[0xbadcafebee]

Somebody please make OkCupid for Jobs. Both employers and applicants fill out questions that the community curates (so they can be job-specific, industry-specific, etc). Both employers and employees get matches via algorithm rather than receiving 100 random resumes or 100 random job suggestions that have nothing to do with you.

Existing sites do this, but very poorly, because they only try to detect information out of a resume and detect keywords in a job listing. There is only so much data you can gleam from those sources and it's highly variable. You need to ask specific questions, like "Is Functional Programming the best form of programming?" or "Do you prefer asynchronous communication when dealing with coworkers?", or "Do you like being in an office?", or "Do you like working in finance?".

[BrianOnHN]

Great idea. Being in the job board business for a decade, here's what I think would be the challenges:

1. Organizational structure mismatch: The people posting the jobs don't know the answers to the required questions. Outside of startups, job postings may be drafted by a manager, but a committee, legal, and/or HR has the final say on what's in the posting. An assistant or VA is often the POC when the job posting is copy&pasted into the job board. This can be overcome, but it's challenge to try to reeducate the organization.

2. Chicken-egg Problem: The challenging part here is that your racing against many existing players with large audiences, and hoping that they can't reverse-engineer the solution before you catch up.

3. User Habits: I can't think of any "normal" person I know that looks for jobs without some negative catalyst pushing them into it. Meanwhile, I imagine that single people are incentivized a bit differently. Not that this wouldn't work, I'm just not sure if it would work the same as for dating. For example, "why would I feel motivated to complete a long profile when I can just spam my pre-made resume, since it will be required at some point in the future anyway." There are good answers to that question, but now you're entering the same re-education waters as in challenge #1.

[feoren]

> The people posting the jobs don't know the answers to the required questions.

That is a feature, not a bug: it means that candidates know that jobs they see posted on that site have actually been at least looked at by the actual team you'll work with. Filtering out jobs offers that are just HR buzzword soup would be exactly its competitive advantage over Indeed and LinkedIn job-spam.

> hoping that they can't reverse-engineer the solution before you catch up

Don't worry; they can't. Kodak couldn't even pivot to digital and they invented digital photography! Huge companies have awful, terrible, very bad, no-good software, and leadership that has no interest in ever adapting to anything. Besides, the niche is different anyway.

> "why would I feel motivated to complete a long profile when I can just spam my pre-made resume, since it will be required at some point in the future anyway."

Because you're catering to applicants who are picky about what job they're looking for. I don't just want any job -- the list of jobs I'm qualified to do is vastly, vastly larger than the list of jobs I'd ever want to do.

You're not trying to replace Indeed: you're trying to replace recruiters.

[BrianOnHN]

100

> hoping that they can't reverse-engineer the solution before you catch up

>> Don't worry; they can't. Kodak couldn't even pivot to digital and they invented digital photography! Huge companies...

Yes, b/c Kodak taught us that, better put: hoping that they can't acquire a better connected imitator before you catch-up.

> You're not trying to replace Indeed: you're trying to replace recruiters.

This is spot-on.

My vision is like "talent agents for the rest of us." There are additional challenges that I won't mention because their knowledge is a competitive advantage at this point. But yeah, that's what's up.

[ineptech]

4. As a hiring manager, I'd need different "profiles" for each team (one team mobs 24/7, another almost never mobs, etc)

5. What I would want as a candidate (a blunt and specific "warts and all" description of the team I'm joining) is not what employers would provide (vague, HR-y generalities like "we use cutting-edge technology" and "we prioritize work-life balance"). This was what turned me off from KeyValues.

But I hope these are tractable, because I very much like the OKCupid model and would use the product if these could be solved.

[BrianOnHN]

Good additions. I'd like to add that these are all challenges that will be faced by all competitors of the existing model.

Personally, that's what lead me to focussing on what might be the most innovative direction I can fathom instead of the incremental direction (aka "this for that" innovation). If I didn't recognize that all of the challengers would face this same uphill battle, then I myself would've built "OKCupid for jobs" by now. It would 10X the pure-spam nature of the current market if the execution was done well, but current players are too big. To conclude, you have to outthink them and create something that can't be reproduced with effectively unlimited budget.

[rglullis]

> The people posting the jobs don't know the answers to the required questions.

The questions should be answered by all the devs in the hiring team. You don't care about the compatibility between HR and employee, you care about compatibility between your peers and your direct boss.

