It’s a familiar story – you post a position on various job boards, get bombarded with resumes (many of whom obviously didn’t read the job description), and have to spend way more time than you’d like sorting through them all. You take hiring seriously, and you of course want to sit down with a short, highly qualified list of people. But getting from 200 candidates down to 10 is time you’d rather spend elsewhere.
This is the problem we were solving at Brill Street, a hiring startup I where I led product development a few years back. We discovered a consistent pain point at the top end of the hiring funnel and set out to solve the problem using an algorithm.
Using a combination of experience, education, skills, competency and culture, our system was able to take the 15,000 candidates we had in Chicago and tell you which would be the best fit for a given position. The system actually worked, too.
But building a system like this is expensive – you need to know algorithm design, machine learning, statistics. You need multiple developers working long hours for several weeks. Not cheap.
Luckily, you can get most of the way there. You don’t need lots of time or money to build your own candidate screening system. In fact, you can reliably screen candidates for any position for less than $100 and less than seven minutes of your time.
Step One: Write down the list of things you want a person in this role to have.
For any given position there is a list of things you’re looking for in an ideal candidate. The first step in your system is to write these criteria down.
As you’re doing this, focus on objective measurements. You want to focus on things that can be verified by the information in the resume or cover letter. You aren’t looking to gauge the ‘quality’ of a given criteria (for example, the quality of a design portfolio) but simply testing for whether the criteria exist. You’ll see why shortly.
Let’s say you’re hiring a front-end web developer. This list might looks something like this:
- Expert in HTML
- Expert in CSS
- Expert in JS
- Bachelors degree in computer science
- At least 5 years experience
- Experience in your industry
- Strong in PHP
- Strong in Ruby on Rails
- Strong knowledge in WordPress
- Familiar with Mobile development best practices
- Experience conducting user testing
- Lives in your town
- Has a portfolio online
- Has a LinkedIn profile
Step Two: Assign a weight to each criteria
Once you have your criteria, it’s time to open up a spreadsheet. You can use whatever software you’re comfortable with, but I recommend Google Docs for reasons you’ll soon discover.
Your column headers will start on row 2. In the first column, put “Candidate Name”. In the second column put “Total”. And starting in the third column, put each criteria you listed in Step One in a column heading.
Above each criteria, assign it a weight. Think to yourself, “if this candidate didn’t have this skill, how likely would I be to hire them?”
It’s important to make sure that your must have’s are weighted much more heavily than your nice to haves. If a web developer doesn’t know HTML, that’s a huge problem. I use a scale of 1-50 to give me enough of a range to weigh certain items much more heavily than others.
Step Three: Create Your Formula
Once you have your weights for each criteria, you need to create your formula for adding them up. You can of course just look up and enter the weight value into the cell directly, but I like to keep the spreadsheet clean by just using x’s. To do this, I use the following formula:
This says “For each column in Row 3, check to see if there is an X. If so, use the weighted number above it. Add all of these numbers up.”
This should give you a total score, which you can then use to rank candidates.
Step Four: Test Your Weights
Before you continue, you want to test your weighting. Drop some made up names into the candidate column, and start randomly filling in X’s. You want to make sure that someone who is missing critical criteria doesn’t end up with a high score because they have a bunch of less important criteria.
You’ll most likely find yourself making some tweaks to your initial weights, which is totally fine. What matters is that you have confidence it will spit out reliable results – otherwise you’ll be looking at every candidate’s resume anyway which defeats the whole purpose of the exercise.
Step Five: Hire someone cheap to look at the resumes
Use ODesk or Elance to find someone to review the resumes. You could even hire the high school kid who lives down the street.
Share the spreadsheet with them via email and explain what each of the criteria means. This is why having a Google Doc is valuable – no passing the document back and forth, and you can check in on their progress or do QA checks with the early resumes if you feel so inclined.
When you hire someone, I strongly suggest you ask them to do it once per day, rather than throughout the day as the resumes come in. People are more efficient when you batch similar tasks together, which will save you money.
Depending on the rate you agree to, you should be able to get 200 resumes reviewed for less than $100.
Step Six: Get some work done.
When you post positions, set up a forwarder in your email client to send resumes along to your contractor. You might also want to set up a rule to have these emails skip your inbox directly, going into your archive or another folder. They’re out of your sight, but available if you need them later.
With the system set up, you can now get to work on the things that matter to you. In the background, hundreds of resumes are coming in. Your contractor is looking at them and comparing them to the criteria you outlined in your spreadsheet. They aren’t having to make value judgments to determine which candidates you should look at, because the formula is doing it for them.
A week or so later, you can drop into your spreadsheet and do a sort on the “Total” column. If you built your formula correctly, you should have a small group of candidates who meet all of your must have criteria and have a solid combination of your secondary criteria.
We’ve found this system to work very well across a variety of positions. Looking for a social media person? Look at the number of Twitter followers or their Klout score. Looking for a backend dev? Ask if they have a Github account. Salesperson? Look for specific quotas hit.
The key is to use objective metrics – you want a series of criteria that a high school student could reliably check. While you can get more advanced, we’ve found the efficiency gains to be minimal. Once you’re down to a short list of 20 names or so, you can take the time to make the subjective value decisions yourself, assessing the quality of someone’s code or design work, etc.
Save Even More Time – Use My Spreadsheet
For less than an hour’s worth of time and less than $100 you’ve saved yourselves hours of time and have a short list of candidates you’re excited to talk to.
To help you jump start your formula, I’ve shared an example spreadsheet in Google Docs. Feel free to copy and paste it into a new document, and edit to fit your needs.
If you have suggestions for improving this process, let me know. I’d love to hear how this process works for you!