Facebook Audience Testing Case Study: Elijah McCoy Campaign

The goal of this project was to see if we could effectively market our content to an audience outside our core demographic in a data-driven manner. We picked this video as the test subject. The success of this experiment validated our strategy for the YEAR project.

1. Identify goals

Success metrics for this campaign included:

  • Engage non-core audience
  • Identify which of the target personas our list resonates with the most.

2. Brainstorm audience personas

We reviewed the video and discussed what kinds of audiences would engage with it.
We identified four audience candidates:

  1. FEE Donors
  2. Black Entrepreneurship Fans
  3. Inventors and Makers
  4. Mike Rowe Fans who like Entrepreneurship

We tested four additional placements to display the content:

  1. YouTube link from the FEE page
  2. FEE.org article on the FEE page
  3. Facebook add (content specified within ad)
  4. Facebook Video Upload

We excluded Donors from the additional placement test, this left us with creating 12 ad sets (4 placements x 3 audiences, each containing 1 ad).

3. Create saved audiences

Each audience was iterated, so the below represents the final combination of criteria. We identified a number of people, hobbies, and shows which our personas might have in their Facebook profiles, but only the below were available as targeting options:

4. Measure test results:

In terms of placement, we saw that ads with the YouTube link, FEE.org article, and in-ad content were not responding at all. The per-engagement maximum was set at 15 cents, and there were only 1-4 engagements out of 40+ reach. Cost per engagement was 3-14 cents. Once we changed format to Facebook-hosted video, reach and results immediately took off, at 1 cent per engagement.

Audience (Persona)
Out of the three audience we tested, the “Black Entrepreneurship” audience responded the strongest. Here are the video engagement stats for the first day:

Across all metrics, “Black Entrepreneurship” outperformed other audiences. Donors also performed well, as expected.
5. Boost winning ad set

Based on the first day, we decided to boost “Black Entrepreneurship” to $100 per day and “FEE Donors” to $30 per day. All other ad sets (and audiences) were stopped.

6: Measure results

Delivery Summary
  • $440 was spent boosting this video, with $38 used for the first day of calibration, and the rest directly on the selected audiences.  
  • A total of 322,000 people were reached. 178,000 of those were from direct paid reach, while the rest were organic. However, once advertising stopped, video views dropped for 30,000+ per day to well under 1,000. The majority of organic views originated from people who saw the ad.
  • After the ad campaign ended, organic views dropped rapidly from over 10,000 to a few hundred (see below). This video is not effective with our existing audience, and did not have a lot of organic momentum.
Return on Investment
  • There were a total of 111,000 video views, and 8,000 reactions. Each dollar thus generated 252 views and 18 likes.  
  • Cost per 1000 impressions was $1.50 for the “Black Entrepreneurship” audience and $8.03 for the donor audience.
  • The FEE Store link was clicked 301 times.  There were 28 Real Heroes sold during the ad period vs 26 for the several months prior, so the ad likely drove virtually all of those sales. Facebook matched 7 purchases via Offline Action tracking (each purchase may involve more than 1 book.) 

Open Questions
  • How many of the users who engaged with the ad will become frequent readers?  We’re not sure how to track this. (Google Analytics? Facebook Insights?)
  • How to measure the value of paid ad exposure? Book sales? Shares? Organic uplift? Are paid views as such worthwhile?
  • Given that our existing audience was not the most responsive to this video, and given that we were attempting to target groups outside of our core user base, is it possible that we need to reach a larger number of people with paid views in order to reach some unknown saturation point that will spark more organic growth? In other words, do we need to spend a lot more money before network effects will take over?
  • Once we’ve effectively identified an audience target through testing, would different means to reach those audiences be more effective than paid social media ads? For example, should we have planned an advertising strategy that included direct outreach to major audience influencers around Black Entrepreneurship, such as Daymond John?

Lessons Learned
  • Targeting segments effectively requires testing different audiences.
  • We can effectively boost content to new demographics and even get them to buy products, but converting them to returning visitors will require a new strategy.
  • We need a clear call to action for ads to measure effectiveness and extract value.

Five principles of effective marketing at FEE

Let’s talk about five principles that help the marketing team at the Foundation for Economic Education (FEE) communicate more effectively with our customers.

1. Clearly state the value proposition

Every message should clearly communicate the value our customer will gain from the product we are selling.

Let’s break that down:

Although FEE is a non-profit organization, we consider everything we do to be a product which can be sold to a customer just as much as multinational behemoths like Apple and McDonald’s

Every marketing conversation begins with a discussion of our audience:

  • Who are they?  What do they care about?
  • What do we have that they will value?

