Automating Your Job Search

Over the course of the past few months I applied to almost 1,000 engineering leadership positions in the search for my next big adventure. If you're interested in how that job search went and the lessons I learned from it, there's a breakdown of the entire thing in data available here.

This article is not about that, however. It's about how I streamlined my application process so I could apply as effectively as possible to the numerous positions available that matched my requirements and expertise.

This article breaks these strategies down in ascending order of difficulty for you, the applicant. So, if you reach a point in this article where you think "that's too complicated for me" that's fine. You can take what you've learned thus far, implement it, and you'll be more efficient in your own job search for doing so.

What's The Point?

When I began my job search it was not my intent to push out a huge number of resumes. I began my 2024 search as I did my 2021 search: careful consideration, purposeful networking, and a heavy reliance on my own professional network.

It got me nowhere.

In several cases I was never contacted at all by companies where I had former coworkers and dear friends. In others, carefully scouted opportunities with warm hand-offs and enthusiastic endorsements along the periphery of my network got back an automated rejection letter within minutes of applying. There's value to the human element but if only if it gets your foot in the door.

Stage One: Automation By Not Doing Stuff

So, step one of my optimization process was cutting out the unnecessary. I pride myself on my writing but cover letters weren't landing me interviews. In fact, when I asked HR and Talent folks about my cover letters most of them confessed to not having read them.

I tried using AI to generate cover letters but found myself dissatisfied with the tone and the AI's tendency to invent job experiences. Cleaning up after it ended up consuming more time than just writing the letters from scratch and neither was doing me any good. So I cut cover letters from my process for the most part.

Along with cover letters went any attempt at 2nd degree networking. If I happened to be applying to a position where I knew someone I took the time to ask for a referral, but trying to see if a 1st Degree contact would make an introduction to a 2nd ate too much time for too little benefit.

Stage Two: Basic Tracking

Of course, the single biggest time waster imaginable is applying to a position you've already applied to. Some kind of application log is an absolute must for any job search. Besides giving you a way to see where you've applied (so you can avoid applying there again), your log can help you identify opportunities for follow-up.

Here's the basic log template that I use. It has some fancy conditional formatting to let you know where you stand on your various applications at a glance. We'll be coming back to this later in this article but, even if you tap out well before then, just doing some basic tracking will save you a fair bit of time, especially if your job search extends past a few weeks or a few dozen applications.

Stage Three: Automating The Easy Parts

If you've been on the job market long enough to need a spreadsheet to track your applications you've probably been frustrated about (at least) two things in HR/Talent software.

Some Applications Are A Lot Of Work

I'm looking at you, Workday. Five minutes on LinkedIn will get you 20 posts complaining about how terrible the application process is through Workday. You have to create an account for each new company; the resume parser just straight-up doesn't; and the "skills" section seems like it's trying to identify mind-readers.

You are not the only person with this problem; automation can help. SpeedyApply is a bare-bones auto-fill solution for most major job application portals including Workday, ICIMS, and Greenhouse. The idea here is that you configure it once, turning those platforms into (very nearly) one-click applications. This saves a TON of time, especially with Workday

All Applications Are Pretty Much The Same

SpeedyApply is a Chrome-only extension and that's probably fine for most people. If you're a FireFox user you might consider Simplify to fill the same niche. I found that Simplify's browser extension worked in some places that SpeedyApply didn't but I also appreciated its ability to serve as a copy-paste databank for filling out applications on sites where automation doesn't work. The extension was well worth installing just so I didn't have to remember my LinkedIn URL.

These extensions are valuable because the vast majority of the information you're entering into a job application isn't relevant to your application. As a consequence, these sorts of questions don't change much between companies. Once we're past the content that's actually necessary to apply -- your resume and (optional) cover letter -- the remaining questions break down into three broad categories.

1. Demographics -- plenty of these questions are required by law but that means the questions take on an extremely predictable form. Automation works well here but you should review the results for accuracy.
2. Tracking -- these are the "how did you hear about us" questions. Again, they are fairly standard across applications but automation solutions are more prone to incorrect responses. On the other hand, that doesn't matter. Correct automated stumbles if you notice them and feel like it.
3. Everything Else -- "why do you want to work at [company]" type questions are fairly common and strongly resistant to automation. Platforms like Simplify will promise to fill those in with AI but the answers will be vague and valueless. While cover letters may not get much attention, these questions are purposefully included in the application by HR or the hiring manager and they're where you should be spending your time.

Stage Four: Defeating Automation

Within a few moments of marking yourself "Open To Work" you were probably bombarded with dozens of self-proclaimed "resume experts" offering to retool your resume to make it play nice with ATS. ATS stands for "Application Tracking System" and it allows Talent folks to search a database of resumes for candidates with specific skills.

Candidates that the system identifies as having the required skills might be considered for an interview.

From a hiring manager's point of view, this isn't the same as rejecting applicants that "fail" the ATS step. The system isn't seeking out "bad" resumes and rejecting them and it certainly doesn't rise to the same complexity level as AI. But, from an applicant's perspective, it sure seems like an AI is rejecting resumes for lacking the right keywords.

If you don't want that to happen to your resume there are a couple of strategies you can try.

Bespoke Resumes

We'll start with the old-school, non-automated approach here just to set the stage. You open up the job you're applying for and you re-write your resume in anticipation of what the hiring manager probably wants. This is going to take a lot of time and it is fundamentally ignoring the ATS parser in favor of the hiring manager.

AI Generated Bespoke Resumes

Plenty of services offer to use AI to reshape your resume on an application-by-application basis. Simplify offers this service in their paid tier but there are plenty of others. By far my favorite is the resume grading and customization service that Jobright.ai provides for free (we'll see how long they can keep that up; it's got to be expensive to run).

