Reflections on the Revolution in Job Applications

A long absence, I know. I can only apologise. The book, other work, a baby-now-toddler: they all pushed this newsletter down my to-do list. But I’m back now–don’t worry.

Three Bits of Housekeeping

Book: The book is available to pre-order. Here is the Waterstones link: Is This Working? The Jobs We Do, Told by the People Who Do Them. It comes out on 6th March. If it’s not your thing, then buy it for someone special. It’s a gift that says, ‘I’m concerned about your relationship with work, and I would like you to consider these other careers.’

Substack: I’m also going to publish these newsletters on Substack, so if they keep getting held up in your spam, try subscribing there.

The future: I have been applying for PhDs in History. If I get funding, then I will start in September. A forty-year career as an academic historian awaits–I hope. If I don’t get that funding, then reply to this email and tell me what I should do with my life. 

Reflections on the Revolution in Job Applications

My PhD applications took a roughly similar form, with ‘form’ being the important word. They each consisted of three parts:

Application number one: the ‘General Application’ for a PhD in History at University X, comprising:

• CV;
• cover letter;
• research proposal of c. 1,200 words;
• writing sample of between 2,000 and 5,000 words;
• two references;
• a form on an online portal, where you detail your skills, experiences, and the training that you will need to undertake during the PhD. 

That General Application gets you into a university, but it doesn’t get you any money. For funding, typically, you need to make two further applications.

Application number two: the internal scholarship applications for University X. 

• lay summary of research proposal c.200 words; 
• research proposal with higher word count, c. 2,000 words;
• statement of purpose and institutional fit;
• application form, where you describe skills and previous experience/achievements.

Application number three: the scholarship application for government funding, which is awarded by regional consortia.

• summary of research proposal c.150 words;
• research proposal with lower word count, c.750 words;
• form on an online portal describing: (i) training needs; (ii) your strategy for knowledge exchange and impact during your PhD; (iii) how you will benefit the consortium and the consortium will benefit you.
• one or more references tailored to the consortium’s requirements. 
• institutional statement of support from proposed supervisor(s) at chosen university within the consortium.

If you were to apply for seven PhDs, you would repeat this three-part process seven times. The people in-the-know tell you to apply to as many universities as you can, because postgraduate funding in the humanities has all but dried up in the last twenty years. You cast a wide net over a shrinking pool. 

When you write these forms, you can’t help but think of all the work you are creating for others. For your referees. For your potential supervisors, who have to write statements of support. And for the panels of academic staff who will read and discuss your application. In this way, the applicant is a machine for generating administrative labour. 

As recruitment processes go, the one for PhDs is par for the course in the modern knowledge economy. Or perhaps better than par. Others certainly have it worse. At most, I will have one interview per university, rather than the three rounds of interviews that are becoming the corporate norm. Nor did I have to do a psychometric test, a coding exam, a two-week placement, a marketing pitch, or a video interview with a computer. I got off lightly, I suppose. But it’s enough to make you think: if several hundred pages of application forms counts as getting off lightly, what does that say about the job application systems that we have created?

I suspect the growing length and complexity of applications is a product of technological change, a disease of modernity like obesity and dementia. The pattern looks something like this:

1. Applicant numbers track the supply of labour.

2. A technological change makes applying for jobs easier. 

3. Applicant numbers per job increase significantly (whether or not the actual supply of labour has increased), 

4. To manage the increase in applicants, employers make job applications more onerous.

5. The time-saving effect of the technological change is swallowed by the more onerous new applications. 

The technologies of the internet are the clearest example of this pattern. Job sites, emailed applications, online application portals, and so on. These technologies lower the effort required to make a job application. There is no posting of CVs, no calling an employer to ask for an application form, no time spent looking through adverts in different newspapers. Instead, you click through to an online form, copy over your details, and submit. Before long, employers suffer a great deluge of applications, from all over the UK, and from all over the world, with applicants asking ‘do you sponsor visas?’. 

The employer is in a tricky spot. They have a pile of CVs and cover letters that is far taller now than it was several years ago. They have more candidates in front of them who are ‘good enough for interview’ by their existing standards. This looks like a good thing, but it isn’t. The typical application is less serious than it was before: each application represents a smaller proportion of an applicant’s total job applications, and a typical applicant is therefore less likely to accept a job offer. 

