Book cover designs need A/B study capabilities. I’m a perpetually aspiring author and a designer. I can’t help watching the Amazon Kindle store become overwhelmed with TERRIBLE book covers by the unending wave of self published authors. People judge a book by it’s cover, and professional covers see a ~35% increase in sales.
Clearly covers are important, but retailers have yet to provide tools to optimize book cover design. There is no real way to experiment. Yes you could upload a different cover each week and see what happens, but even then you cannot see how many glance views your cover had and how many conversions resulted from it.
I’ve often heard that writing is a business and a book, the product. I would push that further and state emphatically that for new authors writing is a startup company and you must be able to experiment to know what your audience will respond to.
Today, I was thinking specifically about how to A/B test a book cover design. If you’re not familiar with A/B testing it’s a process were you put up two versions of something and measure which one results in the desired action more often. This shouldn’t be hard for a company like Amazon who already mercilessly A/B tests even the most minor change.
How to A/B test book cover design in a retail environment
- Let publishers/authors upload multiple book cover designs
- Provide publishers/authors with custom URLs for the different designs so they can market each one separately if they want to.
- The retailer should measure how many glance views/impressions each cover gets on search, similar books, best seller list and every other location available, but group locations by image size to keep the variables separate. It would be interesting to see if image size makes a difference, but also what attributes of a cover design work better across all books in each of the locations on a web site.
- Provide metrics to the publisher/author about which cover is doing best, or simply start using the more effective cover if the author selects an option to do so.
- Allow a new cover to be added and a percentage of all glance views/impressions to be directed to that cover for a limited period of time to see if it has a better conversion rate (for clarity a conversion rate is how often a desire action is being taken).
Much of the A/B testing process could be automated. Just allow for up to 6 cover designs (Facebook Ads anyone?) and then see which one performs well. There might even be a case for different aspect ratios on book covers in various locations to generate higher sales (mobile device horizontal view).
Better covers will probably lead to more sales, but it doesn’t necessarily mean better books. That’s a different topic altogether.