Use Little Data to Generate Test Ideas
Marketing relies on testing to improve results. While testing is a fine example of data-driven decision-making, marketers often fall short in using data to generate test ideas. Test ideas are mostly generated by a creative ideation process, untethered to any data.
For example, a credit card company wanted to improve performance of a landing page that talked about “No Annual Fee.” The conversation centered on critiquing the images and call to action buttons. We decided to bring some data in and looked at the source of traffic. Most of the volume was coming from paid search, using the search term “0% APR.” The page, which talked about “no annual fee,” was completely misaligned with what users were seeking.
The Little Data approach in this case would entail looking at the source of traffic to a site and understanding the intent — data that is easily available in Web analytics reports. Understanding the user intent helps generate a data-driven hypothesis. In this case, the page did not meet the needs of the user and, hence, had a low conversion rate.
Use data to inform your hypothesis and then overlay it with a creative process to generate tests.