- April 22, 2014
- Content Marketing
3 Tips for Technology Marketers to Get Started with A/B Testing
“Test. Optimize. Innovate.” Check!
So while I can’t take credit for this catchy phrase, I can say that I was thrilled to be part of Optimizely’s first conference last week: OptiCon 2014 – all about testing, optimization and innovation. Not only was I able to listen to some of the industry’s smartest people discussing web personalization and optimization, and participate on a panel where I discussed lessons learned from my years of testing, but TechTarget also had the honor of winning an Optie award for “Most Advanced Use Case.” We were thrilled to have been a finalist for the award and even more excited to have won, for some A/B testing we recently conducted which improved conversions 61% for our most common registration experience.
Using A/B Testing to Move the Needle for Your Business
Trial and error is by no means a new method for problem-solving, but surprisingly, many marketers do not employ testing as a critical part of their efforts.
For those of you who are doing online A/B testing, I applaud you. For those marketers who aren’t, you can’t start soon enough. From simple copy, image and button changes to more advanced user experience modifications, online testing can be an easy and quick way to move the needle for the metrics that matter to you most. While tools like Optimizely make it easy to do, you don’t need tools to get started. All you need is data and the desire to learn from that data to make better decisions.
There’s no reason not to try (or “test” ) it…
Top 3 Lessons Learned About A/B Testing
I’ve been responsible for A/B testing and optimization at TechTarget for the past 5 ½ years. In that time I’ve learned a tremendous amount, through some very successful – and sometimes unsuccessful – tests. I thought I would share some of the lessons that I have learned over the years to help you as you get started with A/B testing.
1. Don’t stop at one test wave
This is a tough one, particularly if you’re fighting for resources to help implement tests. If you see a positive result, it’s easy to ring the bell and call it a day. But we know that in many cases, continuing testing with additional iterations can really compound your success. Sometimes you’re developing a new hypothesis based on test results – so one test iteration informs another. This can really turn a good test result into a great one; but even with small wins, when you compound them on top of each other, you’ve suddenly achieved greatness and truly moved the needle.
2. View many different health metrics
With each test, you’re identifying a main success metric by which you’ll evaluate success. But it’s important to also look at other health metrics in context with the test. Is a positive result worth a negative trend somewhere else? You need to decide what is and isn’t acceptable. Conversely, sometimes you’ll see a negative result of a test, but if you look at other metrics you may actually see that something else moved in a positive direction. That kind of information can help inform future tests and ideas. Some basic health metrics you might want to continually monitor are bounce rate, pages per visit, ad click-through rate, etc.
3. Internal PR is very important!
It’s easy to get caught up in the day to day, and while you’re achieving great things, people don’t seem to realize it or really understand what you’re doing behind the curtain. It’s so important to socialize your testing program. Talk to other groups in the business about what you’re doing, what your process is, what great tests you’ve run and results you’ve seen. Invite them to share ideas or bring their problems to you and see how you can help. Try putting together short case studies or one-pagers that show highlights of success in various areas, which will not only serve as a proof point for your work and your team, but circulating this will open the door for people to come to you next time they have a business problem to be addressed. This socialization will also help drive top-down prioritization efforts. For some tests you may need resources from other departments, and without support you have to wait ages to get any help. Showing how successful you can be moving the needle for the business will surely get the powers that be to prioritize the work you need.
Some people tell me they worry that opening things up to a larger group will result in delays and unwanted ideas. The great thing about testing is that you don’t have to be the bad guy – sure, let’s test that idea against something else and the DATA will tell us what the reality is!
I would love to hear more from you about your own testing efforts. Please feel free to leave a comment or connect with me on LinkedIn.