Last updated on October 14th, 2016 at 10:15 am
A/B testing has long been a go-to marketing strategy to learn how certain content or social media posts fare with your audience. But are you looking at the right metrics?
Quicksprout.com recently published, “How to come up with winning A/B tests using data” which features an infographic (below this post) that provides a step-by-step method businesses can use to improve their testing approach and, in the process, work to increase conversions.
Value of gathering and analyzing data
The key to improving conversions is to analyze your data and then develop an action plan, notes entrepreneur Neil Patel, authorof the article. He notes using tactics, such as better headlines or bigger buttons, is not the way to go, stating firms should focus on the process instead.
“If you just base your tests on your gut feelings, you may find a few winning variations, but chances are, most of them will lose,” writes Patel.
He also suggests “drastic changes” usually have the most impact. But before making any changes, it’s important to first gather data and do analysis before running your A/B tests.
According to Patel, A/B testing success is as follows:
[Number of tests run + Percentage of tests that provide a win] + [Avg. sample size and impact/Successful experiment] = Provides idea of total business impact
Key steps to improve conversion rates
The infographic highlights 7 key steps to data gathering and analysis which businesses can use to improve on their testing methods and, subsequently, conversion rates.
- Heuristic analysis: Evaluates each page’s content and user experiences.
- Technical analysis: What browsers are equaling the most conversions? What about devices? Are successful conversions happening with smartphones, tablets or desktops? Website speed is also evaluated. What’s the ROI? Is it worth fixing?
- Web analytics analysis: What are your users doing? What is the impact of each of your website’s pages and features? Looks at where your “site is leaking money.”
- Mouse tracking analysis: How does a visitor use a mouse? What’s being clicked? How far down are they scrolling? If they stop at a certain point, this may provide some insight on what to improve.
- Qualitative surveys: Are you using exit surveys? What about on-page surveys? Ask users about their experiences using specific criteria including which page to place the survey and establish criteria of when to show the survey. Be sure and script your own questions, don’t use a template.
- User testing: Provide tasks and quietly observe what people do in response. Let them take the lead and see what they say.
The final step is to collect all of the information you’ve collected and pull it all together in a master action sheet and categorize problems – place them in “buckets” to determine your next course of action. The first one would be to fix the obvious issues.
Other suggested “buckets” include pinpointing findings to give “standard tests”, one for findings where there is no clear solution, compiling “high-level issues” that require more info and adding new instruments to analytics.
Going beyond the A/B test
Are you getting the most out of your B2B company site or is there room for improvement? If you aren’t seeing the conversion rates you’d like, it may mean the time is right to try a different approach to testing and maybe your overall marketing strategy.
There are numerous benefits to A/B testing to see what your visitors respond to, but keep in mind, as a B2B company, since you will have a longer sales cycle than a B2C, you might want to consider going beyond A/B testing and take an all-inclusive approach to ongoing site improvement.
Statistics suggest B2B buyers may go 90 percent of their way through research before reaching out to a salesperson. Ideally, you want to be that person they reach out to when they are ready.
Whatever your approach, the bottom line is you might not want to narrow in too thin on your testing as you might miss important data in the process. But on the other hand, you don’t want to be too broad as this can turn out to be time-consuming and expensive. By taken a data-driven approach, you can hopefully find that right balance when deciding what to A/B test.
Does your B2B firm conduct A/B tests? If so, have you thought about how you are approaching the task?