The Ins and Outs of A/B Testing
A recent email from Maggie Georgieva, Inbound Marketing Manager for HubSpot, caught my attention. The email linked to a recently published ebook on A/B testing: a topic of particular interest here at Ratio. The following is my summary of the high points from the 50-page guide.
What is A/B testing?
A/B testing, or split testing, is the practice of testing two variations of a particular item (typically an email marketing template, a landing page or a specific website page) to determine which version gains a greater response. An A/B test involves a "control" element - the current page or design in question - as well as a "treatment" element - the alternate, optimized version of the "control." The goal is to determine which design will result in the highest number/frequency of users taking the desired action.
What is the value of A/B testing?
The advantages of A/B testing are many. A few common objectives of A/B testing are:
- Drive more traffic to your website
- Generate more qualified leads through your site
- Convert more of your visitors into customers
A/B testing allows you to make small adjustments to your website or email marketing designs, which can pay major dividends in the area of site traffic and new business leads. A/B testing is also vital for companies that engage in online advertising, Search Engine Optimization (SEO) and/or Search Engine Marketing (SEM). Through testing, you can optimize key pages of your site to give yourself the best chance of turning paid traffic into paying customers.
7 helpful tips for A/B testing:
- Only conduct one test at a time.
Over-testing can do more harm than good. If you are running a paid search campaign on Google AdWords, for example, you should refrain from testing different ad copy variations at the same time you are A/B testing your landing page(s). By keeping the tests separate, you can track the results with greater accuracy.
- Only test one variable at a time.
If you test a new headline AND change the color of the "buy" or "request a quote" button on the page, how will you know which change users are responding to? Remember: A/B testing is not a race - it's a marathon. Take it one step at a time.
- Test minor changes as well as major changes.
The beauty of social media is that it allows the public to get to know the people behind brands. Unless they're in the field or are looking for a digital marketing agency, the average person probably doesn't care about Ratio. However, if they know that I'm the person behind Ratio's Facebook posts and Tweets, they may feel a sense of personal buy-in that promotes interaction.
- Measure, measure and measure some more.
A few of the top metrics you should pay attention to: clickthrough rate, conversion rate, leads, demo/proposal requests and sales. You should also measure offline factors such as your closing percent, cost-per-acquisition and the number of new business referrals. Monitoring all of these factors will help you refine your sales process across the board. Measurement is the key to success.
- Split your sample group randomly.
For the most accurate results, you need to make sure that your sample groups are as equal as possible. For example, when A/B testing your email campaign, you should be sure that the recipients of each version are as similar as possible. If your results are based on two vastly different audiences, they aren't going to be nearly as valuable.
- Test at the same time each day.
Be consistent with the time of day at which you test. People are typically fairly habitual when it comes to email and Web browsing, so consistency in the timing of your A/B tests is key to obtaining meaningful results. This goes back to #5.
- Decide on necessary significance from the start.
BBefore you begin testing, determine how much change in response you need to see before you actually implement the changes from your test(s). A 0.5% change might not be worth the trouble.
Get started A/B testing
Ready to get started A/B testing? There are lots of great tools our there that can help you. HubSpot's software handles split testing; along with a slew of other cool features. Google's Website Optimizer, Adobe Omniture and Optimizely, to name a few.
So there you have it. Thanks to HubSpot for sharing such a great informational resource. I would love to hear your feedback on how you have used A/B testing to grow your business. Likewise, if Ratio can be of service to you for A/B testing or anything else, please don't hesitate to contact us.
Until next time,