I have seen hiring from inside, the middle, and the outside. I have hired technical talent, from software development interns, to QA engineers, to technical writers, to some of the top data science talent, and all stops in between. I have been hiring for almost 15 years. I have also helped build a recruiting platform and seen the inner workings of the process. Recently, I have put myself on the other end of the table.
This process is broken, fundamentally broken. Most job seekers do not have the luxury of sharing how broken the process is, so this post is for them. I have spent the last 6 weeks as an active job seeker and the last year talking to those in the job search.
I am one of the most recognizable names in data science so if anyone were going to have a smooth ride, it would be me. Many people told me that I should use someone else for the experiment because, “People like [me] aren’t having trouble finding jobs.” The truth is anyone who uses the traditional hiring process is having trouble finding a job.
The process is widely known to be dysfunctional and those of us with options avoid it completely. Referrals make up 40% of all hires (JobVite) and almost half of all companies say their best hires come from referrals (LinkedIn). That is telling. People like me are not having trouble finding jobs and companies who depend largely on referrals are not having trouble finding candidates because we bypass the process altogether.
I Have not Had A Resume In 4 Years
I tell people I do not have a resume and they think I am lying. I have seen this hiring process from a completely different, utopian angle. While I run a consulting practice, I am still hired by my clients. Most find me through a referral from someone I have worked with or by seeing something I have written. My reputation is established. While I am often asked for a resume, potential clients immediately accept that I do not have one.
The process is different. No resume equals little outreach from recruiters or hiring managers. Even with as active as my social presence is, recruiters rarely come calling.
I spent a day building a resume and the process is laughable. Reading the best advice online, my resume should be a comprehensive walk through of my career. Anyone with a career spanning 5 years can write a book about their professional experience. At 22 years, mine would be a 4-part series. My resume will be out on Amazon in 6 months.
My 1st attempt was 8 pages long. I pared down as best as I could in the traditional summary, experience, skills, and education sections. I picked 5 sample projects and reduced to the core technologies I have used over the last 5 years. I got it down to a manageable 2 pages.
How anyone determines my qualification for a role with those 2 pages is a mystery to me. I believe the dysfunction starts here. I have distilled a lengthy career, with every step playing a role in my qualifications today, down to 2 pages. After reading my resume, I would not hire me.
If you follow the research on personnel selection, their findings support my opinion. According to multiple studies, only 33% of employee performance can be explained by biographical data such as a resume or work experience. Cognitive capabilities are the best predictor at 62%. That means intelligence is a better predictor of job performance than experience. It turns out to be a better predictor of job performance than any other data point.
Industrial and Organizational Psychology studies personnel selection or hiring (among other things) with the goal of understanding which selection methods choose high performing employees. Modern research uses a technique called psychographic meta-analysis to determine how much of performance is explained by a specific indicator or data point. That is their validity measure (greatly simplified for brevity) and it is the methodology behind the numbers in the last paragraph.
A resume is an incomplete biography of a candidate. It has been shown to be a poor predictor of candidate performance. A job opportunity is an incomplete description of a role. It is known to be a poor predictor of how the employee will produce value for the business. What part of this process does anyone expect to work if the foundation is built on incomplete information? Intuitively, candidates know this so the process of creating a resume feels more like a game that needs to be played rather than a serious exercise.
The Recruiters from A Candidate Perspective
The resume and job opportunity are not useless documents but when they are expected to be comprehensive, that assumption derails everything that comes after it. Most candidate searches are done using keyword matching. The recruiter takes keywords from the job description and plugs them into some sort of search engine. The resumes with those keywords come up highest in their search results.
I have built a resume search engine, so I know a lot about this part of recruiting. Newer search engines add intelligence to this process. Key terms alone do not mean much. However, the semantics of each term can reflect a sort of term IQ or how intelligently the candidate talks about the term. That is more important to matching a candidate to a job than search terms alone.
Unfortunately, most resume search engines use an older method and only match based on the search terms. That leads to a whole bunch of irrelevant candidates coming up in the results. Those candidates get emails and phone calls for jobs they have no interest in applying for.
For in demand technical fields, that leads to hundreds of emails and dozens of phone calls a week. That is not an exaggeration. At first this feels like a good thing. I am wanted. When candidates drill down into what they are being offered, those good feelings disappear.
While positions may have the right job title or matching skills, they are not substantively what a candidate is looking for or even qualified to do in many cases. There’s nuance to technical fields where a single term or requirement in a job description can change the role fundamentally. If that has not been picked up, the result is hundreds of candidate outreaches that miss the mark.
In some cases, I got three emails from the same recruiter, for the same job, on the same email address, in the same day. I got emails where the form they wanted me to fill out was longer than the job description. One job description was “NLP And Machine Learning. Good Communication Skill.” You expect me to have good communication skills when the job description is two lines and one of them has a grammatical error?
I got everything from a computer technician to .NET developer to Salesforce engineer offered. The data scientist roles were all over the map. There was no clear rhyme or reason why I would get one role over another. If the job title and a couple of key words matched it got sent my way.
Many emails seemed like outright frauds. Some requested the last 4 digits of a social security number. Some had names in the email signature that did not match the name on the email address. Some rates were so low that they could not have been legitimate job offers. I Googled unique text in a few of the job descriptions. They came up with cached pages from jobs that were no longer posted. Some had text like, “We’re working with a well-funded, Fortune 500 company.”
Candidates start out reading every email and end up ignoring them all. If you are a recruiter, your emails are in purgatory. The likelihood that anyone reads even the good outreach is minimal because we cannot tell good from bad after a few days.
