Excellent thread by @arnav_kumar on why job classifieds are sticky. Adding my learning below as an investor and board member in Talentpad <thread>
Context: @jainmayank_943, @nikhilvij03 and @jainraghav17 started Talentpad in 2014 with the aim of making hiring smarter. They parsed millions of resumes to find the best fit for specific kinds of jobs.
Their algorithm could figure out that people who have worked for 5 years at IBM with a title of Senior Software Engineer made excellent Project Managers for Accenture, for example. They called it data science,but if they had called it AI they could've raised at a higher valuation
When I first saw the product, I was floored. It was magic - they could find the 20 best resumes for a particular job, not just sort available resumes for a fit. This allowed them to *be the recruiter* rather than a classifieds platform or matching software.
This part was a big deal, because like @arnav_kumar says, recruiters close the transaction, not just supply a lead. I had concerns about scalability, but was willing to bet that the founders will figure out how to automate many of the processes and be a tech-enabled recruiter.
The recruiter made most of the money in this value chain. For say a software job with a 20 lakh salary, a recruiter makes 1.6 lakhs for a single hire. If you could reduce the people-dependency through technology, you could scale this very, very fast.
So we gave them a seed check, and they were away.
The business scaled faster than I expected. Within 9 months of funding, they were a doing a $1M run rate, and barely burning any cash. Their clients were large startups and MNC product companies, and they loved them!
The business scaled faster than I expected. Within 9 months of funding, they were a doing a $1M run rate, and barely burning any cash. Their clients were large startups and MNC product companies, and they loved them!
The product improved very fast. The matching was so good that the recruiter had to do very little manual work to source the candidate, and their conversion rate was 3-5x of an average recruiter. We truly had product/market fit.
Yet, in the middle of 2015 the founders and the board decided it was not working and we needed to shut down. We did not think a $1M business with almost zero burn was viable. It was the boldest and most intellectually honest call I've seen founders take in my career.
The issue was this: their product kicked ass in product companies, but they found no traction with mass recruiters. IT companies, BPOs, banks, retailers, insurance firms: these form the backbone of India's white-collar workforce. And none of them were biting.
Initially we thought these are slower-moving, traditional companies and that they will eventually convert. But there was another troubling sign: the product didn't work for mass recruitment.
The same data science engine that could find the best candidates for a product manager position at Amazon was little better than chance at predicting the right profile for a bank teller job. We thought it was a problem with the product, but the truth was quite sobering.
The fact is that, outside the rarified world of software product jobs, a resume matters very little for recruitment. An IT company mostly hires fresh graduates on the basis of tests. A BPO based on your ability to speak English. Retailers and banks based on your "attitude".
An auto company recruiter told me they find recruits from a particular state very hardworking, something possibly illegal but nevertheless a fact of recruitment in India. None of this information is in the resume.The best data scientist in the world can't sort on these attributes
That's why human recruiters are needed for most jobs. The judgments they're making are not arbitrary - they can be coded in software. But there's no data to apply the science on.
Talentpad didn't want to create a 1000-person recruitment company. They wanted to build a tech company, but the tech only worked for a very small number of jobs. When they couldn't go big, they decided to go home.
That, I'm afraid, is why this industry is still stuck with classifieds. Because "closing the loop" requires thousands of skilled people. A platform can feed resumes to these recruiters, but doesn't have enough data to replace them (yet).
Postscript: @jainmayank_943's family couldn't believe that he was shutting down a 6 crore revenue business! Who does that? Ambitious entrepreneurs, that's who. <fin>
(hope you don't mind me sharing this, Mayank 🙂)
(hope you don't mind me sharing this, Mayank 🙂)
