Tuesday, 16 July 2013

Learn the Secrets and Power of Web Analytics Driven Testing

If you are a reader of this blog, there is pretty big chance that you are using web analytics data to monitor and track your website. But are you doing a great job of using web analytics data to power your efforts to test and improve your website? My name is Rich Page, and I’m here to reveal the power and benefit of web analytics driven testing!
I used to be a web analyst for many years, including for Disney Online, and I learned many great web analytics tricks, and was always fascinated with using these to test and improve websites. After making the leap into the testing and conversion rate optimization world, I quickly realized just how important web analytics were to effective website testing.
To help give you a great kick start on this, here is an overview of the benefits of web analytics driven website testing and key things to learn. These will help improve your testing efforts and give you much more ROI from your website test results (and make you look like a rockstar!)
Ben­e­fits of using web ana­lyt­ics to power and improve your test­ing efforts
First let’s get started with the benefits of using analytics to drive your testing efforts:
  • Web ana­lyt­ics helps you dis­cover, mon­i­tor and set tar­gets for your key suc­cess met­rics (for exam­ple shop­ping cart aban­don­ment rate or signup form com­ple­tion rate). These will be vital to use and mon­i­tor dur­ing your test­s to under­stand how well your test efforts are performing.
  • Web ana­lyt­ics helps you under­stand and pri­or­i­tize what pages need opti­miz­ing most on your web­site, rather than sim­ply guess­ing what you should test. For exam­ple, you shouldn’t be test­ing pages that have very low traf­fic or have a low impact on your con­ver­sion goals.
  • Web ana­lyt­ics helps arm with you infor­ma­tion that can improve web­site deci­sion mak­ing, which is often crit­i­cal to over­come prob­lem­atic HiP­POs (High­est Paid Person’s Opin­ion) who often only allow the testing of things they want, or and often incorrectly think they know what is best for your website visitors.
  • Ulti­mately, web ana­lyt­ics pro­vides amaz­ing vis­i­tor insights to help cre­ate even bet­ter test ideas. These web analytics driven test ideas will usu­ally have a greater chance of see­ing more sig­nif­i­cant con­ver­sion lifts than random test ideas (based on my experience with clients).
Key things to do with your ana­lyt­ics to help power and improve your testing:
To help give you a kick-start in using your analytics to improve your testing, here are some effective simple things to analyze and look out for:
  • Ana­lyze your key con­ver­sion fun­nels. Pages within your checkout flow and sign up process are enormously influential on your conversion rates, and you should analyze them to look for ones that have par­tic­u­larly high exit rates or drop-off rates. In many analytics tools you can even analyze form abandonment to find problematic form fields. Once you have found offending pages or page elements, run tests to better find variations that improve them and increase con­ver­sion rates, for example by focusing them better, repeating benefits and risker reducers, and also removing un-necessary form fields. These con­ver­sion fun­nels can easily be set-up and analyzed in Adobe Site­Cat­a­lyst using the ‘Fall-Out Report’ and in Google Analytics as Goals.
  • Start higher up the con­ver­sion flow fun­nel. When ana­lyz­ing your con­ver­sion flows, find the top entrance paths to the first page of your con­ver­sion flow. For exam­ple, if you are try­ing to opti­mize your reg­is­tra­tion flow, you should find out the top pages on your website that most of your vis­i­tors arrive from, and opti­mize those pages too. This will amplify any effects of the improvements you have made to your conversion funnels.
  • Review your top entry pages for any pages that have high bounce rates. These are great can­di­dates for test­ing and opti­miz­ing because so many visitors go through them and they have much higher traffic, and therefore will yield you better conversion improvements, quicker. Again, focusing and reducing clutter on these pages will improve bounce rates, along with making sure they help solve your visitor’s major needs on them. Remem­ber to also check the key­words or sources they are arriv­ing on these pages to see if there are any issues with con­tin­u­a­tion of messaging, as this can often cause high bounce rates too.
  • On pages that you are con­sid­er­ing test­ing, use your visual ana­lyt­ics reports. These will help you under­stand what vis­i­tors are click­ing on and their pos­si­ble intent (and quite often this is different than you may have expected!) This helps you under­stand and pri­or­i­tize which page ele­ments may need test­ing and opti­miz­ing the most, rather than guessing and potentially risk wasting time on elements that have low impact on conversions. Google Analytics has newly improved ‘In-Page Analytics’ as their visual report for you to use, or you can use low cost tools like CrazyEgg to do more in depth visual analysis. 
  • Improve your test plans with ana­lyt­ics insights and data. When cre­at­ing a test plan for each of your tests, you should always include the web ana­lyt­ics insight you found that led you to think of the test idea. You should also include key success metrics that will be improved (like signup completion rate or average order value), the likely impact on revenue and estimated targets to beat). This will help oth­ers in your organization under­stand your logic for the test, and help with pri­or­i­ti­za­tion efforts to get it launched.
  • Ensure strong col­lab­o­ra­tion between web ana­lysts and website test­ing man­agers.  Doing this will help increase the amount of test­ing insights being generated and increase the amount of learnings to improve testing process in the future. Therefore you should encourage regular weekly meeting and reviews with these members to review current and proposed tests to try and improve them further. Keeping a project plan with your tests is very important to review in these meetings too. This collaboration will also help to build a testing culture in your organization – key to a long term effective testing program.