Marketing is driven by testing new ideas and tactics, and more importantly, measuring to see which ideas and tactics are keepers (and which you toss away forever). Recruitment marketing follows suit. In this new era of talent acquisition – where candidates are like consumers – it’s imperative to think outside the box, stand out and constantly test and measure what works for your target audience.
I recently heard a presentation from Julia Spano, Senior Director at Bounce Exchange, on how to strengthen A/B testing, which is simply running a test on two different pages, campaigns, etc. to determine which performed the best and why.
Here’s how you can apply it to your digital recruiting and talent acquisition strategies.
- Make it “clean.” A clean test is testing one variable at a time; too many variables “dirty” the results. If you’re testing an email nurture campaign to leads in your talent network, one variable would be testing two different subject lines (for instance, one with a number and one without). Another test could be calls-to-action in the email, like switching the placement of an Apply Now Using one variable ensures you can attribute successful results to one key change. If you change the subject line and the call-to-action and the length of the text, you can’t really be sure what accounted for better or worse results.
- Ensure it’s statistically significant. Getting enough data to back up your results is essential, as it quantifies repeatability for a larger audience and eliminates the probability that results happened by chance. If you test a job description title, but only 12 people view the job ad, it’s not statistically significant to say that your change influenced four people to apply (although that’s a great conversion rate!). So what is the magic number that gets you to statistical significance? It can be either sample size or period of time. If you consistently have multiple openings for a Customer Care Technician throughout the year, but close the req after 50 applications, that’s not large enough to get statistical significance on your one test. But if you continue to run the same A/B test for every opening in six months, testing the same two job titles, you should be able to see a statistical significance of your test based on responses over time.
- Always have an end goal. If you can’t figure out the clear change to make after your A/B test, then it’s all for naught! Some people get test-crazy, but it’s important to have a reason for doing every test and a clear action you can take from it. Testing the color of call-to-action buttons on your career site is actionable: you should eliminate any color from your site that isn’t performing. Testing the use of a candidate’s name in an email subject line is actionable: it tells you if personalization can help open rates for key emails. For every test you conduct, consider what you want to learn and what action you would take based on the results.
For some addicting fun and great insights, check out WhichTestWon, where people can submit their A/B tests and have others vote on which version won.