What about all the phone calls? I started taking calls during my experiment. I quickly learned to qualify calls with a single question, “Tell me about the projects they’re working on?” A serious recruiter has an answer to this question because they have spoken to the hiring manager. While they cannot always drill down into specifics, they can explain some project details. I went full days without getting a single call from someone who knew the answer to my question.
Some callers started reading from the job description. Some said the job had just been posted so they did not have that detail yet. Many said their manager had those details. Others asked why that was important to me? A few were angry that I had asked them a question.
I took calls for a full week. The process was offensive. Tata Consulting Services, Cognizent, and many other consulting houses use these robo-callers. Clients quoted to me included Fidelity, Macy’s, GE, and State Governments. Their clients don’t know that some of the data scientists they’re paying $300-$500/hr. for are recruited by a caller who’s only qualifying question is, “We have a role for data scientist in [city]. Are you available?”
Many candidates start screening recruiter calls after a day or two. If you are a recruiter getting voicemail from almost every candidate, it is not because we are that busy. Other recruiters have soured us before your call.
I Know Great Recruiters
I have met your kind before. I say you because if you have made it this far you care.
The recruiters (more order takers than professionals) who turn candidates into recruiter haters do not care enough to read resumes or job descriptions. They do not want candidates to pick up the phone. They just want a resume, so they can submit it and move on. For them, it is a pure volume game without consideration to accuracy. They have soured the pool before you get to it.
Why? You spend the time getting to know the role. You call hiring managers and sometimes people in the role currently. You spend time coming up to speed on new terms and technologies. When you search, each match you are shown is viewed with healthy skepticism. You review the resume and look them up on LinkedIn.
Unfortunately, your email is 23rd in line behind a tsunami of raw sewage all titled with urgent, immediate hire, or critical. Your outreach is guilty by association.
Companies are paying sometimes $25K - $50K for order takers to harass candidates with no interest or capability for their roles. They are not seeing the quality, or lack thereof, provided for that fee. The average job requires 4 - 6 interviews to fill and those candidates are selected from 250 resume submissions (Glassdoor). That is as broken as it gets, and it is a volume of garbage that quality recruiters are fighting to wade through.
I do not blame legitimate recruiters or the companies who utilize the services of low-quality recruiting companies. My guess is both are just as frustrated with the way some do business as job seekers are. However, the reality is that several bad actors are degrading the reputation of great recruiters and any company associated with the bad actors.
What Comes Next?
For most candidates going through the process, the next step is silence. They fill out the form, send in a resume, have a follow up call with the recruiter, and then…nothing. I have been on the corporate side of hiring. The HR team screens most of the obvious bad matches and few of the resumes that make it past that screen are ever called for an interview.
That means for candidates going through the process, it’s hurry up and wait. That is where many candidate journeys end. They have spent a few hours on the process of applying for a position to hear nothing back. An active job hunter can be doing 4-5 applications a week which adds up to a lot of wasted time.
Most qualified candidates eventually find a job, but the process is soul sucking. I have not covered the issues with time to hire, poor interview processes, and all the other little things that make getting hired more difficult than it needs to be.
I have not gone through that from the candidate side in almost 5 years and I did not get any job outreach that I would consider. There are roles and contracts I would have applied for. I was ready to find my next client through the traditional hiring process. Over 6 weeks I did not see a single match.
I think that is the biggest take away. For niche talent, the traditional process does not work at all.
From the stories of others, I can tell you the interview process is broken as well. I will not dive too deeply into the research, but most businesses are not conducting interviews in an ineffective way. Unstructured, free fire questioning interviews have a validity of .20 for predicting employee performance. Traditional, structured interviews have a .37.
Those that are conducting effective interviews, have spent a lot of time and money to get the process right. It is not easy to build an interview process that results in high performing employees. Job interviews should be behavioral structured, with a focus on job performance and job knowledge questions. You will see these at the higher levels of some Fortune 100 companies.
Assault with A Deadly Process
The entire process is built to discourage and eliminate. That is the governing dynamic of the hiring process. The root cause is those 250 applications per job opening. When only 2% of candidates make it to the interview, the fewer candidates the better. Given that reality, you really want candidates to self-screen. The process is difficult to weed out the candidates who are not that motivated or interested.
The deeper problem is the number of under qualified or qualified in a tangential field candidates who apply for jobs. The process is meant to discourage them from overloading screeners. From the 250 applicants per job statistic, that is not working.
In the end, the root of a job seeker’s woe is job seekers themselves and the robo-calling order takers are not helping. If only qualified, well matched candidates applied for jobs, this would not be necessary. The process would be built to include those candidates rather than weed out the masses.
If there was more certainty around qualifications for both the job and the candidate, we would not be in this mess. The dysfunctional process that is hiring was built to account for that uncertainty. Like most processes built to manage uncertainty, it does not work very well. Hiring is like a gambler who swears they have a system to beat the house.
It is neither measured, nor managed. That is our biggest problem. If we had more raw data to show how much this problem costs each business, there would be more traction for real solutions over Band-Aids. Based on SHRM and BLS numbers, the cost of hiring is close to $272 billion a year in the US alone.
What percentage of those hires stays for over a year? What percentage are high performers? What is the ROI on hiring and recruiting processes? We cannot improve any of that since most businesses do not measure it. What is not measured cannot be improved. Most businesses are not aware there is such a thing as a better hiring process.
With access to talent becoming a critical success factor for businesses entering the machine learning, blockchain, IoT, infosec, and any other emerging technical space, those who start solving this issue will have a competitive advantage. There are technology and process solutions available but until businesses take the problem seriously, the mess will probably get worse before it gets better.