So, under the old model, order 200k engineers applied to us. Because we were exclusive gatekeepers, around 3% of those were "accepted" onto the platform. Around 2/3 of accepted candidates received an offer, and around 1/2 of offers were accepted.

So you accepted 6K onto platform, about 4k got offers, and just under 2k accepted. Roughly a 1% conversion rate for those who applied? Yikes.

I don't know much about them, but the post does say their goal was to be "exclusive gatekeepers". If you consider it from that perspective, they had more like a 1/3 conversion rate (2/3 * 1/2). But conversion kind of seems like the wrong metric.

What metric would make more sense here, though? I can agree that from TBs perspective, they may not factor total applicants as their base # (although that is clearly their market). However this exclusivity of only using the small percentage they accept doesn’t seem to be a scalable business model. Having 1% of the devs who try their service actually accept an offer would not entice me to try their service.

>However this exclusivity of only using the small percentage they accept doesn’t seem to be a scalable business model.

For sure. That's very likely why they're making this pivot. Exclusivity was their goal, but it just turned out to not be a very profitable goal for them, it seems. I was basically just saying it seemed to be a "theory" issue rather than an "implementation/execution" issue.

No, the conversion rate is the fraction of companies who sign up which eventually hire someone.

I noticed that too, and when you combine it with "Triplebyte has hundreds of thousands of engineers on our platform" it means they aren't doing great with the rate either.

in the comments:

dev)

> I forgot how to implement the zig operation in a splay tree.

employee at Triplebyte)

> To put some hard numbers to this: further down this thread, there's a post about how "any engineer" could answer a question the poster thinks is too easy. I looked up the question in our back end and, in fact, barely a third of people who take our quiz get it right (the correct answer isn't even the most common one!).

So there 1000 job matches and at most (("hundreds of thousands of engineers on our platform" / 3) - 1000) who were incentivized to answer these contrived problems, did so correctly and still couldn't be matched.

So even among the large pool of engineers who have gone thru the process, met some arbitrary threshold of engineer-ness, there's still a huge mismatch between corporate/prospective employee expectations, that I'm not sure will be able to be overcome quickly even with these new initiatives, but it's interesting that they are being pursed now (not surprisingly after the shift in working environment after massive government restrictions on freedom uber alles).

I agree. We tried using TripleByte for hiring but what they screen for and what matters are entirely different. A founder we knew got an angry missive from one of the TripleByte founders because they’d rejected candidates during a final culture screen. Apparently the only qualification should be whether the candidate can do contrived coding tests and programming jeopardy, but whether you actually want to work with them is beside the point!

> Apparently the only qualification should be whether the candidate can do contrived coding tests and programming jeopardy, but whether you actually want to work with them is beside the point!

Yeah, different companies want different things and are ok with different things. My best experiences have been with companies that have taken less than 2 weeks to hire me with no testing whatsoever: just a couple of calls. Other companies want and only select for the contrived coding tests and programming jeopardy, and that's fine, but I want nothing to do with them at all.

I would be interested in using triplebyte if they explored something like this: https://news.ycombinator.com/item?id=27545446

Maybe even allowing 3rd parties to bet on specific hiring decisions (in addition to what is described in the link) could add another layer of incentivized feedback loops to the process (esp if the people making the hire/fire before 3 months are made known) and add another layer of market driven signals for them to pay attentions to. Maybe make it more expensive to chase certain types that are in higher demand, and less expensive to chase others types that are in low demand (not the salary to the employee, just the payment for filtering). Maybe less expensive for shorter refund windows, and more expensive for longer refund windows.

yea, I went through their process on the employer side, and my guess is an average hire is 10k, that means 10m in revenue total

They work on a retainer basis, not a commission basis, so you really can’t predict income like that. See other comment: https://news.ycombinator.com/item?id=27543931

When I went through it was like, ~1k/month and somewhere from 5k-15k/hire over the top. Industry average is around 10k/hire and their pricing suggests they're probably banking that amount with pricing used to encourage loyalty and appease investors with the Saas narrative. They're not pulling in much greater than industry average, and even if they're double, its only 20m, which is still bad.

Especially when they're probably making on average less than $25k per engineer they place. I can't imagine they're making even close to $100k on average.

$25k*10k = $25M / 10 years = $2.5M / year. That is peanuts compared to how many people they employ. Plus the ads they run...

How much is a referral worth to companies? $1,000? If so, their revenue is only ~$1m, but they've raised $48m.

I have no idea what Triplebyte gets or even what their business model is, but typical recruitment companies get a lot more than that -- on the order of 15-20% of the first year's salary, so on the order of $15-25k. I suspect Triplebyte is less than that, but could be an order of magnitude more than you've guessed.

2 months salary, like an engagement ring ;-)

When I last used headhunters they charged 20-30% of a years salary. So they are likely making $25-50k each. So that could be as high as $25-50M