A/B Testing Your Brand

Using Facebook ads to sharpen your startup’s positioning


Brand is one of the elements of product building that is intangible. Many components of marketing can be tested scientifically with funnel analytics and A/B testing, but brand is all about the feeling that your company, your product, and your website evoke.

I’ve worked on public relations and digital marketing for enough companies and startups to know that even the most innovative products can hit an acquisition wall in a hurry if the branding isn’t resonating with its target audience.

I’m in charge of marketing for Clue, a fertility tracking app. Women are our primary target users. It would be easy to reach for the standard girly (and standard menstruation) brand elements in creating our website and our product: butterflies, flowers, and lots and lots of pink.

Ida, our CEO and founder, made the decision to design away from traditional brand elements. As she wrote in one of our job posts, “Make no mistake, we are in a world of pink body parts, but that does not make us love pink and pastels. We are modern, sexy, confident and scientific.”

We already had anecdotal evidence that our modern approach resonates with users in a powerful way. It’s not an exaggeration to say that at least 50% of the feedback we receive in app reviews and support tickets can be distilled into “Thank you for making something that doesn’t look like a My Little Pony-Lisa Frank-Hello Kitty orgy.”

But as the marketing person, it’s my job to transform the CEO’s brand vision into action. I wanted to know: could we test this brand hypothesis rigorously, and get results that help us execute it with more confidence in user satisfaction?

Yes indeed. Let me show you how I did it.

Testing the Message

Below is a snapshot of the Facebook ads dashboard for the ads we’re currently running. We targeted women ages 16–25 in four different countries. Obviously, the goal is not just to test brand but also to acquire users. The nice thing about Facebook ads is that we can do both at once.

We had a hunch that the “not pink” messaging would be the most compelling element to call out in our ads, but we also wanted to test out other claims like “best,” “best designed” and “most accurate.”

In the screenshot you can see that the click-through rate on all the ads (far right) is about the same: around 3.6%. But of course we don’t just want clicks; we want people to be compelled by the brand to the point that they actually install the app.

Facebook automatically adjusts to display the higher-performing ads more frequently. We see which brand messaging is resonating the strongest with our target audience by which ads are have a higher install rate once Facebook users click them.

Not Pink

So far, the “Not pink” messaging has been the most successful ad in 3 out of 4 countries.

As an interesting aside, the one country where “best designed” was the top performer was in Denmark. This was true for both the 16–24 age group and the 25–40 age group. Perhaps the Danes as a nation have very high aesthetic standards — or believe they do?

Either way, both ad types that were successful were based on the appearance of the app, rather than its features or performance. These results verify, quantifiably and on a large scale, the approach we’ve taken of making the “not pink” brand a critical difference in our app.

Going forward, they will inform and reinforce how we will continue to target our messaging across platforms, from our website to our tagline to our communication with users and the media.

And it guarantees that you’ll never, ever see any pink in Clue’s design.

Are you interested in working on branding and design challenges like this? If so, we’d love your help. Drop me a line at lisa@helloclue.com if you have any thoughts or comments on this article, or if you’re interested in open positions on our team.