The short answer
Because applications are reach, not pipeline. A pipeline is people who can do the work, are open to moving and would accept your offer. Volume tells you only that your advert was seen, and triaging it costs a small team the hours it needed for shipping. Fewer, better conversations beat a bigger advert.
A VP of Web Development I spoke to last month has received more than 9,000 applications for a single entry level engineering role. If you are hiring Elixir engineers, this will be familiar. He has not hired anyone. If you are hiring Elixir engineers, or engineers on any specialist stack, that sentence probably lands somewhere between familiar and infuriating.
He is not doing anything obviously wrong. He runs a small, senior team inside a US healthtech business. He posted a role. The internet answered at volume, and the volume was worth almost nothing. This part of the market never shows up in hiring reports, because on paper it looks like abundance.
Why Is Nine Thousand Applicants Not a Pipeline?
Reach and pipeline are different things, and applicant tracking systems are very good at making the first look like the second. A pipeline is people who can do the work, are open to moving, and would say yes to your offer. Nine thousand applications tell you none of that. They tell you your advert was seen.
The cost falls hardest on the smallest teams. This engineering function is, in his words, three and a half people. Every hour triaging applications is an hour not shipping, so the flood does not merely fail to produce a hire, it slows the team that needed one.
Then the signal itself. He described applicants misrepresenting their identity and the depth of their technical knowledge, some of it AI assisted. His response has been live code review, sitting with candidates while they talk through their own work.
That is a sensible answer, and worth noticing. He has built a verification layer purely to establish that the person on the call did the work they claim. That layer did not exist five years ago because it did not need to.
Why Is the Engineer You Want Being Screened Out?
Three days later I spoke to a staff engineer in the American northwest. Twenty five years in the industry, six or seven of them full time in the same language.
He was laid off in May with about fifteen others, nearly all staff level, when contracts his employer expected did not materialise. Four months on he has six rejection emails and, by his own account, most look automated. At one company he applied for both a senior and a staff role and got two rejections, both naming the senior one.
Then somebody explained why. His CV described outcomes, what he had built and what it changed. It did not list technologies in enough detail. The advice was that what he had was good for the person he actually wanted reading it, but it would never reach that person. So he rewrote it as a list of tech stacks.
Now read that next to the nine thousand applicants, and resist the obvious conclusion. These two are not a match. One is an entry level opening, the other a staff engineer who was earning $245,000. Nobody here solves anybody else’s problem, and pretending otherwise is the tidy story about this market rather than the true one.
What they share is the screening layer, and it fails in opposite directions. At the junior end it produces volume nobody can verify, so a manager builds a live check to prove applicants are real. At the staff end there are few roles to begin with, and those are matched on keywords by software that never reaches the achievements. So a strong engineer rewrites his CV for a parser.
One end drowning in noise, the other filtered out as noise. Neither is a talent shortage, and neither is fixed by a bigger advert.
What Does the Exact-Stack Filter Actually Cost?
The exact-match requirement is the most expensive habit in engineering hiring, and it is expensive precisely because it feels responsible. The same US hiring manager had previously hired a senior engineer with no background in his stack at all, and that engineer ramped successfully. Stack experience, in his view, is helpful rather than mandatory. Few teams hold that position. Most write a specification describing the person they lost, then wait for them to reappear. That is the same habit that makes a job spec smaller than the market.
The cost is not abstract. It shrinks an already small pool, removes candidates who would be strong within a couple of months, and lengthens time to hire in a way nobody traces back to the requirement that caused it, because the empty chair does not send you a bill.
Meanwhile the engineers make their own calculations. One in Spain, who ran systems handling hundreds of millions of transactions a day, would accept a lower level and a lower salary to move into a new language. One in Dublin would take a boring job to work in the stack he cares about. These are people offering to absorb the ramp-up risk themselves, and few employers take them up on it.
What Should You Do Instead of Opening the Floodgates?
Three things. Define the smallest true must-have list. Not the wish list, the things someone genuinely cannot learn on the job in your environment. For most teams that list is shorter than the job specification suggests, and it is about depth of engineering judgement rather than years in a language.
Decide who does the ramp-up. If nobody has capacity to bring someone up to speed, exact-stack experience really is a must-have and you should price the role accordingly. If someone does, you have widened your pool at no cost.
Go to the people rather than waiting for them to come to you. Specialist engineering communities are small and relational. A VP I spoke to in Israel knows roughly four of the seven engineers working in his language in his whole country, and was surprised I knew one he did not. That is the real shape of these markets, and a bigger advert does not solve it.
Why Do Fewer, Better Conversations Win?
None of this is a criticism of anyone drowning in an inbox. Worth noting that the same business cannot yet open the staff role it will need, because that headcount is a business plan waiting on board approval. The senior end is thin not because nobody wants those people, but because the roles are stuck a layer up.
The signal has degraded and the tooling has not caught up. But the way out is not more applicants. It is fewer, better conversations with people who were never going to apply. The healthtech search written up here was run that way.
Questions this search answers
Why do thousands of applicants still not produce a hire?
Because reach and pipeline are different things. A pipeline is people who can do the work, are open to moving and would say yes to your offer. Nine thousand applications tell you none of that. They tell you your advert was seen.
What does a flood of applications cost a small team?
More than the empty seat. In the case described here the engineering function is three and a half people, so every hour spent triaging applications is an hour not shipping. The flood does not merely fail to produce a hire, it slows the team that needed one.
Are candidates misrepresenting themselves in applications?
The hiring manager described applicants misrepresenting their identity and the depth of their technical knowledge, some of it AI assisted. His answer has been live code review, sitting with candidates while they talk through their own work, a verification layer that did not need to exist five years ago.
Why are strong senior engineers getting automated rejections?
Because the screening layer matches on keywords. One staff engineer with twenty five years in the industry had written a CV describing outcomes rather than listing technologies, so it never reached a human. He rewrote it as a list of tech stacks to get through the parser.
Does requiring exact stack experience actually protect the hire?
It is the most expensive habit in engineering hiring, and expensive precisely because it feels responsible. The same hiring manager had previously hired a senior engineer with no background in his stack at all, and that engineer ramped successfully. Stack experience is helpful rather than mandatory.
If you are hiring on a specialist stack and the inbox is not producing, half an hour working out where the real pool sits is time well spent.
Arjun Gillard
Founder, AG Talent
agtalent.co.uk