What should the employer do? If they interview the same number of candidates as they used to–before the applicant deluge–then they stand less of a chance of hiring someone good. If they interview more candidates, then they will be taking on more administrative work for themselves. And are they not already spending more time than ever before reading applications?

The sensible thing for the employer to do is to somehow (i) reduce the total number of applications and (ii) increase the ‘seriousness’ of the remaining applications. You can see where this is going. The sensible thing to do, in other words, is to make the applications longer and more onerous. 

And so, the moment a clever software engineer first created an online job portal, the long application form became inevitable. ‘Describe a time when you applied teamwork to solve a problem.’ [250 words]. ‘How do you embody our values of customer-focus, excellence, and delight?’ [500 words]. Seeing these questions, some will close the application tab. Others will sigh and begin writing, showing their iron commitment to the job that lies ahead. Applicant numbers down. Seriousness up. 

Until 2023, a state of calm had settled over the job application system. In entry-level jobs–like graduate schemes–which were most susceptible to internet-enabled application spamming, employers thus constrained applicant numbers through a series of Byzantine application processes. Online tests. Long forms. Assessment days. Several rounds of interviews. These took up a huge amount of the applicants’ time, inadvertently creating a paranoid style in the way that students came to approach university (Have I done enough internships? Have I demonstrated leadership skills? Can I evidence my empathy?), but they served their purpose in limiting unserious applications.

Recently, the introduction of generative AI has noisily interrupted this equipoise. Just as with the arrival of the internet, jobseekers have been handed a tool that increases the number of applications they can make. The long application form is not so daunting when ChatGPT can write it for you. The online exam or the pre-recorded interview is not so difficult when you can ask Claude to come up with answers as you go. The AI’s efforts are obvious and turgid, but they may not always be, and these chatbots are at least completing applications: application numbers are shooting up once again.  

“Securing new jobs is becoming harder: Job applications surged by 31%, significantly outpacing the 7% growth in job openings compared to the same period in 2023.” (Workday’s Global Workforce Report Q1-2, 2024)

This is a problem for applicants, obviously, and for the employers who have to wade through a growing swamp of applications to find good candidates. Even if employers change nothing about their recruitment process, it will take more work to recruit for a given role. If they try to outrun the chatbots by making applicants write more, and with more specificity, it will take more work to recruit for a given role. If they try to spot the chatbots by introducing AI detection tools and invigilated remote assessments, it will cost more to recruit for a given role. And if they banish the chatbots by introducing more in-person assessments for the soon-to-be hundreds of applicants, it will still cost more to recruit for a given role.

Complexity runs through all of these options, and it is complexity that creates more work for employers and applicants, the cruel type of work that brings no real world benefit. But what’s the alternative? Perhaps companies could introduce application fees. That would reduce applicant numbers and increase applicant seriousness, but it is also an idea that only a psychopathic economist could support. 

Or companies could bring in randomised sifts, which would reduce the number of applicants to a manageable one. You could then invite the sifted applicants in for an in-person interview, exam, or coding test. But could we really sift like this? I don't think so: we believe too strongly in the ideal of meritocracy, in the notion that there is a perfect candidate out there, a candidate that our recruitment systems are set up to find. The thought of idly sifting away ‘the one’ is too painful. 

With no other good options in front of the company, I suppose applicant-cutting will instead be done by AI screening tools, which will spend happy days sorting through piles of AI-generated applications, creating a never-ending loop of generic applications receiving generic feedback, beautiful in the way it reminds us that the symbol for infinity is a snake eating its own tail.

This thirty-year recruitment revolution is, I suppose, another example of the shimmering deceptions of technology. The people creating the first online job boards must have sincerely believed that they were simplifying the world of job applications, making it easier for applicants to find and apply for work, and creating a much more energetic and meritocratic labour market, where anyone could easily switch careers. And yet the opposite has transpired. Job applications are more onerous–for everyone–than ever before. Of course it’s too soon to say what the new chatbot recruitment world will look like, but will it be any better? Really?

All of this points to a wider idea, one that appeared time and again in my interviews. When we introduce technologies that speed up the production and distribution of information, there is often a corresponding rise in the amount of information that we come to view as ‘necessary’ for a given task. Put another way: the quicker we can write documents, the longer we expect them to be. That is as true for job applications as it is for, say, planning applications, or the pitch decks of investment bankers, or the notes written by healthcare workers. It ought to scare us: you can’t build a great economy on long documents.