Overview
A good referral program drives quality new customer acquisition. A great program activates a significant and broad base of participants. The different types of advocates that refer (and their friends) represent new types of segments that are essential for a healthy brand.
Extole allows you to test and target content, rewards, and rules to motivate varying roles and segments or to compare configuration options. Our goal is to increase overall activity by tuning programs to your segments, new countries, or even new audiences like influencers, advocates, affiliates, and partners.
If you're looking for more specific information on how to test your offer, please refer to our article on A/B Testing Your Offer.
Quick Glossary
Program: A set of campaigns targeted to a given audience for a given outcome. A company can have a single program, like an evergreen Refer a Friend, or can have additional programs targeted to different groups of participants (i.e., influencers, affiliates, international customers, etc.).
Campaign: The content, rules, and reward configuration for a specific version of a program. There can be more than one campaign in a program (for testing). All campaigns in a program target the same audience for the same given outcome. It is important to note that you would only have more than one live campaign in a program if you were trying to A/B test.
A/B Test: A user experience research method. Typically A/B tests involve randomly exposing your customers, or a segment of your customers, to two variants of an experience. The purpose of the test is to determine which variant produced a better result (i.e., more click throughs, more registrations, etc.).
Burst: A limited-duration version of campaign typically used to run a special offer. For example, if you wanted to promote an elevated reward to coincide with a holiday, you could create a burst of your evergreen campaign that would last for a few weeks before reverting to your typical offer.
Targeting: The mechanisms for determining which program is given the control over handling an event in the system.
Journey: The mechanism we use to make sure people who see a specific campaign continue to be served content from that campaign unless explicitly moved to a new one. For example, if an advocate shares on a campaign whose reward is $10 off, then the friend who clicks will be targeted to a landing page that supports that reward unless that particular campaign is stopped.
Knowing When to A/B Test
The need for an A/B test typical comes from the desire to test a hypothesis about your program. Let's say that you and your Extole Team have set up your core Refer a Friend program and you are now looking to optimize it. To do so, you may identify a particular part of your customer journey where you'd like to see better results.
For example, you may want to know if changing the subject line of your Share Email can positively impact your open rate. In this case, it could make sense for you to set up an A/B test with two versions of your Share Email that have slightly different subject line copy. The test will allow you to determine which copy helped drive a higher open rate among your included customers.
To determine if an A/B test is right for you and your program, reach out to your Extole CSM for more information.
Setting Up a Test Within a Program
Please take the following steps when configuring your test in My Extole.
Log in to your My Extole account.
On the Programs page, find the program for which you would like to run your test. Hover over the campaign you'd like to test and hit the duplicate icon.
Once you've created your duplicate, be sure to rename it in a way that makes it clear this is the "variable" campaign in your test. You can preview and make changes in our Campaign Editor and, at this point, your campaign will not be launched.
As your are editing your duplicated campaign, the Apply Changes button on the top left corner of the editor allows you to save your changes as you work. If you are not immediately ready to go live, but would like to save and return to editing at a later time, click the Publish Changes button at the top of the page (this will not launch your campaign). Note: If you are testing an offer, please refer to the checklist in our A/B Testing Your Offer article.
When you are ready to launch your test, open it in your Campaign Editor and find the menu in the top right corner of the page. From the dropdown, select the option to Go Live.
A message will pop up to ensure that you want to Go Live as a Test. Clicking this button means that all other live campaigns within your program will continue to stay live and the campaigns associated with your A/B test will begin splitting traffic. Because you are running a test, you will want this to be the case.
Monitoring your Test
Once your test is up and running, you will see two (or more) live campaigns running against each other on your programs page. To compare performance, navigate to your program stats dashboard to view your test campaigns side-by-side.
Concluding your Test
Once your test has run its course, you can choose to either pause your test campaigns or end them. Pausing your campaigns means that new customers will not see them, but customers who are already participating (i.e., friends) will still be able to finish their journeys and earn their rewards. Ending your campaigns means that no customers will see them, but rewards that have been earned will still be issued. If you are unsure of which route you would like to take, contact your CSM.
You can then work with your CSM to determine the outcome of the test and next steps. Using the program's dashboard and Extole reports, you will be able to draw conclusions about your test and which experience performed better according your metrics. For example, if you tested two different Share Email subject lines, you will be able to determine which copy correlated with a higher open and/or click-through rate.
Frequently Asked Questions
How is traffic split between my A test campaign and B test campaign?
For A/B tests, the Extole campaign targeting engine uses a pure random number generator when each user begins their session. No other inputs are considered. This is true for all users who interact with your test, regardless of whether they do so on browser or mobile.
Will customers who have already interacted with my program continue to see the campaign they have already experienced?
If you would like current customers to continue to experience the campaign they have already seen, the steps detailed above will be perfect.
If you would like to ensure that any traffic, new or old, gets split for a test regardless of prior history, you will want to complete steps 2–6 again to test two new campaigns against each other. You will also want to end your evergreen campaign so that it stops receiving traffic. To do so, navigate to your Programs page and open the menu on the right of the campaign you want to end, then select the End option.
What if I host the promotions I want to test?
If you control the promotions or calls to actions (CTAs) you would like to test, we have two options for you.
Make your promotions and links generic and set up an A/B test.
Recommended: Change your promotion for a limited time and run a burst. A burst will pause traffic to your live campaign in favor of a new campaign that features the promotion you would like to test for a limited time. Once your burst is launched, you will want to end your original campaign so that it stops receiving traffic. To do so, navigate to your Programs page and open the menu on the right of the campaign you want to end, then select the End option.
Do I need to re-tag my site in order to run a test?
No. If you have already added Extole tags to site you do not need to do any additional set up in order to run an A/B test.
What if I want to target a specific audience for my test?
This is possible using our Promote to Audience feature if you would like to initiate your test via email.
You will need to set these up as their own programs in My.Extole. You can select the audience for which you would like to run the test, then send group A to test A and group B to test B. For more information on how to set up this scenario, please reach out to your CSM.