As with any AI driven content, you need to review everything it generates. AI is prone to hallucination and may just make up experiences that you do not have. Automation is great, but you do not want to lie on your resume; it could go very badly for you. Jobright's model highlights the changes it makes, which streamlines review. On the other hand, the Jobright AI is tightly coupled to the Jobright job boards, so unless you're going to conduct a one-board search, you'll need other solutions.

White-Fonting

White-Fonting is when you grab a large block of text -- frequently the entire job ad -- and drop it onto your resume in all-white in a font-size suitable for Enron or Beelzebub. The ATS parser will happily see it and read it but the hiring manager might not notice it.

I shouldn't have to say this, but this is wildly unethical.

You shouldn't be claiming skills or abilities on your resume that you don't have and you shouldn't be trying to fool HR or a hiring manager into believing that your resume contains information other than what it actually does contain.

Keyword Banks

That said, there is value in openly acknowledging that your resume is going to first be read by a machine. One of the first optimizations I made with my resume was the inclusion of a keyword bank at the bottom under the heading "Selected skills for machine parsing." The idea here was pretty straight-forward: I wanted to make sure that my resume included specific keywords without having to weave those keywords into my work history.

Being honest about what I was doing eliminated the risk that I would be seen as misrepresenting myself or "cheating" somehow but it also made my resume look "a mile wide and an inch deep." I was hesitant to put more niche skills into my keyword bank for fear of looking like a jack of all trades but a master of none.

Automated Bespoke Keyword Banks

The solution I ultimately arrived at was an automated, bespoke keyword bank. Where I couldn't leverage a quick-and-easy AI-generated resume, this option offered a optimal blend of speed, ease of use, and tailoring.

The system consists of two parts: a Google Spreadsheet (the keyword bank and logic) and a Google Doc (the resume). To use it, I would input key skill phrases extracted from the job into the spreadsheet, hit the "Tools->Linked Objects->Update all" button in the resume document, and save the result as a PDF.

While using this system is easy, setting it up is a bit of a pain. While you're welcome to make your own copies of the template Document and Spreadsheet, you'll have to set up the linkage between the two yourself.

• Delete the existing "Selected Skills From..." sections from the copied resume
• Go to the "Linked Cells" tab of your copied spreadsheet
• Select the two rows in column "B" which correspond to the header and skills for the skill block you want to embed
• Copy them (Ctrl+C)
• Paste them into your copied resume document (Ctrl+V)
• Be sure to select "Link to Spreadsheet" when prompted
• You may want to then resize the tables and reduce their border widths to 0px.

The setup of this system is non-trivial but what it got me was the ability to take any job posting on any site, extract some keywords, and generate a resume which would ethically and transparently provide a good match for any ATS query. What's more, plenty of job sites list keywords that describe the position in a format that pastes effortlessly into the spreadsheet. LinkedIn in particular is great for this.

Stage Five: Closing The Loop

The efficiency gains unlocked by using automation to fill out the boiler-plate questions and replacing bespoke resumes with machine-generated ones will massively increase the number of positions you can apply for. This, in turn, will create two new bottlenecks.

The first is your ability to read and digest job descriptions. While many of the above tools could be used to just indiscriminately blast resumes out into the void, the point here is to make you more efficient. Even if it only takes 30 seconds to apply to a position, that's 30 seconds you don't need to spent if you design computer networks for a living and the position in question is for a molecular biologist. Tools and search terms on your favorite job board will help with this somewhat but, in general, it's still worthwhile to read before you apply and that will take time.

The second is your ability to keep track of the positions you have applied to. Application log from Stage Two will be a huge help here, but pretty soon you'll find yourself spending a lot of time searching or paging through it trying to figure out if that posting looks familiar for a reason.

There is not a good solution to this problem. I began building one myself in the form of a browser plugin called "MarkApplied." This served both as a distraction from my own job search and as a project to keep my skills sharp. But -- and this can not be overstated -- the resulting product has a long way to go before it's ready for general distribution.

If you're feeling brave, you can find the GitHub repository here. Instructions for getting set up are available as part of the project.

In broad strokes, this is how it works

MarkApplied pulls a simplified version of the Application Log from Stage Two when you visit certain pages. It attempts to parse out the company name and job title from several job boards and then compares that against the log.

If there's a complete match, MarkApplied colors the job posting red and tells you when you applied (and the status of your application). If there's only a match on the company name, MarkApplied colors the job posting yellow and tells you the title of the position to which you most recently applied.

Finally, just to reduce friction a bit, MarkApplied also supports a one-click copying of a job posting for pasting into the Application Log.

You should probably not use MarkApplied

If you are not a software developer you probably shouldn't be using MarkApplied. Honestly, even if you are a software developer, downloading browser extensions off GitHub and running them locally is probably not the best idea.

That said, if you have some front-end expertise and have strong opinions about how bad MarkApplied is and all of the things you think should be done with it, drop me a line and let's talk. I think there's a niche for something like this in the market but I don't have the time or front-end chops to develop it.

Conclusions

The reality here is that automation is a double-edged sword. One of the reason the hiring process is so broken right now is that companies are trying to lean back on automation strategies in lieu of adequately staffing their HR and Talent departments.

As an individual candidate, you can respond to that problem with automation strategies of your own. But know that, as you do so, you are making the problem worse for everyone else. This is a classic "tragedy of the commons" scenario and so it seems inevitable that someone (indeed, many someones) will lean into that automation strategy, forcing everyone else to keep up or be left behind.

And maybe that will finally be the catalyst that forces a reevaluation of how we handle hiring, staffing, and talent acquisition. Or maybe it won't.

Good luck with your job search. 🥂