Analytics and A/B testing: It’s a match made in data heaven.
Just like Tom and Jerry, Simon and Garfunkel, or peanut butter and jelly (depending on your taste buds), each is fine on its own, but combining the two makes the real magic happen. But while the other pairs’ magical synergies may remain inexplicable, it’s obvious why analytics and A/B testing go together so well.
Analytics provides the rationale and the guiding light for your testing decisions. It allows you to decode your users’ behavior and see where your testing will have the biggest impact. After all, if you’re going to make the effort to set up A/B tests, run and monitor them, and analyze results and implement changes, you want to make sure you’re focusing on the right elements. Forget the “finger in the wind” approach where tests are identified based on your gut feelings – if you want conclusive results with a tangible impact on your business, you must select tests based on tangible criteria.
Analytics data helps you establish an efficient, data-driven strategy upstream of your tests. And it also enables you to get more downstream from your tests: it permits you to dig deeper into your results, see how different user segments are reacting to your different versions, and implement more effective optimizations and personalization.
To illustrate the power of analytics and A/B testing when combined, here are four great ways to use your website analytics for more effective tests, and more actionable results.
#1. Identify pain points and optimization levers
Your analytics performance dashboard should give you an overview of your website’s main KPIs. Is there a certain KPI that’s not performing to your expectations? Even if all is performing within your targets, perhaps there’s one KPI that’s not growing at the same rate as the others. Once you identify the site KPI you want to improve, use your analytics reports to zoom in on specific points like:
- Performance over time – Has this KPI stagnated or decreased gradually over time? Or was there a specific point where things began to go sharply downhill?
- Traffic sources – Are certain traffic sources performing worse than others for this KPI?
- User pathways and page sequences – Where are your users leaving your site, or going back and forth between pages?
By studying your analytics data, you can zero in on what to address via testing, and precisely where on your site to run experiments. (A performance audit performed by analytics experts is an effective way to comprehensively review your site and identify any problematic areas or opportunities for improvement.)
#2. Uncover new pockets of potential
Let’s assume your KPIs are all up and to the right, and you see no need for immediate and urgent correction. (Great news!) Another way analytics can help you identify impactful tests is through competitive intelligence insights. Certain analytics solutions like AT Internet’s Analytics Suite feature integrated industry benchmark data, so that you can compare your site’s performance against your industry competitors and see where you have room for progress.
For example, your 30% bounce rate might seem acceptable as a KPI, until you notice your direct competitor’s bounce rate is only 20%. You see that you have room for improvement, and it only takes going back into your analytics reports to identify which levers to pull with your tests.
#3. Verify and qualify your test results
After identifying which tests to implement, you’ve ideally prioritized them and transformed them into scenarios (learn more about this phase). After your tests have run and you begin to get results, digital analytics data is also essential for your analysis, as it permits you to see the full picture of what your users are doing – and not just their behavior at your test location. You’ll be able to see if your test variations are having an impact further down your funnel or elsewhere on your site.
For example, if your A/B test involves optimizing the very first step of your conversion funnel, you can use your analytics reports to verify whether test variations have had a positive or negative effect on later or final steps, and therefore qualify the full scope of your test’s impact.
#4. Segment results for more pinpointed insights & personalization
Another key way to use analytics to truly understand your A/B test results is visitor segmentation. Segmenting your test data can reveal deeper and unexpected insights which may vary from the overall test results. You’ll be able to better understand how your test variations are affecting different user populations… and better optimize and personalize for those different groups.
For example, you might segment your analytics data based on customer status (new or returning) to see how different customers are responding to a certain test feature, or segment based on device used to see how mobile users interact with a test variation versus your desktop visitors.
As we’ve seen, by making digital analytics an integral part of your testing and optimization cycle, you can get even more value from your testing efforts. And thanks to the integration between AB Tasty and AT Internet analytics, all the heavy lifting is already done – get your testing data directly in your analytics reports!