Web designers know the power of imagery. A well-placed picture or a splash of colour can be the difference between someone clicking on through the site, or leaving without a second thought. Anyone can have an opinion, but it’s not always clear which images will have the best results. Happily, some cheap and easy testing can tell us the answer.
A/B split testing is the process of creating a different version of a web page, and measuring how well the new version performs compared to the original (or control group). Because it measures results with minimal effort, split testing is a blunt but decisive tool – it will clearly show you which version performs better, but will not tell you why.
Split testing has been around for a long time. Google Web Optimiser first made it possible, but the additional tracking was a bit onerous. Google has made split testing much easier by introducing Google Content Experiments– a tool that combines split testing with other analytics within the usual Google Analytics interface. Content Experiments has just been released to all Google Analytics accounts, but here at Quirk we were lucky enough to get early access to the product testing in beta phase, as a Google Analytics Certified Partner.
Quirk Education is a specialist online educator that offers digital marketing courses to students around the world. The target audience includes current and aspiring digital marketers in all industries, as well as business people and professionals in related fields such as PR, creative industries and management.
For Quirk Education, the most important conversion page on their site is the course details page. This is where visitors can sign up for a course or download a course info pack. It is the core of Quirk Education’s online efforts.
Quirk Education needs to maximise their conversion rate to get as many course registrations as possible. The analytics team took up the challenge to improve the conversion rate on a course page with minimal effort and expense. Google Content Experiments was used to test a simple image swap to discover whether changing one image on the course information page would increase the conversion rate on that page.
The original pages featured a photo of a smiling student; the test page replaced that student with the Digital Media and Marketing Association (DMMA) logo. (The DMMA is a South African body that represents the digital marketing profession and also endorses the courses provided by Quirk Education).
Original version: Smiling student
Variation page: DMMA logo
Two different test pages – one with image of the lady, and the other with an image of a DMMA logo – were created for each of these four courses:
- Digital Marketing course
- Search Engine Optimisation course
- Writing for Digital course
- Social Media course
After these pages were created, a few lines of tracking code were implemented on the variation pages. The Google Content Experiments code then served the different versions to different people (taking care to ensure that each visitor would only ever see one version to avoid confusion) and recorded the results.
Results And Analysis
The pages with the DMMA logo received a 3.69% conversion rate, compared with a 1.29% conversion rate for pages with the image of the student. Of course, our results don’t tell us why the visitors who were served the DMMA logo were more inclined to convert. We can infer, then, that the logo of a well-known digital marketing body lent an additional level of credibility to the Quirk Education offering. These results are not sufficient to conclude that any respectable third party logo will improve conversion, but they do tell us that in context, this particular logo was very effective.
To conclude, the Google Analytics Content Experiments module looks to become a real asset to our business, and we’re looking forward to more experimentation!