Applicant tracking system are crushing workers dreams of getting hired

It’s 2011, and I’m in the corner of a musty coffee shop in a Hilton hotel in Northern Virginia. And I’m losing hope. Having just figured out the science of our generative AI platform, one of the first of its kind, and with a content deal looming with Yahoo Fantasy Football, our startup needed to hire a dream team to make the AI part happen at scale.

I led the AI side of the company, and I was failing at hiring my dream team. There were no data scientists, there were no machine learning wranglers, there were no large language model enthusiasts – at least any who wanted to work for a startup that planned to make money by teaching computers how to write sports articles.

I had just bought breakfast for the head of stats at a well-known sports media network – my current angle of attack being to find and hire “Pete” from Moneyball (stream it). He didn’t get us. There was smirking.

As I got in my car to drive the four hours home, and yes, I’m using the last of my shoestring budget to find this unicorn, I told myself I’ve got one more shot, a kid just out of college who was a self-proclaimed sports and numbers nerd, who was waiting tables at a local Chili’s.

I mean, you saw the title of this article – you can pretty much guess what happened next.

Everyone Gets Fired

No, not at our company. I hired the kid, built the team, we crushed Yahoo Fantasy Football and NFL.com fantasy football, and then went on to work with the Associated Press and others, generating not just sports articles but all kinds of articles, and we exited to a private equity firm four years later.

That was fun.

But before you label me an old-head calling for a return to the days when we found talent needles in haystacks by using our bare hands and elbow grease, I want to tell you one more quick story, about how half of an entire HR department got fired last month for using automated candidate screening, poorly.

Short version: Company needs an Angular developer. Goes three months without finding a single qualified candidate. Company’s HR manager decides to submit a tailored resume for the position. Gets rejected immediately. Discovers automated candidate screening had been programmed to search for “Angular JS,” a framework that was discontinued 14 years ago.

No. No. It’s not the typo. It’s the method.

Why My Chili’s Guy Worked

This column would not be worth your time if I hadn’t hired the Chili’s waiter. And it would be funnier if he had been working at Chi-Chi’s or Flinger’s, but I can’t change that.

You need to know that I am not an HR person and that I have very few HR skills. So I’m not going to solve all the HR problems in 1,000 words with some jokes. I’m going to do what I normally do. I’m going to connect the dots on how the rise and reliance on automated candidate screening – part and parcel of applicant tracking systems – is betraying both the talent getting screwed over by it and the company using it.

Maybe it’s time to make some changes.

The Chili’s guy worked because it forced me – company leadership – to be not lazy.

Now, this exact scenario probably won’t happen to you often, if ever. It rarely happens to me. But having to basically invent the skill set of the person I needed based only on the results I desired, made me realize how little work goes into actually defining the qualifications for the role itself.

We basically take it off the company HR shelf or copy one from the internet, right?

Thus, we put garbage in, we get garbage out, and we use AI to filter the garbage into better garbage.

This, not a perennially overworked HR department, is where the problem begins.

We Need Some of That Sweet, Sweet Tech

More specifically, the problem begins when company leadership decides they need a skill they know very little about. This happens most often in tech, but it can happen with every flavor of talent.

At a startup, you’re basically wearing all the hats, much like me scribbling pseudocode into a notebook before sitting down to bad coffee and toast with a smug nerd who was still dreaming of becoming Scott Van Pelt someday.

And also like me, you can’t wait to hire people to take some of those hats off your head. But that doesn’t mean those hires can take on the knowledge responsibilities as well. 

So in today’s world, the CEO decides that the company needs to be on the AI gravy train, but doesn’t care how they get there. They tell the CTO to build a team, who mandates it to some tech lead, who wants nothing to do with it, and thus pushes the request to “someone in HR.”

That’s your first mistake. 

You Can’t Get What You Want Until You Know What You Want

You know, I’m not a huge Joe Jackson fan, but dude wrote some bangers.

With time of the essence and no real overarching philosophy in place, everything except the actual necessary qualifications goes into a two-page job description that reads like a vague self-help manual written by a cult leader.

They need a self-starting, team-player who has vision, is strategic, but is not afraid of the execution. Someone who can drive innovation while ensuring alignment with business goals and outcomes. Someone with demonstrated expertise in the technology, strong understanding of the business, and a passion for mentorship and leadership of cross-functional teams to encourage competitive differentiation. 

That is 1) the biggest word salad I’ve ever tossed and 2) taken almost directly from the first job description I found on LinkedIn. 

But wait. There’s no AI in it. At all. 

What do we do?

Got it! We’ll stuff the rest of the job description with AI buzzwords. Doesn’t matter how. Just have someone in HR do it.

Then they program the ACS on the ATS and hit the “OK Find Me a Unicorn” (OKFMAU) button.

That’s why your resume isn’t getting seen by a human. And it’s actually the fault of many of those humans.

Can’t Find Needles? Make the Haystack Larger!

You don’t have to be a data scientist to see how quickly this metastasizes into a full-blown carnival of hoop-jumping recruiting processes, painfully mismatched hires, sound and fury signifying nothing, reductions in bloated workforces, and ends up with truly talented people on the sidelines not getting a sniff from companies that have thorny and growing problems getting anything done.

The Chili’s kid was proficient enough with data and passionate enough about sports to learn whatever he didn’t know provided I could get him across the finish line. And it’s not that no one has time to coach anymore so much as it is no one knows what they need out of the player in the first place.

Someone in HR will figure it out. And the ATS will find it. With keywords.

Today, my biggest struggle with new hires is getting them to unlearn everything they think they know about how to get stuff done. Because what they know got them the job, because the job description required that they know that. 

It’s infuriating. I’d rather go back to the days when I was driving through Richmond, punching up the digits of a phone number of a waiter who loved sports and numbers, and deciding it was up to me to teach him what was eventually going to make us successful.

INC