Discover more from Culture of Engineering
Five Reasons companies do not get quality candidates.
ATS systems rule the candidate world, but hiring is human. The more a company allows technology to replace humanity the fewer humans a company will attract.
This is my third tech tsunami. In 2001 I watched as people took their Herman Miller chairs as severance and walked them down the block of the March First building. I 2008 I watched companies collapse overnight as people lost everything they recovered from their loss in 2001. We thought the industry had matured enough; we were done with the mass layoffs and dominos falling. Now it’s 2023, and the wave begins again. Some call it a bust; others blame Covid, still, others call it “right-sizing.” However, the tech sector is having another one of its “moments.” Founders Fund needed to put a run on SVB, now? One difference at this moment is the saturation of ATS (Application Tracking Systems) with AI capabilities, think Workday, Greenhouse, or Workable, that companies bought, hoping to streamline the search process so you could find the orchids through the weeds.
Sadly, the weeds are proliferating. Why? Because hiring is a human job. Hiring is not something you can automate. Screening is not something you can 100% automate. An ATS can assist with routing automation and tracking, but you can’t change the straightforward fact that hiring a human being requires another human being to be involved. The more we try to AI our way out of humanity, the more we will find we live amongst the weeds, not the orchids.
I have spoken to candidates and recruiters, people with 20+ years of experience in their fields, and they all say the same thing.
The candidate: “It doesn’t seem like anyone knows what we do, and even if we are qualified, we’re getting auto-negged because we missed keywords or the JD was basically - everything, just do everything, and of course, I can’t do everything.”
Culture of Engineering is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
The recruiter: “these systems are killing us because we’re judged on metrics that make no sense, and we simply can’t take time to assess someone, or the metrics make us look slow and inadequate. It’s worse if you’re Junior or Mid because sometimes you’re left setting up the tools and don’t have clear direction from management.”
I am not suggesting dumping Workday or Greenhouse. Organizations need to be operating these systems well and be focusing too much on reducing the number of candidates to be reviewed while alienating candidates from applying, leaving room for more weeds. This is not to say that the wrong people will not still apply for roles for which they have no business to apply. However, assuming it’s the rule, not the exception, the organization is leaving great people in the dust.
So as a current “candidate” who has also been on the hiring side, going through this process again, here are five things that everyone needs to improve so candidates can find homes and organizations can find talent.
1. If your organization’s system can’t parse a resume and refuses to read it, YOUR HIRING SYSTEM IS BROKEN!
TL:DR - you don’t want to read a resume that isn’t formatted as your system likes; you don’t want to employ people. You want robots.
On Friday at 6 pm, a former colleague refers you to their company. The following Monday, which incidentally is a holiday, the candidate receives an auto-rejection email explaining that the candidate's resume is impressive but needs to be a better fit for the open position. Because of poor automated resume parsing, a robot denied a candidate with 20 years of experience. This would not have happened if a human being had been involved.
There is no perfect resume. There is no one size fits for anything in tech; when your organization is designing the workflow, there will be trade-offs; in this case, an administrator chose to let robots determine if a resume was formatted correctly—bad choice. A candidate who has likely written and rewritten their resume hundreds of times is lost to the virtual paper shredder for the crime of using columns.
Why does this happen? Mainly due to time constraints. Based on some stats I have read, recruiters can get 1000 resumes daily, and only 5% are from qualified applicants. The really good recruiters can scan and hone in, but they are still scanning.
How do you fix it? Share small items with the candidates about how your system works. If it’s an “easy apply,” let candidates know how you want documents delivered. “Easy apply” should not require a candidate to spend 45 minutes correcting poor parsing. Scanning does not consider the nature of the internet, where companies can rise or fall with blinding speed. Perhaps the companies cited on the applicant’s resume are one of the fallen. Perhaps the applicant was considered a “genius” at one or more of these defunct companies. Help the candidates get your organization what’s needed to make your decisions; we’ll do it, we are taking the time to apply for the job, and we don’t want to fall into the dead letter office over technicalities.
