The algorithm did it

The internet has identified the latest suspect in the murder of the mid-career job search.

The ATS.

A British woman called Stacey Duguid recently went viral after sharing what it was like to be 52, unemployed and watching her savings disappear.

She has 30 years of executive-level experience in fashion and publishing.

She has also spent 16 months looking for work.

It’s a gut punch of a story.

Within days, headlines and social media commenters had solved the case.

It was algorithmic ageism!

The robot saw her age, rejected her and went back to plotting the downfall of middle-aged women.

Case closed.

Except nobody outside the hiring organisations knows what happened to her applications.

Could hiring technology have screened her out?

Absolutely.

Could a human recruiter have made assumptions about her age, salary or willingness to report to someone younger?

Yep.

Could the roles have attracted stronger candidates, been filled internally, disappeared during a restructure or never really existed?

Also possible.

And could parts of her job search strategy have worked against her?

We don’t know.

Her story proves that her job search has been bloody difficult.

It doesn’t provide a forensic report.

That’s why a new audit of a semi-automated hiring system caught my attention.

The audit was conducted by algorithmic-auditing company Eticas.ai and published this month as a preprint.

It examined roughly 497,000 candidate–vacancy pairings handled by Barcelona’s public employment service.

At first glance, the system looked fair.

Women represented 51.5% of registered candidates and 49.8% of recorded hires. Statistically, there wasn’t a meaningful difference.

Then the auditors opened the bonnet.

Women were less likely to be shortlisted for full-time, mid-paying and higher-paying positions. The largest gender gap appeared among candidates aged 46–55.

Yet the audit couldn’t identify what caused it.

Employers defined the requirements.

Human analysts turned those requirements into keywords and filters.

TalentClue searched and ranked the profiles using proprietary logic.

The analysts reviewed the results and built the shortlists.

Employers made the final decisions.

Nobody recorded the analysts’ searches. Nobody could inspect the vendor’s logic. Nobody documented why particular profiles were excluded. Final employer outcomes weren’t systematically collected.

The hiring process had all the accountability of a group assignment where everyone swears they did their bit.

However, candidates aren’t powerless bystanders in this mess.

Applying for poorly matched roles, sending generic resumes, burying useful evidence and relying exclusively on online applications will all make a difficult search worse.

Now that doesn’t excuse age discrimination, employer ghosting or crap recruitment processes. But blaming every rejection on recruitment technology lets everyone off cheaply.

You stop examining your strategy.

The recruiter blames application volume.

The employer blames the vendor.

The vendor blames the employer’s settings.

Meanwhile, nobody can explain where a capable candidate disappeared.

The ATS and AI have become the perfect scapegoat in all of this.

Both mysterious, they don’t return calls, and can’t write a furious LinkedIn post defending themselves. Although, I suppose AI can, if a human prompted it to.

But before we declare that the robot did it, perhaps someone should investigate the rest of the suspects.

~Lisa

Your Q&A

A question from LinkedIn

Hi Lisa,

You posted about applying for roles where you genuinely are a good match. What is a “good match”? I’ve only been applying to those roles I’m an 80%+ fit for.

Bob

​

Hi Bob,

The 80% rule sounds scientific, but professionals can be very generous markers of their own homework.

I’ve reviewed ads where someone believes they’re an 80% match, but they’ve glossed over a required qualification, specific industry experience or another criterion the employer considers essential.

A good match means you can do the core work and meet the genuine non-negotiables.

But let’s say every role really is a good match and you’re still not getting interviews.

That’s when I’d stop staring at the percentage and look at the rest of your application data.

• How old was the ad when you applied?
• Did you follow the instructions and send every document requested?
• Does your resume clearly prove the match?
• Did you follow up?

Being qualified is only one layer of a job search.

Using the 80% rule is a useful guide.

But it isn’t a diagnosis.

Thanks for writing in,

~Lisa