Skills Index,
NASSCOM FutureSkills |
✓ Materialized —
reshaping curricula at thousands of institutions worldwide |
|
Predict
which company posts which job, when |
2013–15 |
Chief Statistician / Rohini |
Eightfold.ai, Beamery, Predictive Index |
✓ Materialized —
unicorn-level companies; deployed by Google, Unilever, dozens of governments |
|
★ Data for education and policy planning |
Oct 2015 |
HRD Ministry, NSDC, NITI Aayog |
NSDC, AICTE, Skill India, World Bank India |
✓ Materialized — NEP 2020
partially shaped by labour market analytics |
|
IF-THEN
expert system from job data |
Aug 2013 |
Rohini Damahe, LTIT |
ML embedding models — BERT, GPT-family |
~ Partial —
same outcome, different method; neural nets replaced explicit rules |
|
★ National unified job portal with deep analytics |
Oct 2015 |
Chief Statistician of India |
NCS Portal, Govt of India |
~ Partial —
exists, but still lacks the analytical depth he envisioned |
|
Open Indian job-ad dataset for public
research |
2013–16 |
All three recipients |
Nobody — yet |
✗ Still open — the white space he identified in
2013 remains unfilled in 2026 |
|
★ Real-time salary benchmarking from live ads |
Oct 2015 |
Chief Statistician of India |
Nobody — yet |
✗ Still open — live city-wise, role-wise pay
benchmarks from job-ad data do not exist in India |
Score:
9 ideas fully materialized.
2 partially done.
2 still open.
★ 6 of the 9 materialized ideas were pitched directly to the Government
of India — and ignored.
The Detail
That Stings Most
Of everything Parekh proposed to the Chief
Statistician of India in 2015, the most striking is this:
He was not asking for funding. He was not
asking for a government contract. He already had the data. He
already
had the schema. He had been running the analysis —
partially, crudely — on his own websites for years.
He was
asking for collaboration. For the Ministry of Statistics
to lend its authority to a project he was ready to
begin immediately.
The Ministry did not respond — or if it did,
nothing came of it.
By 2018, the United States Bureau of Labor
Statistics had begun formally supplementing its own surveys with
data mined from online job advertisements. The UK's
Office for National Statistics followed. The OECD published
a methodology paper on it.
Today,
job-ad data mining is considered a standard complement to official
employment statistics in every serious
economy
India's Ministry of Statistics and Programme
Implementation — the very ministry Parekh wrote to — eventually
began
exploring similar approaches, nearly a decade after he
put the proposal on their desk.
The One Gap
That Remains
Despite all that has materialized, one idea from
those three letters has still not been built — not in India, not
anywhere in the form Parekh described.
A single,
open, longitudinal, publicly accessible Indian job-advertisement database —
clean, structured, freely
downloadable
— available to researchers, policy makers, journalists, and students without a
login, a fee, or a
data-sharing
agreement.
Naukri has it. Indeed India has it. LinkedIn has
it. None of them have opened it.
This is the white space he identified in 2013. It
is still white in 2026.
What the
Three Letters Tell Us
Taken together — the 2013 letter to a student, the
2015 letter to the Chief Statistician, the 2016 letter to a tech
startup —
these documents reveal something beyond a single good idea.
They reveal a man who understood, with unusual
clarity, that the most valuable economic data in India was not
being collected by the government.
It was being generated, voluntarily, every day, by
thousands of companies trying to hire people — and then
promptly ignored by everyone who should have been
paying attention.
He had no institutional backing. No research grant.
No IIT affiliation. No consulting contract. He had a database
he had built himself, a series of carefully worded
emails, and the stubborn
conviction that he was right.
He was.
The market proved it. The companies that acted on
his insight — Burning Glass, LinkedIn, Eightfold,
Indeed
became global businesses worth billions
of dollars. The governments that eventually adopted his methodology
improved
their understanding of their own economies. The institutions that ignored his
letters simply lost
thirteen years.
What Comes
Next
If the past thirteen years validated the vision,
the next five years may complete it. Here is what is already taking
shape:
Real-time salary benchmarking from live ads. Combining daily job-ad inflow with RBI and MCA data to
produce
live, city-wise, role-wise compensation benchmarks — updated weekly, not
annually.
Student-to-job matching at national scale. Every graduating student matched automatically to live market
demand, with a personalized skill-gap report
generated before their final semester begins.
Agentic job-description writers. AI agents that draft, refine, and post JDs
across all portals simultaneously,
from a single spoken prompt — exactly the "Wipro Business Analyst" scenario Parekh
described in 2013.
The open
Indian job-ad dataset. Someone will build this. It is too valuable not to exist. Whether it is a
government
institution, a university consortium, or a private company that eventually
opens its archive remains
to be seen —
but the case Parekh made in 2015 has only grown
stronger with time.
And then there is the most fitting possibility of
all.
Parekh, who turned 90 in June 2023, has built a
digital avatar of himself at www.hemenparekh.ai
— trained on
decades of
writing, thinking, and correspondence. The man who wrote to the Chief
Statistician of India with a
proposal to transform how the country measures its
own economy may yet become — through that avatar — the
oracle that
the next generation of policy makers consults when they want to understand
where the Indian job
market is heading.
He ended his 2015 letter with a characteristic
flourish:
"The fact that they have, so far, ignored this line of examination, will
work to your advantage — making you the very first person in the entire world
to come up with a Prediction Model in the area of Jobs."
He was offering the advantage to the
Ministry.
They passed.
The world took it instead — and built a billion-dollar industry on the foundation he had already
laid.
The original 2013 emails are archived at: marcomhcp.blogspot.com
The 2015 proposal to the Chief Statistician is
archived at: myblogepage.blogspot.com
Hemen Parekh's digital avatar: www.hemenparekh.ai
Tags: Labour Market Intelligence · Annual
Employment Survey · India Jobs Data · Data Mining · HR Technology · Predictive
Analytics · Ministry of Statistics · Hemen Parekh · Then and Now · Chief
Statistician of India
Related
References :
>
YOU
WILL LOVE THIS CHALLENGE.. ………………………… 10 Jan 2015
> Datamining
Big Data ………………………………………………. 23 Sept
2014
No comments:
Post a Comment