A/B-testing automations

Once you've whipped up your E-goi follow-up series or sales funnel, you may want to find out whether some tinkering can drive better results. For example, say you're nurturing your leads with kitty tips to set them up for downloading your "All about cats" ebook. Will you get more downloads with five nurturing emails or simply two? An A/B test nails it down for you:

1) Start by ensuring your E-goi automation workflow (ie, your automation) includes an "Add tag" action which labels anyone who reaches your goal (in our case that tag would be something like "got ebook").

2) Create the whole email series. This will be path A. Then, right next to it, create an alternate email series (with perhaps fewer emails, punchier content, etc.). This is your path B.

3) Head over to the left-hand panel and drag "A/B test" onto whatever kickstarts your A and B paths. In our case, since the email series is triggered right after a subscription, we'd drag-and-drop the "A/B test" action over the "Subscription" trigger.



E-goi will auto-connect "A/B test" to the "Subscription" trigger and then ask you when the test ends. Choose "Until a number of contacts are split", enter the amount of contacts to test and pick the segment which will decide the winning path. In our case, that would be "got ebook" - which means the email series driving more contacts to "got ebook" wins. If there is a tie, path A is always the winner. Hit "Done" to finish the test setup.

4) Now connect the A/B test's left circle to the first action in your email series. It'll be marked with "A". Then connect the A/B test's right circle to the first action in your alternate emails series. It'll be marked with "B".



5) You're ready to roll! After the test splits the amount of contacts you selected, E-goi will check which path brought more people to the "got ebook" segment. From then on, every new contact shall be routed through the winning path.



The animation below shows you how the whole thing works.



Awesome! Are there any other ways for the test to end?

- Until a specific date and time: Similar to "Until a number of contacts are split" above, except the test ends on the date and time you select.
- Forever, don't stop it: Whatever happens, the test always runs. Any contact funneled into it will be randomly taken to path A or B.


Any examples of cool stuff I can do with A/B testing?

Sure! You're able to:
- Test major changes in your marketing pipeline (eg. comparing an email-only series against an SMS or webpush-powered funnel).
- Create control groups for small changes (eg. split-testing a different subject line in your welcome email).
- Run a sweepstake (use the "Forever, don't stop it" option under "The test will run..." to randomize your contact split).


How can I split-test two emails for 1 hour so the the one with the most opens is sent to the remaining contacts?

This is a special type of split-test. Check out how to do it!