Anatomy of a Job Search - A Journey In Data

About this time three years ago I authored an article breaking down the data I accumulated over the course of a month-long search for my next adventure. That search ultimately led me to Tracer , but all good things -- especially startup things -- must come to an end. So, three months ago I once again found myself unexpectedly on the job market and I once again took the opportunity to detail my search in data.

In this, the resulting analysis, I hope to give some insight to other technologists on the search for their next adventure but also provide some insights for my friends in HR and recruiting. It's easy, from that side of the hiring desk, to lose track of the way the aggregate job market looks from the standpoint of a candidate.

Speaking of candidates, if you haven't been one in a while, boy-howdy have things changed since the "talent wars" of 2021. What was then a cut-throat race to recruit and retain talent at any price has transformed into a buyers market and nowhere is that more true than in startup-land. So, this search was very different from the one I conducted three years ago and what started out as a careful, considered, and highly-networked inquiry rapidly morphed into a bulk process driven by automation, AI, and process optimization. But that is a story for another time.

The Sound of Silence

In my 2021 job search I found that about half of my resumes were met with silence. The same is true today. If anything, the current talent market has left even more companies either unable or unwilling to respond to every application. This is... disappointing; while recruiting and HR departments have suffered some of the deepest cuts over the past few years, the rising availability of automation tools should make some kind of response possible, even if it is only notice that a position has been filled or taken down.

In both 2021 and 2024 I considered a resume to have "No Response" if I never received a rejection or interview request from the company. That long view of response rates is great for statistical purposes but doesn't tell someone in the midst of a job search much. So, while I was applying and interviewing, I started considering a resume to be "dead" after about two weeks. In 2021 this was a fairly reasonable position to take; the average response time then was 7 days and so, after twice that, a response of any kind was unlikely.

That's not true today. Automation has improved the responsiveness of the typical company (the statistical mode) but the enormous tide of applications has swamped many HR departments. That's pushing average response times out across the board but especially for companies which aren't leaning into HR automation. We can see that in the distribution of response rates. Despite the massive difference in application volume, the most common (mode) response rate in 2024 is two days faster than it was in 2021. But companies that responded slower responded much slower .

Yes, you're reading that right: one company took 12 weeks to get back to me. If you're in HR and reading this, the lessons here are pretty straight-forward: you need automation. The average turn-around time for a resume is about 14 days: double what it was in 2021. That's not surprising; many postings now receive several hundred applicants a day. Automation really is the only way to deal with it.

There is another explanation we should consider for that long tail in the response graph, however.

Holding Out For A Hero

Anecdotally, it seems like a number of companies are treating the glut of applicants as an opportunity to search for a "unicorn." Long gone are the days of "hire great people and see where they fit;" searches today are targeting a candidate that checks every possible box and has every desired skill. It's a buyers market out there.

We can see the impact of that in three statistical areas. First, as alluded to above, the long tail of rejections suggests that companies are going through multiple "classes" of candidates before finding the fit they seek and are holding on to applications longer because the positions are staying open longer.

Second, a higher percentage of resumes are being rejected outright. Indeed, this was the most significant difference between my 2021 job search and today's. In 2021 about a third of my applications resulted in an invitation to interview. In 2024, my interview rate has dropped to just 5.4%.

Third and finally, we can see the "unicorn" effect in round-one interview scheduling. While in 2021 a sizeable chunk of HR departments were setting up round-one interviews on the same day as an application, today it takes about a week-and-a-half to get a call together. That is probably just a consequence of the number of applications, but hiring managers are also scheduling round-one interviews late, some as much as a month after an application. This speaks to a long-running job search and, given the realities of the market right now, one with exacting requirements.

Drilling into that first few weeks we can really see how swamped HR and recruiting departments are. In the most common case companies can manage to line up a chat inside of the first week but the entire process has shifted right since 2021. As a consequence, there is a substantial first-mover advantage to be gained in those first four days.

Third party recruiters know this and act on it, setting the hook with candidates quickly. Pay special attention to the X-axis on the below image; the slowest third-party recruiter I worked with was still almost twice as fast to first-contact as the average internal HR department.