Help the recruiters: Allow them to have reasonable timeframes. Sure, you want someone ASAP, but you want someone good, and it’s not right to give a recruiter 10 seconds to figure out if someone is good. Offer your candidates the chance to provide feedback on your application process so you can improve their experience and as a result, their opinion of your organization. ATSs have two masters: the candidate and the organization; focusing only on your organizational needs will alienate candidates. For example, how many candidates might express annoyance that a company requires a base salary entry when the proposed budget for the role is not even posted? The information will help your organization improve the candidate experience.
Give as much as you take. Candidates aren’t doing this for fun; they are investing their time trying to do nothing more than impress you; give them the best chance to do that.
2. Tell us what you REALLY want.
Manage expectations: The more the job description particularizes the responsibilities required by the position, the more likely only qualified candidates will apply. An overly broad description frustrates many applicants who are perfect fits for the actual responsibilities of the position. Others may take a shot. For example, someone with 15 years of cybersecurity experience may apply for a B2B Fashion Marketplace position. Why? Because it wasn’t made clear that beyond the basics of the role, someone with fashion experience was needed.
Everyone wants a high-achieving team player. At the highest level, I have 20+ years of experience in product, operations, and project management. However, I could sooner fly to the moon than successfully perform all of the responsibilities required to fulfill the following:
“Lead everything, responsible for budgets, team management, client management, high-quality solutions while doing user research, product marketing, performance marketing, social, wireframes, then figma/sketch designs from those wires, also that you will likely need to code some of a proof of concept, while gathering requirements, negotiating SOWs, managing resources, and then be ready to lead the company with your Go To Market strategies, and have 10+ years experience with NFTs”, that weren’t created until 2014. Then your organization is not looking for a role. It’s looking for a Unicorn. A Unicorn will probably want a much higher salary than what is offered.
Suppose the organization is not looking for a Unicorn but honestly thinks it is the responsibility of a project manager also to be a front-end developer with some VR skills. In that case, the organization is why good candidates are not applying because those candidates don’t exist. Take a step back. Break the responsibilities into the three roles that they are. If the organization cannot afford three roles, it is unprepared to hire.
3. Keywords are not a career
TL:DR - don’t have the first part of your application funnel be so inhuman you alienate great candidates.
Example: A resume submitted through an ATS Tuesday morning. The auto-rejection arrived three hours later because keywords were missing. If only the candidate had known about REZI. Three days later, the CEO of that company directly reaches out, asking for an interview. Why? The candidate, knowing that only the keyword Pacman machine would review their resume, cold-emailed the CEO.
A Federal employee responsible for hiring recently told me that their system passes every veteran applicant. Any veteran with “writing” or “editing” is automatically moved to the top of the sourcing list.
Keywords must mean something within the context of the needs of a company. The keywords writing and editing define the basic skills needed for an editorial position but only filter out applicants whose writing and editing skills are scaled for a particular company. The veteran with the background to write and edit military reports is not a good fit for an editorial position with a fashion magazine. To avoid the resume avalanche, your organization prioritized “writing” without considering the difference between writing the Presidental Daily Briefing versus a fashion review.
Surprised? A good deal of the ATS systems will allow you to filter and sort by specific dimensions to check the resume hits the basics before a human reviews it. AI does the first pass to check the resume meets the basic requirements; this is how companies manage the 1000 resume avalanche.
Who defined the basics? Did the hiring manager give a list of keywords? Did someone with experience in the role help design those AI checks, or are you just pulling words from the JD to see if there is a match? Did you take something out “of the box” because the provider probably knows the best practices? Did you adjust everything for each role, or are you just saying, “use the same thing for the project manager and the product manager role?” We are seeing businesses being explicitly created to answer the demand of how to game an ATS keyword system so humans can interview candidates; THERE IS NO MORE SIGNIFICANT SIGN THAT THE SYSTEM IS BROKEN.
Clarifying the scope of keywords will likely result in fewer applicants, but the quality of your applicants will increase. However, be careful not to turn the application into an interview when looking for more nuance in your candidate screening. While candidates don’t want to be rejected because they missed a keyword, they also don’t want to write five essays about projects they have worked on to get a phone screen. Your recruiters also do not wish to spend hours reading five essays about projects in the candidate's resume. Asking, “how do you manage conflict” on a job application indicates your organization is trying to automate the interview, making a human process inhuman. That’s not attractive even to the most desperate of job seekers.
