The Value of A/B Testing Online Job Posts

A/B testing might be something you’ve heard of in a marketing or design context before, but have you ever considered applying the method to your online job posts to get more applications or to improve the quality of the applications you receive?

Many recruiters and small business owners feel frustrated when their job posts don’t perform as well as they had expected, but are often left scratching their heads about how to improve the number and type of applicants they receive. At LocalWork.com, we’ve seen firsthand how effective A/B testing can be for companies seeking talent, and have some advice for anyone looking to improve their posts without spending a lot of money in the process.

What is A/B Testing?

A/B testing is a simple, cost-effective optimization strategy where two versions of the same element are tested simultaneously in order to see which performs better, according to a metric that defines success.

For example, an ice cream company might A/B test two designs for a pint of mint chip ice cream to see which one sells better: A) a design that emphasizes the mintiness or B) one that emphasizes the chocolatey chip-ness. In this case, the A/B variable would be the packaging, the ice cream would be the unchanged element tested, and sales would be the performance metric. After conducting the test, the shop would move forward with the design that sold the most ice cream.

Using an A/B testing method for job posts can help you optimize your ad to attract the most number of qualified candidates.

What Should You A/B Test in a Job Post?

Here are some variables that you can consider testing in a job post to attract and convert the most candidates into applicants AND draw the proper type of quality of applicants.

• Job Title/Headline
• Position Title
• Feature Photo
• The wording of the Responsibilities section
• The wording of the Benefits or Compensation section
• The wording of the Requirements section
• The amount of text on the post
• The layout of the post
• The order of the sections in the post
• The “voice” of the post
• Including elements that describe your company culture, like employee quotes

Creating Your First A/B Job Post Test

Once you’ve decided what to test, the next step is to build two or more versions of your job post to run simultaneously. If this were a science experiment and you wanted to determine precisely which element (job title, feature photo, amount of text on the page) produced better results in version A than B, you would only test one variation at a time. For example, you would run two identical posts that had one difference only, let’s say the feature photo.

Since this is HR and not chemistry (a wonderful fact, in our opinion!), we can enjoy the freedom of testing as many variations as we want at any given time. All we care about is finding the best way to present our job online as quickly as possible. With this in mind, consider running multiple versions of your post at once, like an A/B/C/D/E/F/etc… test. The more variations you run, the more opportunities you have to learn what works.

NOTE: Since it’s imperative that you run these variations side-by-side, or simultaneously, be sure you have access to a job board service and account type that offers an UNLIMITED number of posts at one time. If you’re currently constrained by the number of jobs you can post simultaneously, check our our Unlimited Package, or inquire with your current provider.

Dos & Don’ts

Dos

• Give your tests at least a week to run. Don’t make up your mind too quickly before allowing job seekers time in their weekly searching routine to come across your post. Remember that some job seekers work full time, attend school, have busy schedules, or take care of family. Not every job seeker looks for a job on the same days or times, so be patient and give posts a week to run before pulling them.
• Experiment with really different versions. Got a crazy idea on how to present your company benefits? Go ahead; try it out. Sometimes it’s the off-the-wall ideas that end up attracting job seekers best. You never know until you try. Just be sure to uphold your brand standards as you would with any form of official company communication.
• Take what works and create more variations. If you find that job title A works better than job title B, try to figure out why. Was it more action-oriented? Was it shorter or longer? Test your theory by creating similar iterations that are even MORE action oriented and see how they do.

Don’ts

• Run tests back-to-back rather than simultaneously. It does no good to run Version A first for two weeks and then run Version B for two weeks. First of all, you’re losing valuable time not testing Version B sooner. What if Version A is a lemon and you miss two weeks of job seekers? Additionally, if you don’t run the versions simultaneously, you run the risk of misinterpreting the effect that an outside element, like a competitor’s new job post, the time of the month, or a news report, has on your application numbers.
• Let your gut feeling overrule your results. Sometimes the results of an A/B test are surprising or confusing. You may have been sure that one version would outperform another, but the opposite is true. And let’s be honest, you may experience a slight ego-bump if you and a coworker had different ideas about the job title, and their version actually out-performed yours. Let the market (job seekers) determine which posts you run and which you re-think. Don’t fall in love with a post; fall in love with getting great results.

Ready to Get Started?

Upgrade to an Unlimited Package now. You’ll have the benefit of A/B testing an unlimited amount of job posts simultaneously for just twice the price of running only three ads at a time. By saving time, you’re saving money.