For example, when we segmented the market for FEEcon (our big conference this summer), we:

  • Split our customers into three groups: college students, young professionals, and older supporters of FEE.  
  • Brainstormed the values that each group gets from attending FEEcon.  From these values, we identified 4-5 value propositions for each group.

Here are the first two value propositions for FEE donors:

  • Personally witness how your investment in educating the next generation transforms their lives
  • Interact with the leaders of dozens of partner organizations and major philanthropists across the freedom movement

And for students:

  • Real skills for professional success and solid theoretical education
  • Networking opportunities: Engage with successful entrepreneurs and student leaders

These value propositions inform all the messaging that we create for this product, whether the medium is an article, email, or Facebook ad.

Consider this ad for students:

And this ad for young professionals:

And here is an ad for donors:

2. Present a single, clear call to action

The motto of FEE’s marketing process is “Always Be Closing.” Everything we do as marketers is designed to move our customer further toward closing the deal. Every message we make, be it a landing page, email, SMS, facebook ad, flier, whatever, asks the customer to take an action that moves them further down the conversion funnel. Each marketing communication has to have a large, prominent request to take a single action which will give the customer some sort of value.

This doesn’t mean that every marketing message has to ask for a deep commitment. We have a conversion funnel for each customer person which consists of a series of small messages. If we want students to apply for a three-day seminar, we first ask for their email so we can send them a free book or guide. The goal of each communication is to deepen awareness of our products and lower the barriers they have to the next step.

3. Make messaging personal 

The essence of our communications strategy is to make every message we send feel like it was written just for you by a real human being who cares about your concerns and is eagerly awaiting a reply, then use marketing automation to scale up that personal feel to thousands of people.

There are a few ways we do this:
A. Every email comes from a real human being. We don’t use any noreply@, support@ or sales@ emails. This includes transactional emails such as payments, registrations, reminders, etc.
B. Emails use a first person informal tone. We never say “we.” Messages take the tone of “I would really appreciate if you could do x.”
C. We sign all emails with our name just as we do with our personal mail.
D. We use plain text format whenever possible, especially if we expect a reply. When we use CRM tools to target messages, we export the names into Gmail or send a plain text message whenever we can. Marketing messages sent in plain text from Gmail avoid both the junk mail and the “Promotions” folder.
E. We make messages short and to the point. Because we can speak directly to the recipients values, we can offer something that we know they’ll care about and don’t have to waste space addressing everyone.
F. We talk like a normal conversion. We experiment with short, informal, lower-case subject lines such as “quick question:” or “you’re missing out.”

Marketing email from our CRM tool sent via Gmail

4:  Focus messages on specific customer personas

It’s easy to say that every message should push the customer down the sales funnel. The hard part is to track who each customer is, and where they are in their journey.

At FEE, we use HubSpot to split our users into a lifecycle stage funnel (lead, subscriber, opportunity, customer) and a customer persona (college student, parent, interested donor, casual reader, etc).  HubSpot tracks every visitor’s web and email interactions, and dozens of workflows use specific triggers, e.g., visiting the donate page recently, to classify people into personas based on recent behavior.

This allows us to tailor messages to the specific customer profile and offer them a product that we think they are most likely to be interested in. This minimizes our unsubscribe rates and keeps followers interested in our content.

Additionally, we make heavy use of retargeting for our advertising. We show Facebook and Twitter ads based on specific pages people visit.  If you’ve visited FEEcon.org recently, you’ll start seeing more FEEcon ads in your Facebook feed until you register, which will switch you from our “promotion” (please register) to the “nurture” campaign (please share this with your friends). If you’re not on our daily email list, you’ll see a lead form in your feed, otherwise, you might see donor messaging if you’re a heavy user of the site.

Partial snapshot of personas and lifecycle

5: Experiment to identify the best strategy, then automate it

We don’t plan a grand marketing strategy for each product.  The fact is, we have no idea what kind of message will resonate with our audience. Our marketing strategy is basically this:

A. Define audience
B. Define value proposition
C. Experiment with campaigns based on A and B at a small scale until we find something that works
D. Scale up C.

We build workflows which capture the most effective strategies and automate them for each customer journey. We use HubSpot to build one or more workflow for each product which contain a series of calls to action, triggers, messages and rules. First we capture leads with a CTA on a website or ad, then we enroll customers in a workflow and nurture them until they convert (buy the product, register for an event, donate to us). This enrolls them in a new workflow which is designed to deepen their commitment and cross-sell other products, starting the process over again.