FMCG house of brands spent a ton of money on market research crafting the persona of their customers for each brand in their portfolio. I even had an opportunity to run a data strategy workshop for the personas, which was a super fun endeavour on its own (ask me about it please). The workshop had around 10 participants which led to 8 different messaging ideas that could be apt for the each persona. You are probably thinking that’s a lot and you are absolutely right. We had a lot of discussions and friendly banters and brought it down to 6 messages, phew. 

6 was still a lot and multiplying that by 12+ brands, it is a whackadoodle amount of creatives. So we took a data-driven approach to find a way to cut down production cost. The client was recently transitioning from TV to targeted programmatic video. So, I took the opportunity to bring on display channel to test out the messages, mainly because it was cost-effective (yeah, less than 2 USD CPM) and we drafted our hypothesis as to which one was expected to work and why and ran a test for 1 month. 

Iterative tests were conducted, swapping images, titles and CTA all whilst sticking to healthcare regulations (not a small feat). We made interesting observations:

Personalized messaging did not work for healthcare brands, generic messaging won the best engagement rate. 

Making CTA shoppable i.e, sending the users to retail websites like Amazon boosted e-commerce sales while also boosting brand lift. Win, win I would say. 

We used a cool technique on a dating website where smokers matched nonsmokers for a smoking cessation brand that displayed a message saying “Would you like to quit smoking, your date doesn’t smoke” → So cool, so native, it worked.

Interstitial banners were so intrusive, they caught attention, they shot up engagement rates, but the on-site metrics were terrible. Do not spam your customers for metrics please. 

Positive messaging worked for NA and EMEA and negative tone of voice (for example: don’t miss out) worked really well for the APAC region. Interesting cultural context right there. 

All in all, if you are not testing you are missing a lot of learnings on how to talk to your customers. Work with me, I am great at building long-term learning that we can carry over the years.