Similar to how search engine optimization works for websites, a resume screening algorithm is looking for keywords, phrases, and text strings within the document itself. These keywords are usually the knowledge, skills, and abilities a company is looking for in the ideal candidate, and often include the duties of the job. The algorithm will sift through a resume to match against the keywords. Often, from here, a resume is scored according to keyword match and fit. Often, hiring managers can emphasize the order of magnitude for knowledge, skills, and abilities. So, one skill might have a higher weight in whether a candidate is a fit for the job.

The downside to using AI to sift through job candidates is that machines, rather than people, perform the tasks. It is going to find the candidate according to the pre-determined criteria. This limitation puts a lot of pressure on the hiring manager to be thorough about the requirements for the hire. If jobs become too standardized by these algorithms, companies can miss out on a home run hire. Maybe a candidate applies for a job in project management, and on the resume, we learn that this individual was a contract employee that lead a massive system integration at a company that leads to an additional $2M in revenue at the risk of $5M in losses. Learning this alone speaks volumes about this candidate. And there’s no good way to create quality tags for that kind of experience.

A candidate will want to ensure that the resume contains the keywords that are showing up in the job description, specs and competencies. Does the company want highly competent in specific software? Make sure to have a section that speaks to that particular software. In my industry, there tend to be specific skill sets within a skill cluster (multiple regression, econometric modeling, cluster analysis, etc). They could roll these up and call them “statistical modeling,” but a hiring manager often opts not to because he or she knows these are the core techniques we use. So, a candidate will want to match this tit-for-tat. “Experience in statistical modeling, specifically in (insert the specific ones you asked for).”

There are three keyword clues usually found in a job posting, and they are on a continuum of observability. Surface-level attributes are often found within the job requirements and include the length of experience, educational background, licenses, and other related skills. These are things that are easily observed and easily verified. The next is the back – does the candidate’s employment history match our desires? Has he demonstrated an ability to write surveys, conduct analysis, etc.? Third are the less tangible skills, often called “soft skills.” Things like “people management,” “project coordination,” or other phrases that suggest something about an ability to execute a task. A job description might say they want a candidate to “manage a project from initiation through completion.” So you want to hit on how you have produced work from end to end.

There is often some level of expressed magnitude for a skill or task. It’s a clue for the level of desire for this skill or task. If a job posting says it wants “mastery” in a software program, that’s a clue that the software is weighted heavily by the algorithm. Other key emphasis words might include: “significant,” “strong,” or “compelling,” the adverbs “successfully,” “effectively,” or “efficiently” (note the “ly” at the end) or emphasized nouns, like “mastery,” “proficiency,” or “competency.”

In the arms race of rising up the ranks in web search, early SEO professionals would hide keywords throughout the website so that someone searching would be more likely to find this site. Google and other search engines figured this out quickly, so SEO folks had to adapt to work up the ranks in search properly. Some candidates might try to do something similar with resumes. It’s called the white font trick. Candidates will simply list as many related keywords as they can fit in white font. The naked eye won’t see them but the algorithm may still pick them up. This is not a smart approach for job candidates because many algorithms have picked up on this trick and will flag this. Plus, if certain keywords come up too frequently, it’ll likely get flagged as well.

Now, cover letters are still highly important. While a smart resume speaks directly to the job posting, the cover letter is a way to talk directly to the hiring manager. A way to speak to the job in more intimate detail in ways a resume cannot. A candidate can reference to those intangibles through experiences and stories, and to why he or she is a fit for the company. It is an opportunity to get noticed by the hiring manager and is often the key talking points when folks get in a room and decide who to bring in for an interview.

The biggest mistake job seekers make when submitting their resume digitally is not writing a resume to the job posting. Sending out the same generic resume is a sure way to get called by a bunch of jobs that are just looking for warm bodies. In terms of speaking to the algorithm, if the resume cannot be read on a basic word processor, it won’t do well in front of an algorithm either. Any images, unique fonts, tables, charts, are often not great for submission. I have seen some impressive resumes with graphics that detail a candidate’s skills and capabilities, but that’s better served in an interview. You only get so much space, which means minimal real estate to hit the keywords and tags the algorithm is looking for.