October 22, 2015 How to Test and Optimize with Change.org October 22, 2015/ Mallory A. Russell There are a lot of times in life when the best advice is to go with your gut: choosing a partner, figuring out what city to live in, or deciding what to watch on Netflix on a Friday evening. And there’s one very important time when your gut should be completely ignored - when it comes to your digital engagement strategy.Here at Change.org, we talk a lot about the importance of testing on a number of different levels, and that’s because you can make a much more informed decision about your campaign with some data to back you up. And I don’t know about you, but I find it really difficult to get data from my gut. We’re focused on ensuring that you put your best foot forward as an organization, with the most compelling campaign possible. We have a number of different tools that will help you do that, and as always, your client manager is available to answer any questions you might have. Optimizing your Sponsored CampaignWhen you’re running a sponsored campaign on Change.org, there are a number of variables that we can test to find the most compelling and highest converting version of your petition. Depending on what you choose to test, your client manager can report back on which combination of headlines, photos, and body copy performed the best on the platform - and that winning combination will automatically be shown to our users once it reaches statistical significance.Content VariablesHeadline / title:We often use the term “headline” or “title” interchangeably. It is a component of a campaign that can be tested using multivariate testing. We recommend testing up to 3 versions of a headline at a time.A headline should contain between 30-50 characters (including spaces).Strong headlines typically use compelling, action-oriented language to grab the reader’s attention and draw them in.Possible headline testing strategies may involve opposing themes such as “hope vs. fear” or “shocking statistic vs. a more personalized appeal”Sponsored Campaign imageAn image must be a landscape (horizontal) image that’s at least 1600x900 pixelsThe image will display in thumbnail size on the sponsored campaign, making text overlays or captions difficult to read. Avoid including text in the image whenever possible. Instead, leverage the headline for your most important text.Possible image testing strategies may include “photo vs. graphic illustration,” or determining whether a photo of a person looking directly at the camera performs better than an image of someone looking away from the camera. Alternate testing may involve a group photo vs. a photo of an individual, or emotional expression such as “happy vs. sad.”If you’re testing images of people - test attributes such as:EthnicityAgeNumber of people in the photoLooking at the camera versus looking awayLocation of peopleAttitude (e.g. happy vs. sad)If you’re testing images of animals - consider testing up to three different animals (as long as the description is still applicable)Other examples may include: a baby wolf, a family of wolves, an adult wolfBody copyThe 300 characters that are shown as the first paragraph of your sponsored campaign on Change.org are the most important, but you might not know exactly what resonates with potential new supporters.Try testing different versions of this copy, with similar strategies to the headline tests above: personal stories, voice of authority, positive vs. negative, etc.Campaign testingDepending on your organization’s goals and contract size, it’s possible to run multiple campaign issues simultaneously to test which resonates best with our audience.Possible campaign testing may involve a simple test of petition framing versus pledge framing. Similar to headline testing, overall campaign testing may seek to identify stronger performance between content framed around fear versus hope, or statistics versus a personal story.You might even take a look at your content calendar for the next six months and want to test around different issues entirely (i.e. one legislative campaign, and one campaign targeting a corporation or other entity).To take it a step further, depending on your data resources, supporters’ email records can be tagged to identify their campaign preference to help drive segmentation strategies.It can be really interesting and useful to track the long-term performance of supporters recruited on a variety of topics. Perhaps you find an issue that consistently attracts people who become monthly recurring donors or consistently active advocates. In that case, because you tested and then tracked your data, you know how to finesse your recruitment strategy for higher ROI over time.Of course, testing on Change.org is just a small sample of the testing that a robust digital engagement program should be undertaking on a regular basis. Once your campaign has wrapped up and all your new supporters have been delivered, don’t just rest on your welcome series laurels, as there’s plenty to test in every email. You can read more about it in our "How to build a welcome series" guide. And don’t forget Change.org’s motto: A.B.T. (always be testing, of course).