How FlipGive Used Facebook To Drive 10x ROAS

Daniel Phillips

The Problem: is looking to grow its user base and drive sustainable revenue. Other acquisition channels such as Google, through partner, affiliate, proved costly and in-efficient.

Partner:, FlipGive is an online shopping platform that allows groups to shop online and earn cash back

Scope: 100K budget, 8 months

Summary: How we used Facebook to drive targeted leads and drive revenue

Results: 10.4x ROAS in revenue driven from Facebook leads, 20% year over year organic growth

Creative Expressions

The art of connecting with an audience is the most important step in learning how to grow your business in Facebook. Who are you targeting, and what messages are you delivering? If you’re not targeting the right people, or delivering those people the right message, it’s likely that your business will miss the mark. We began our optimization journey by connecting with our audience. 

Creative connection is driven in simultaneity by creative (imagery, text) itself and audience target sets.  Our process began by experimenting and ‘properly’ tracking our results to understand what message and imagery resonates best – and with whom. This ‘resonance’ was reflected by a decrease in click/engagement cost.

After testing (we’ll elaborate on our testing process below) we evolved from a team based messaging set that targeted members of youth sports teams that needed to fundraise, to a brand based messaging targeted to consumers that purchase specific brands in the FlipGive offering. We evaluated success by reporting on time-bounded acquisition and activation results with our intended target. You can see, that by our third iteration of our ads, we had started to begin to producing meaningful engagements with our target audience (reflected by actual engagement statistics seen below).

Increasing Early Wins With Lookalike Targeting

Once we had found a way to connect with our optimal and activating audience, our next step was to scale. To do this we leveraged the power of look alike audiences. Look alike targeting can help businesses reach millions of new customers similar to the existing client base. We used look-alike to build lists of over 2 million unique profiles in Facebook that match similar web conversion signals to those of optimal performers in the FlipGive ecosystem.

Specifically, we took a look at the users who were satisfactorily completing registrations, and purchasing from specific brands and target sets. For example, a lookalike audience of 10,000 baseball users that purchased from a key apparel brand in our offering – and Facebook’s ability to find us 2.1M similar consumers.

Testing Part 1: ICE Scoring

So now we have our audience, we’ve scaled up to a larger size. How can we effect end revenue for the business? Here’s a unique method in which we prioritized ways to drive revenue from the leads in question, ICE Scoring.

ICE stands for Impact, Confidence, and Ease. It’s a scoring system used to ‘grade’ business ideas, and qualify impact for immediate implementation and testing.

ICE scoring was pivotal in allowing FlipGive to create a rapid testing environment to execute Facebook specific acquisition tests with minimal friction.

ICE is used to evaluate the worth of implementing business experiments. ICE scoring can encompass tests as easy as changing a line of copy in your ads, to as difficult as overhauling entire business infrastructures. Corresponding scores that go along with ideas you would like to test, help organizations select and classify the ‘next up’ ideas for implementation.

TLDR; If we couldn’t pull it off in 7 days or less, it wasn’t a project we deemed worthy of our immediate investment – ICE allowed us to determine our time frame and impact on such experiments. 

ICE Scoring example:

Project 1: Ad Copy Change + Landing Page Experiment

We’ll give project 1 an impact score of 7/10. Changing the ad copy and reflecting landing page likely won’t change our business overnight but it will give us a decrease in lead cost if we connect. Confidence 10/10, we’re very confident in our hypothesis that copy changes can effect lead cost. Ease 10/10, to change ad copy and landing page copy is a simple manual task, 10/10 reflects just how easy this is. Project 1’s end score is made up of an average of the three scores. Project 1 has a score of 9

Project 2: New Referral Program + Accompanying Facebook Outreach 

For impact, we’ll give this a 10 as a referral program successfully carried out can have a significant effect on a businesses bottom line and Facebook has proven to be a driver of such programs. Confidence is another 10, this is a highly successful tactic amongst all business demographics. Ease 1, because building out a new referral program requires the shift in resources of our development team – and while the impact may seem significant, the attention drawn from other areas of our business will be an overall detriment to bottom line. Project 1 has a score of 7.

The outcome of this prioritization logic, is that we work on Project 1 as it has a greater score. This is not to say that project 2 is not valuable, or not something we should work on. But it is to say that we can get a greater immediate impact to the immediate activation of our users, if we focus on learning from project 1 first.

Do not be discouraged from larger ideas – the ICE scoring methodology is a way of validating minute optimizations rather than overall business trajectory. There are also unique methods known as ‘smoke tests’ to validate larger ideas, and support the argument for implementation of more difficult projects. 

Testing Part 2: AARRR Framework

This is a unique framework in which we justified the levers of our experimentation. We’ve scored the ideas, it’s now important to understand how these play into the overall roadmap of the businesses goals.

The framework stands for Acquisition, Activation, Revenue, Retention, Referral (or, the pirate funnel). An important pillar of optimization that allows businesses to focus on key actions that drive users through each funnel of your business.

At specific times in your business’ life cycle, the organization will be prone to lean heavier on a specific pillar of the framework. For instance, a growing business may lean extremely heavily on the acquisition pillar, and eventually work it’s way down the funnel. Or, a well established business may focus exclusively on the retention and referral funnels in order to optimize results from it’s existing client base.

At FlipGive, we focused on driving experiments that effected all points of the framework from Facebook driven acquisition campaigns, to Facebook driven referral initiatives. It was only by doing this, that we could unilaterally effect the activation rates and costs of our users. Our selections from ICE framework projects reflected this, as at certain times in the lifecycle, one may choose to disregard a higher scoring project for the purpose of attending to a much needed area of business concern. I.e. Acquisition is not much of a concern when your “R” revenue stream is broken.


Over our testing period we produced a 10.4x in revenue on the initial $100K facebook investment, and counting. We’ve driven initial user activation costs down over from$188.06 to $38.82/activation. We aim to deepen our understanding of audience performance with ever more testing.

As the optimization journey truly never ends, we’re looking forward to providing improved results over time to the FlipGive organization.

Daniel Phillips