How to double your Facebook Display Ad ROI with CRM data and AI

Facebook Custom Audiences

Customer Experience Director

March 1, 2018

As Customer Success Director at Tinyclues, I spend a lot of time talking to clients about how campaign intelligence can help them optimize their CRM campaigns, email being one of their favorite channels. But in our new digital marketing world, marketing cannot live by emails alone; as customer reactivity wanes and fatigue sets in, email campaigns are delivering diminishing returns, pushing marketers to diversify their CRM channel portfolios. The world of new channels can seem like a vast, expensive unknown, but these outlets can be a valuable addition to your marketing strategy, if you’re armed with the right tools. As part of a new blog series on omnichannel strategies, I’d like to  explore just one of these options: leveraging AI-based campaign intelligence to optimize Facebook display ad performance.

More bang for your buck..

Facebook display ads can yield significant ROI, but “batch & blast” methods are very expensive and can blow up your budget; it’s best to set your sights on the right audience from the get-go. Leveraging the knowledge you have about your customers, using first-party marketing data, can make a huge difference in optimizing your Facebook Display strategy much more accurately. With a paid channel, optimal targeting is the name of the game.

What’s the best method to find the perfect Facebook audience for the product you want to sell? Facebook does offer its own targeting tool, but (predictably, perhaps) this would not be my first pick. Facebook also has a lookalike feature that creates audiences based on customer behavior data points such as “likes” and “shares”; liking a picture on cats may not be the most reliable purchasing predictor if you want to sell a watch or promote vacation packages. It’s also worth considering that, in terms of audience selection, FB’s success criteria (clicks & impressions) might not always dovetail with your own (ROAS, CPA…).

CRM data graph

To find the most efficient Facebook audience possible, some of our clients have started to use Tinyclues to go beyond email or push notification targeting. Why? Because they need to identify the customers who will very probably buy the product after seeing the Facebook ad.  For that, they’re best off using a targeting approach that will comb through their rich -and free- customer data and detect the implicit, reliable indicators of future buying behavior.  Tinyclues uses deep artificial intelligence to find customers’ latent preferences and needs. It’s also completely channel agnostic, so the same solution that optimizes your email campaigns can be applied to any other CRM channel.

So how does it work? How can clients use the highly predictable nuggets in their marketing database to optimize their Facebook Display Ad strategy?

Facebook Diagram

A new world of exciting, and bankable use cases

Once you’ve zeroed in on your A-list buyers in Tinyclues and created a cream of the crop custom audience in Facebook, you can proceed with confidence and create the optimal campaign. I’d like to share with you a couple of ways our clients have successfully used Tinyclues and Facebook Custom Audiences to enhance their CRM and Facebook display strategies.

  • Classic Facebook display campaign

Many clients use Facebook not just as an acquisition media but to drive repeat-purchases, using their  first-party data (rich data on people they know). With Tinyclues, they find the future buyers for product(s)/offer(s) which Facebook custom audiences then matches with Facebook profiles. A couple of our clients showcase multiple products on rotating carousel cards to Facebook profiles that match Tinyclues’ targets.


  • Following up on an email campaign: eat, sleep, email and don’t repeat

How many times have you launched follow-up emails with poor results, or even worse, only to see them backfire on you… in the form of e-mail-weary recipients clicking the “unsubscribe” button? Facebook display ads can be a tactful alternative, allowing you to send a subtle reminder (for that well-earned vacation package?) without incurring anyone’s wrath. Even better: reach out to the specific slice of customers very likely to buy who opened their emails but didn’t click in the first place.

  • Reaching out to the un-reachable

How to re-connect with customers who have unsubscribed from emails altogether? Chances are you can catch their eye on Facebook, with targeted offers suited to their unspoken needs and wants. That’s what some of our customers are doing; they take their list of opt-out customers and home in on the 5-7% most likely to purchase the promoted product, then pitch a display ad specifically to this limited tailored audience.

 LightbulbTip: When uploading your target customer data to Facebook, maximize the number of matches found by filling out as many fields as possible. Average Facebook matching rates are 50-60%, but if you include details like phone number and zip code (in addition to email address), you can boost matches to over 80%.

A ground-breaking strategy with exciting business results

Across all the examples where our clients have used Tinyclues to enhance their Facebook Display campaigns, I’ve seen the same pattern: the intensive A/B test campaigns have all proven that Tinyclues outperforms both internal, ad-hoc targeting methods and… Facebook targeting and Facebook Lookalikes! Tinyclues’ campaign intelligence capabilities improve results across the KPI board: increased total revenue, more impressions, lower bounce rates, and better returns.

One of my clients, a travel & hospitality company, uses Tinyclues and Facebook Custom Audiences to display ads for regional travel destination campaigns. After having selected the audience from their CRM using in-house targeting methods (recent clicks, behavior and client destination preferences) and uploading it to Facebook to find profile matches, they then compared it to an audience selected with Tinyclues’ Deep AI.

In these A/B tests, Tinyclues provided a revenue per dollar spent boost of +88%. While fewer users clicked on the ads with Tinyclues, the conversion rate was 61% higher, and the cost per acquisition of the ads was less when using our AI-first solution!

In House Data

Once Tinyclues outperformed their internal data scientists’ targeting, my client decided to test our solution against Facebook’s targeting capabilities based on customers’ interests.

The results were impressive! “While the budget spent on Facebook’s target was almost 3 times greater, Tinyclues’ conversion rate was more than two times higher, and cost per sale was 29% lower. Revenue per click and per dollar spent were both higher, by 112% and 53%, respectively.”

As the CRM campaign playing field takes on new depth and dimension with Facebook Display Ad optimization, this is just one of the as yet under-exploited channels available to marketers. My team and I are excited about the growing possibilities this opens up for marketing teams, so please contact us if you’d like to discuss. Meanwhile, watch this space for more tips about developing your omnichannel marketing campaign strategy.