> I can't think of any "normal" person I know that looks for jobs without some negative catalyst pushing them into it.

You can also solve it by making the match between team members. This way, companies could even ask their currently employees to participate, which means that people would be using the website even if they are not looking for a job directly.

[BrianOnHN]

I think you might be on to something.

What might a B2B employer-only network look like?

[giantg2]

A network of employers colluding on employment would look like an antitrust violation.

[inasmuch]

Yes, please.

I'd also like to see cover letters replaced by something more useful for both parties, like a small set of direct questions to answer.

When an applicant applies for a job, she's left to make her best guess as to what information the company might find useful, and has to find a way to cram that into an awkward, stiff, rather bullshitty epistolary format. The hirer is then left to try to parse and interpret that to arrive at some guesswork notion of whether the applicant might be a fit.

If the hiring company were to instead come up with a list of guided questions, they'd communicate and request up front what information they need to determine whether a candidate is worth meeting (and thereby would be forced to figure this out themselves before beginning interviews—an important step often skipped). Seeing these questions, the applicant would first make a quick judgment on whether she's interested in a company or role that wants to know these things, and then if she is, proceed to answer them, demonstrating her ability to communicate in a much more contextual, engaged fashion, versus modifying her own cover letter template or laboring over the blindfire of a custom one.

I see only advantages to this approach. It helps companies avoid an inbox full of inappropriate applicants who didn't even read the job description, and it enables applicants to learn more about the company and role up front while breaking the monotony of the application grind and having a greater opportunity to put their best feet forward by applying their personality, knowledge, and experience contextually, rather than packing it into a shotgun shell and blasting the job board with buckshot.

This may ask for more of an applicant's time than the current approach, especially in an era where every job site is trying to make applying into a one-click process, but as someone who has spent a truly unbelievable amount of time and effort looking for jobs (and a fair amount of time hiring), I can confidently say the dragnet application approach is ineffective, inefficient, and demoralizing compared to preparing fewer, better applications for jobs that actually interest you.

[memling]

> I'd also like to see cover letters replaced by something more useful for both parties, like a small set of direct questions to answer.

This "small set of direct questions," in my opinion, is implicit in the job description/duties and desired qualifications. Employers lately have had challenges in job definition, of course (they seem interested in advertising for the company rather than a job, and the job descriptions have become increasingly elastic; in their defense, retention is hard, and it can be difficult to slot people into single roles anymore).

My general practice in responding to job postings is to prepend my resume with a table containing the job requirements vs. my experience with an estimation of whether I meet the requirement (check, dash, X) and a brief explanation.

In cases where there aren't any job postings, but I have an interest in the company—or if I want to make an impression that I'm targeting them—I'll send a cover letter by postal mail with a printed resume. Usually it's a few sentences: here's why I'm interested, how I might be useful to you, and here's a resume with contact information if you want to reach out.

[inasmuch]

> prepend my resume with a table containing the job requirements vs. my experience

That's an interesting approach. But what I think gets somewhat lost in this, and why I don't quite think a job description serves as that set of questions, is that there's so much more color and character (culture?) that can come through if the questions are good.

For example, a bad question, or something that should satisfied by the JD and résumé, would be "Do you have experience doing X?" If that's important, just say so. A better question might be "Why are you interested in building X for Y industry?"

I think résumés and job descriptions as they're currently written do an okay job of checking off raw qualifications. What's missing is clarity and direction for the less quantifiable bits. Done right, it could be like a low-pressure, no-expectations first interview in lieu of a crap email that's tedious to write, tedious to read, and likely only helps a hirer understand whether the applicant can write a good cover letter.

[memling]

> A better question might be "Why are you interested in building X for Y industry?"

If the JD is building X for Y industry, I'd think a cover letter might address this. Sometimes, of course, we're in the position of doing a bit of spray-and-pray when it comes to application...but generally I try to make sure the cover letter justifies sending the resume at all: here's my interest, your needs, and some more information if you want to read further.

> What's missing is clarity and direction for the less quantifiable bits. Done right, it could be like a low-pressure, no-expectations first interview in lieu of a crap email that's tedious to write, tedious to read, and likely only helps a hirer understand whether the applicant can write a good cover letter.

