Automated Split Testing

Magazine Cover Testing

February 16th, 2011

Test Your Magazine Cover

Through resources available on line, it is more possible than ever before to
get a quick, inexpensive read on consumer response to your newsstand
covers. Whether you are testing your covers on line via your website,
through an email survey, or through a newsstand split, there are a few things to keep in mind:

1. Isolate Your Test Elements. If you want to tell whether an image or
concept will work best for your cover, a person or object, a full bleed or a frame, or whether it is better to highlight a personality profile or a section on health tips, you need to decide what it is that your are trying to ascertain and change only that element. Every other element on the cover should remain the same. Too often a publisher, while trying to get a read on a specific cover element, will create two completely separate covers and test them against one another. This approach will not yield any results that are actionable over the long term.

The exception is when you do have two completely different cover concepts proposed for an upcoming issue and are trying to decide between the two of them. In that case it is perfectly valid to run an on line test to see which has greater newsstand appeal, and to use the winning cover on the issue in question. It won’t give you information for future issues, but it will show you the direction to take for the issue coming up.

2. Test One Element at a Time. Once you have isolated your test element(s), you want to make sure that only that element is changed in your test split. Don’t change background color and type of image, or type and size of image together, or image and bleed together. Doing so will make it impossible to quantify the impact that each element has upon sales.

3. Test a Significant Group. The online program Decision Analysis provides a statistical model that can help in setting up price groups. Using it, you can see that for a desired confidence level of 99 percent and maximum error of only 2 percent, at an average sale of 5,000 copies, you will want to test about half your draw; at an average sale of 10,000 copies, you can reduce your test group to about 33 percent; and at an average sale of about 15,000 copies, you can test about 20 percent of the draw.

In other words, there is an inverse relationship between the size of your newsstand sale and the percentage that you must use for a test group in order to get statistically significant results. If you up your scale to 100,000 or 200,000 you can test as little as 16 percent and still feel confident. And that’s good, because then you can get to multiple splits.

For more tips, visit: http://www.singlecopysales.com/covers/

Linda Ruth helps publishers with audience development online and at retail. Her books can be found at www.bookdojo.com.
Article Source

Read more about...

Magazine Cover Testing