4. Candidates are not your demographic data pool
TL:DR - Don’t you dare ask a perfect stranger who they like to sleep with; it’s none of your damn business.
There are precise questions that candidates 100% understand will be asked of an applicant. They are:
1. Citizenship status/work eligibility status
2. Disability status for accommodations
3. Veteran Status
Beyond those three items, basic contact info, and maybe a LinkedIn profile, all you’re doing is being intrusive for organizational data mining. I will be highly blunt about this.
There is no reason you need to know anyone’s sexuality; it’s irrelevant to the job. Asking so the organization can produce a report showing inclusivity and diversity in the hiring process is offensive. If you want to prove that your company cares about inclusivity and diversity -- hire inclusively and diversely. There is no circumstance where a candidate should be asked if they are heterosexual, homosexual, pansexual, asexual, or anything else on a job application form. It is illegal to ask during an in-person interview; trying to skirt the legality to data-mine by asking in the application is transparent and insulting.
Respect my pronouns; thank you for asking vs. assuming. After that, you do not need to ask me if I am transgender or cis; you do not need to ask me how I identify, then later on ask again for EEOC purposes what my gender is. While I’m ranting, if you’re going to have the audacity to get this personal before you meet someone and potentially reject via automation, be consistent . Research what you are asking. Sex and gender are not the same, so asking my “legal gender” is not acceptable. By trying to be inclusive without taking the time to understand what the very question telegraphs you did not take the time to understand this issue. It doesn’t imply inclusiveness at all.
5. Be Human
Automation can’t be relied on to make decisions about an applicant; automation doesn’t know what is good, bad, better, or best. Granted, automation makes redundant tasks easier. It eliminates the distractions of work that does not benefit the end goal. Machine Learning? Sure, one day, but not today. No Machine Learning can determine the ideal candidate for all the roles and the company ethos. Yet every day, new tools pop up to help you get that “edge,” “trick the system,” and “master the application process.” It’s the adult Stanley Kaplan industry for the new SATs (note SAT and ATS are the same letters).
So what to do about it? Banishing Workday or Greenhouse won’t do it. These systems 100% make the internal management of many applications manageable. As an organization, you invested in these tools to simplify and streamline your hiring process, and improve the candidate experience to represent your values and ethos. Your hope is to attract, excite, invite, and keep. This application is the candidates’ virtual introduction to the organization. It is not your website or your product. The application process is the opening salvo, so make sure it’s the best experience both you and the candidate can have. Using the two worst words in all of the technology, as a hiring organization, you are going to focus on quality, not speed. At every point of designing your application, ask, “If I had to apply for this role, would I and why?” The answer can’t be “because my company is awesome.” Candidates only know what they have heard or read. They are going to pitch themselves, just like you’re going to pitch them the company and the role they can have in it.
Slow down, and let the humans who you already hired do what they do best, find and source talent. Craft Job Descriptions with your recruiters that both entice and inform. Make sure the automation you set up is not blocking but facilitating. You are going to rule your ATS system and not let your ATS system rule you.
The answer to the problem is human, not tech - it takes human beings time to review and assess other human beings. A recruiter, a sourcer, a hiring manager all need time and consideration to review potential hires. A dashboard reflecting 1,000 applicants and only ten interviews might lead you to believe you’re so selective you must be getting the cream of the crop. I guarantee you more than ten qualified, and one of the 990 auto-rejected was probably the candidate you should have hired.
Technology doesn’t replace humanity, ever; it can amplify it (sometimes destroy it), help reduce distractions, and provide endless hours of entertainment, but it simply can never be human. In its current incarnation, it is not ready to make very serious human decisions for you, especially about who you might spend the next 15 years working with or for. Technology can solve your storage issues, routing issues, and “omg, did we email this person back yet?” issues, but it can not solve the question of “is this person the right fit for my very human business.” Only humans can do that and make your ATSs do some of that for you.
Candidates are not the enemy; we are the answer - help us answer your call.
_ Alison Grippo is a fried of Dustin’s and thinks we need a way to make technology humane _
Culture of Engineering is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.