There’s the old adage, “If it ain’t broke, don’t fix it.” But what about when it’s working, but perhaps not as well as you’d like? One of my favorite things about being a hybrid author is having total control over tweaking certain aspects of my indie-published books that I feel might need a little spicing up.
In this case, I have a series that I enjoy writing, gets good reviews, and sells fairly well. All good, yes? But much like needing to try on four (okay, five) different dresses before going out to dinner, I couldn’t help but feel like my covers could use a fresh look.
With the goal in mind of appealing to new readers unfamiliar with my series, I knew I couldn’t just poll my current readers on their favorite cover options. Instead, I turned to BookBub Ads, and ran a series of A/B tests to help me see which cover would win out. I ultimately launched new covers for each book in the series, which has already driven stronger downloads of my permafree first in series and has increased my read-through rate.
The Rules of Battle
My business manager Alan Burness and I decided that the goals for the cover rebrand were quite simple — we wanted to increase my sales and my read-through rate for The Althea Rose Series. Another one of my full-priced series has a 59.9% read-through rate, while this series with a permafree first book only has a 7.9% read-through rate. Yet my reviews and feedback are comparable to some of my more successful series. So we decided to take a critical eye to the covers.
We chose to A/B test the cover designs using BookBub Ads because our primary focus was to see how many people unfamiliar with my brand would click on each cover design. A poll to my existing readers wouldn’t do.
The beauty of using BookBub Ads is that it takes out many of the variables that you’ll find when you run ads on other platforms like Facebook. When you run ads in a BookBub email, you’re being handed a dedicated audience — a narrow list of readers who are already there with the sole purpose of clicking on books that might interest them. With Facebook, for example, you have no control over when your ad gets shown, and it might be sandwiched between two political ads that leave your potential buyer in no mood to purchase a book.
We did run a Facebook poll for my readers, which was an already known demographic for us, but that didn’t help us gauge interest from new readers.
When creating the ads, it was vitally important that there were no variables other than switching out the book cover for each A/B test. When designing the ads, we chose quotes from previously successful marketing campaigns, mimicked BookBub’s synopsis style, and used Goodreads as our social proof by including the number of five-star reviews in the ad creative. We did swap out the background for the ads in different rounds, in case for some reason people didn’t respond to color or font choices.
When setting up the ads, we targeted readers interested in big name cozy mystery authors who commonly showed up in my “Also Boughts” on Amazon. We also used yasiv.com to discover books linked to my series — this shows where my books appear in the “Also Boughts”.
To note, for the first round of testing we also included my name in the author targeting, but quickly realized this could confuse my current readership who might click the ad thinking that I had a new book out. In the subsequent rounds we removed my name, targeted 1-3 other authors, and also targeted the Cozy Mystery category. The goal of targeting multiple audiences (e.g. more than one author) was to appeal to a wider group, so that one readership didn’t solely click on a particular type of cover.
For a budget, we kept it low and started at $20 per ad. If the results were close, we doubled the budget to gather more data. If there was a decisive winner, we deemed it unnecessary to spend more. Also, let’s be clear — the A/B ad testing was not a waste of time or money in regards to generating sales. Advertising is advertising, and we saw a solid number of downloads on my permafree first in series, which justified our cost per click.
When choosing CPC vs. CPM bidding, we experimented with both, but decided that CPM was more effective for us because we could show the ad to a similar number of people (if the ad was shown to 5,000 people, we could see which was more effective based on how many clicks it generated). For the CPM, I would go over the maximum bid recommended in order to get the ads out simultaneously as fast as possible, again to reduce any variables on time of day. We wanted all ads to be served around the same time.
The Battle Begins
The original One Tequila ad with the large tiki head was already running, and we replicated it with the tequila shot cover.
Cost: $40 total
Results: The original had a 2.94% click-through rate, and the new version with the tequila shot had a 2.97% click-through after getting 6,000 impressions each.
Verdict: We didn’t totally love this first new cover, and felt the original was strong. After dipping a toe in the pond, our stats suggested that it would have a similar click-through rate and that didn’t meet the goal of this battle. Luckily, our designer Keri Knutson at Alchemy Book Covers & Design had more options for us.
In this round featuring a completely different design, we removed my name from targeting for the reasons stated above.
Cost: $80 total
Results: The original cover had a full percentage higher click-through rate — each book generated the same number of clicks, but original cover only needed 8,000 impressions to achieve these clicks, while the new cover needed nearly 12,000 to achieve these clicks.
Verdict: We needed a different cover. However, I personally found this new cover appealing and decided to play with it more. Unfortunately, BookBub Ads do not report statistics for, “I don’t have a reason, I just kind of like this new cover.”
Yes, I’m aware that I only had Keri add a border. A girl likes to accessorize, okay?
Cost: $40 total
Results: The newer cover had a slightly better click-through rate at 3.12% compared to 2.85% on the original. The new cover received 6,964 impressions with 217 clicks and the old received 5,684 impressions with 162 clicks.
Verdict: At this point, I’d thought we honed in on a cover, but I still wanted to try one more — this time with a typography dominated cover that removed imagery. In a genre where everything from witch hats, cats, knives, and typography-only designs dominate covers, I figured that I had some room to play with different options.
At the same time, we also ran a Facebook poll to my current readers asking them to vote on the covers in Round #3. The results leaned toward the new cover, but only by about twenty votes or so. Again, our goal was to grab new readers’ attention, not current readers’ attention. Additionally, the ensuing slew of commentary on what I should do for the covers instead left me reaching for a cocktail. All in all, I’d say the Facebook poll ended up being the most confusing way to solicit feedback that I tried.
Round #4 – Wild Card Round
In this round, we decided to give each advertisement a solid run with $40, doubling the original cost of $20 that we started with.
Cost: $120 total
Results: This time, there was a clear winner — the version with the large typography had a much higher click-through rate than the others with a 5.2% click-through rate.
Verdict: Who would’ve thought a fourth-round wild card would bring the competition down?
Just to be certain we were heading in the right direction, I had Keri mock up a typography-heavy cover for the second book in the series, which we then submitted to its own A/B testing in the fifth and final round.
Cost: $40 total
Results: The click-through rate was 0.58% for the original and 1.21% for the new after serving 2,200 impressions each. Again, the big typography was the clear winner!
Verdict: We knew the click-through rate wouldn’t be as high as the first test, since this book wasn’t a first-in-series, but I am still happy we tested it. It helped me to solidify my direction with the covers.
Thanks to the tests we ran, we confidently launched a new cover design across the series. The new designs have only been live for a few weeks, but we’ve already seen a 40% increase in free downloads of One Tequila from the weeks before launching the new covers. Additionally, I’ve seen an increase of series sales across all vendors, and will be monitoring my read-through rates over the next few months as the results filter through the series!
All in all, I found BookBub A/B testing to be a useful tool to help lead myself and my designer in the right direction. In some respects, it may have just solidified my own gut feelings on what I personally found appealing. However, it was enough to push me toward an end result that, ultimately, I am quite happy with. I think that no matter what, a huge benefit to having these tools as an author is being able to test and tweak your products. There’s no reason to give up on a series that might be underperforming for you, even though it gets good reviews. Testing out new covers can be just the thing to put the wind back in the sails of a slow-moving series.
Want to share this post? Here are ready-made tweets:
Click to tweet: This is a such great way to A/B test a book cover design! #pubtip http://bit.ly/2KH57kI
Click to tweet: See how one author A/B tested her book cover designs to boost series sales! #writetip http://bit.ly/2KH57kI