That speed matters if you're trying to land a unicorn. Even if the perfect candidate applies to your position, if you can't get them into the interview pipeline for three weeks, they may end up under time pressure with another offer. Candidates in this market are unlikely to be risk tolerant; while you're looking for a unicorn, they're looking for an income.

Call Me Maybe

Even after the considerable resume rejection and non-response rate, the most common outcome for interviews is still a "thank you for your interest" email. All total, about 2/3rds of my interviews ended this way.

The standout item here is obviously "Withdrew," which makes up almost a full quarter of interview outcomes. Of these, half were interviews still in-process when I signed with my next employer. The other half represent places where there should never have been an interview in the first place. Here there is another cautionary tale for HR and recruiting professionals.

There were three overriding reasons for these "interviews in error." Either...

1. ... my compensation expectations were wildly out of alignment with the range budgeted for the role.
2. ... the role required relocation but failed to indicate this requirement in the job ad.
3. ... or the hiring manager required a specific skill or skills which I did not have and did not claim to have.

All three of these can be largely addressed with better, more informative job ads. Being forthright about the expected compensation, location, and skill set allows candidates to self-exclude themselves from the applicant pool at the outset.

Fast Car

The time those unnecessary interviews take might seem trivial but it adds up. On average I found that I should expect to wait 41 days for a final decision on any application that received a request for interview with about 32 of those days falling after Round 1. That average conceals a lot of variation in terms of time to make a decision, number of interview rounds, and the spacing of those rounds.

It is tempting to imagine some correlation between the length of the interview process and a positive outcome but the data doesn't support it. Long-running interview processes were more likely to end in a forced withdrawal under time-pressure from an offer than they were to convert into an offer themselves.

As in 2021, it is difficult to overstate the importance of the first mover advantage. Of the four offers I ended up receiving, the quickest turnaround was in just 22 days after two rounds. That's just over half of the average time to act overall and just under half of the time I waited for my slowest offer.

Here again we see the "unicorn" effect in play. In 2021 the slowest offer I got took about the same amount of time as the fastest 2024 offer. Companies are able to move quickly when under pressure -- note the three interview stages in six days represented by the third row in the above graph -- but that proved to be the exception rather than the norm.

Don't Stop Believin'

If you've made it this far you have probably realized that I applied to a lot of positions over the course of the last several months. Some of these were, no doubt, the "fake" jobs we keep hearing about in the business press. Others were a stretch given my current experience or a down-market play, either in terms of title, compensation, or both. All of them, however, were positions I genuinely felt that I was qualified for.

I treated the applications process like a full time job, splitting my time between interviews, applications, and the development of increasingly sophisticated automation strategies to streamline the entire thing. All total, I applied to almost 1,000 positions sourced from numerous job-boards, professional networks, and third-party recruiters.

There's quite a bit of detail in that image and obviously a lot of job boards that I only used in passing. Nevertheless, I found that the best return on my investment of time came from Wellfound . Applications on Wellfound netted me an interview rate of 11.7%, beating out the overall average by almost three full points. Indeed, the only route with a higher success rate than Wellfound was a referral, and that with less than a percentage point advantage.

Those numbers are worth taking with a grain of salt, however. Referrals and Wellfound did not produce a statistically significant number of applications from which we can draw much in the way of inference. Jobright.ai 's edge over LinkedIn, on the other hand, while somewhat more modest, has the advantage of larger numbers behind it. The least successful job board I used was probably Indeed . Out of the 41 positions I discovered there, none resulted in an interview.

As far as the interviews themselves go, each round was a mixed bag. My interviewers made it clear that I was up against other candidates even late into the process. I had more than one "blessing" interview at the C-level after which I was told some variant of "everyone really liked you and thought you were very qualified but we're going with another candidate."

If you've recently been laid off or are still struggling through your own job search, "this too shall pass." Based on my own crude figures, hiring is picking back up. I saw a 20% higher interview rate in May than in April. Recruiter contacts are also up month over month. Hang in there and don't give up.

And watch this space in the coming weeks for a follow up article discussing the automation strategies and tools I used to both generate and track this data.