I think you're basically right here: the interview process is challenging because it's a bit like a first date and you really want to know what the marriage is going to be like. That's why, I think, some companies have shifted away from being very clear about what they're asking for and more towards trying to find personality, adaptability, and retention fits: it maximizes their ROI by ensuring they get people they like, who can change as needed, and who will stick around. The challenge for applicants, of course, is that they no longer really know "What am I going to do?" and "Am I qualified for this?" or "How much do I need to learn?" There's a lot of comfort in competence!

I wouldn't underestimate the value of communication—including the ability to write a cover letter. I resented, at first, that I found myself in more forward-facing positions in my career, because I wanted to spend more time in technical work. But it's been a suitable way to get paid while enjoying time programming, too.

[altdataseller]

What is the response rate to those postal mails you send? And who do you address them to (do you find the name of the recruiter or just address it to the general company)

That’s a very interesting tactic.

[memling]

> What is the response rate to those postal mails you send? And who do you address them to (do you find the name of the recruiter or just address it to the general company)

I don't do this often, so I'd be wary of reading in too much. In each case, I got an interview and one job, though that took awhile. (The hiring manager did not comment on the letter, but rather on the table at the top—probably because I couldn't find his name.)

I try to find the appropriate person to contact through LinkedIn, ideally a manager in the department of interest; I wrote to a (small) company president once, too. I try to avoid recruiters if I can help it, though my bias against them is perhaps a little uncharitable. The goals are two: to sidestep the usual application process through a little elbow grease and to differentiate myself in the candidate pool. I'm probably a pretty average programmer, and this has been a useful tool for me.

In the past, cover letters were usually a weak point because I would use boilerplate language that comes across as a bit canned. I've had the luxury lately of not having to do this: the letters were all written for companies I thought I really wanted to work for, so the interest was genuine and (hopefully) came across that way.

[ladybro]

KeyValues does just this: https://keyvalues.com

"Find engineering teams that share your values"

(No affiliation, just a fan)

[BrianOnHN]

As a job board owner, I'm also a fan of keyvalues.com as an example of innovation in the market (no affiliation).

That said, I don't see it ever scaling to the mainstream. It serves a great niche, though!

[bag_boy]

What’s your job board?

[yellow_lead]

I tried to use this site before but there seems to be very few companies and little to no validation of the values (most are subjective as to whether a company meets them).

[robertlagrant]

Wow - I love it. That's how I'd hope to build a job site. Love the long form company descriptions. Will have a think about DDG's idea of a "DRI".

[rglullis]

Oh, man. Where were you 15 years ago?

Around 2007, when there was this wave of "37signals job board clones" and I was still living in Brazil, my friend and I started a job board focused on the Brazilian market (job4dev.com). It was free to post as long as the name of the company was provided and the contact was direct with someone from the team. No agencies and no clueless recruiters allowed.

At first, there was very little traction. We were doing most submissions ourselves. After a few months it started to pick some audience (a few thousand unique users/day). We did all the "Web 2.0" things from the time: auto-classification and filtering by tags, integration with social networks, a company wiki, etc.

Our site was featured on some tech magazine as one of the best sites for tech jobs. We had a couple of "repeat customers", companies that were posting regularly on the site and were giving excellent feedback. The one thing that we "forgot" to do: sales and marketing. By 2008 I was moving to the US, right before the big crash, so I was focused on finding any kind of job which could sponsor a visa. My friend on the other hand was more focused on his day job, and he was just too afraid to go to companies that were facing a recession to ask "maybe you'd be interested in some paid features"?

I continued moonlighting on job4dev for some years, but because I was in the US and because much bigger companies started cornering this market, it became mostly a hobby project for me.

All of this to say: guess what was one the things that I did start to prototype?

Yes, it was "OkCupid for Jobs".

The algorithm is actually simple to implement. I didn't take it further because I pitched the idea to some and and no one "(in Brazil) knew of OkCupid, and few seemed to care about using "algorithmic compatibility matching". Credentials/Education Level/Networking were the most important filters.

Anyway, after a while I left the US to move to Germany, and put aside job4dev for good. The idea of "OkCupid for jobs" still seems to me a good one, and I still wonder why it wasn't picked up by someone more competent than me.

[0xbadcafebee]

Maybe if somebody gets bored enough they'll start it as an open-source project, fund hosting via donations?

