HACKER Q&A
📣 throwawaynay

Would recruiters be interested in detecting devs who are likely to quit?


I just had an idea, I don't know if a product like this already exists, if you know something similar or have feedback feel free to tell me please :).

It's apparently possible to detect which employees are the most likely to quit their jobs using public data and machine learning.

So I'm thinking of making two tools for recruiters:

-a saas where you'd input your criteria(technologies, experience, location etc...), and you'd get a list of top profiles who are the most likely to quit their current job but who are not yet on the market(or who are never publicly on the market)

-a browser extension for linkedin, which would give each profile a score rating how likely they are to quit.

I know linkedin have some kind of recommended candidates algorithms, but It seems that tons of recruiters are still wasting their invites/inmails on linkedin for a really low conversion rate, so I think this might interest some of them, I see people who got hired a month ago and who are still getting spammed.

+If you find a candidate quickly, you're less likely to pay for inmails, right? So linkedin have much to gain from showing you candidates who look good on paper but who are not that likely to answer, does my train of thought make sense to you?

I'm also maybe thinking of a tool using the same concept but for your own employees, so you can know which employees are likely to quit and really need to get a raise/promotion if you want to keep them.

Do you think those might be interesting ideas?


  👤 brudgers Accepted Answer ✓
Crowd wisdom is just people imagining what other people might think. What they say to you about what they think is constrained by social mores.

It matters what actual recruiters think.

More specifically it matters if recruiters actually pay money for it.

So the test is asking recruiters to pay money. If they pay knowing you have not built it yet, you are probably on to something. If they won't pay for it to be built, you probably are not.

Good luck.


👤 leros
I read something 5-10 years ago about some Netflix engineers who created a startup to do this. They looked at all sorts of signals to determine how likely an employee is to quit. I remember them looking at things like posting more on social media being an indicator.

Sadly, I haven't been able to find the company again.


👤 darrenwestall
My SaaS is used by 4000 recruiters every day and I’ve worked in the industry for 15 years so I feel I can chime in here :)

What you’re proposing has been done before, but it’s never been done well. For example, Chosen AI set out to solve this but ended up pivoting.

It’s still a problem that is interesting and I’d be keen to chat with you more on it if you’re open to it - if we can nail the tech, I know the route to market and can help kickstart it.

Recruitment/staffing agencies don’t tend to buy tech the way others do - direct sales is best, at least, at the beginning to get feedback.

TLDR. Recruiters will absolutely want this, but the execution is key.