I think the key is the right vision to make it work. I see replies of "I did the same thing!" but they seem to be missing the killer feature: high-quality data. Good data (and lots of it) enables better algorithmic matching. To get the data, you need gamification. IIRC, OkCupid grew fast because bored people would fill out fun surveys, and that generated tons of data. So a big part of their site was probably just dedicated to "how can we convince people to answer all these questions and submit new ones?", and that's much harder than just creating a job board.

[danielvaughn]

What I'd really love to see is a way to find candidates via their former employers. As a CTO who recruits in a specific industry, I need to find employees who used to work at companies A, B, C.

[warrenm]

sidebar - if your hiring criteria demands people who've worked at A,B,C companies before (either one, multiple, or all) ... you're artificially limiting yourself to a tiny portion of the potential candidate pool

[warrenm]

Doesn't LinkedIn search more-or-less accomplish this? Look for people who have worked for J, Q, and U companies?

[Mezzie]

This is very industry dependent. Some industries make heavy use of LinkedIn, but others never touch it.

For example, if they need educators, LI is pretty useless.

[htrp]

Not very well.... even when you pay for their recruiter packages.

[danielvaughn]

maybe I haven't looked hard enough, but I find it to be pretty difficult to search by that dimension

[stanislavb]

I’m working on building a generic “reverse job board” for developers - “LibHunt /DEVs” https://www.libhunt.com/devs. I’d be more than happy if you give it a go.

[ahmed_ds]

I like this idea.

I have posted on whoishiring's who wants to be hired threads and each time I got half a dozen cold emails from generic job boards startups. The entire industry needs to take a step back and evaluate their entire process.

The process of evaluating a candidate is technical skills first, personality test second. If we flip the entire I wonder how effective recruiting could be. At least for entry level jobs, I am not sure why technical skills are prioritised over personality, enthusiasm and adaptability of a candidate.

[giovannibonetti]

Maybe because companies that grow often delegate hiring to the HR department, which can't really filter for technical skills (only buzzwords in the resume). So, it's no wonder HR folks end up using generic personality tests to select candidates, at least in the early stages of the hiring funnel, before the engineering team gets involved.

I'm not saying personality isn't relevant for hiring, but the people conducting a hiring process can (perhaps inevitably) add bias to it, knowingly or not.

[jackcosgrove]

OkCupid succumbed to the swiping apps, and I think a similar thing would happen except instead of swiping based on photos you would swipe based on salary.

[hunterb123]

OKC and Tinder-apps serve a different purpose.

You use Tinder when you don't know what you want or your wants are superficial, you use OKC when you know what you want/don't want.

Job searches would be similar. For those who don't know what they want, the salary is the pretty face. For those who have suffered from bad relationships, they have a checklist.

[dr_kiszonka]

I wouldn't mind swiping through job matches generated by an OkCupid-like algorithm. I am guessing people would swipe based on what they care about and it is fine. For me, the salary and the mission matter equally as long as the salary is above some reasonable threshold.

[bglusman]

This was the original vision of the founders behind recurse.com (formerly Hacker School) when they applied to YC back in 2010...

https://www.recurse.com/manual (citation using this exact term)

[rsweeney21]

Shameless plug: I started Facet (https://www.facet.net/passive-search) to do just this. You set your preferences for jobs you'd be interested in (salary range, tech stack, location/remote, etc) and our algorithm matches you to jobs that fit your criteria as they come up. We're adding more and more criteria as we go to get the targeting super precise.

I'm a software engineer and I noticed that LinkedIn's business model was pay to spam. They charge $3 to send a message to a candidate, so their incentive is to have recruiters send as many messages as possible so they don't want to make it easy to target. Our business model is to charge per hire, so we try to be as efficient as possible in matching people to the right job and send the fewest candidates we can to hiring managers.

We require companies to give salary and other key information in the job postings too.

We're bootstrapped and still early, but it's' been an incredible ride so far!

[dgfitz]

It would be nice to be able to set some filters and browse before being forced to make an account. I was intrigued but you’re probably losing people like me by introducing that kind of friction. I understand why you’re doing it, I really do, just some hopefully helpful feedback.

[alexeldeib]

Were you not able to? I clicked "Jobs" in the menu bar and was able to search without an account.

[hash872]

I'd be interested to know what makes your business model different from Hired.com, which was basically the same thing, right? Just curious as someone who follows the space

[nameisname]

This already exists.

https://www.plum.io/

[balls187]

Doesn't OKCupid use algorithms for matching?

[NiagaraThistle]

This is a really good idea.