Facebook Ads: Hip? Hype? Or Huh?

In our column last August, we wrote about “Facebook, Fundraising and the Dawn of a New Age.” We asserted that with the changes to Facebook advertising, Facebook was poised to play a much more important role in integrated and even stand-alone campaigns. So, now that we’ve been through an end-of-year campaign and a major disaster (Typhoon Haiyan in the Philippines), what do the results look like? Short answer: small but promising.

First, let’s talk about the scale of recent tests
According to Adam Gerber of online fundraising and communications agency M+R Strategic Services, six of M+R’s clients (including CARE, PETA and U.S. Fund for UNICEF) conducted year-end tests with Facebook ad campaigns. Across these clients, the efforts raised $25,000 from Facebook Custom Audiences and Lookalike models. (CARE and another international relief client raised more money a few weeks earlier in response to Typhoon Haiyan.) What are these? Facebook Custom Audiences allow you to target supporters in your database with ads on Facebook; Lookalike models provide the ability to target people who are similar to your Custom Audience list(s).

Twenty-five thousand dollars? Total for six nonprofits? Isn’t that still peanuts — especially given the large size of these nonprofits? Perhaps, but these campaigns were in a test phase, and M+R believes there are significantly larger opportunities in the future — especially as we get a more comprehensive view of their impact. This data doesn’t take into consideration the impact these (and other) Facebook ads had on increasing response rates to emails, mail pieces and other donation vehicles as part of integrated campaigns.

M+R found, for example, that the ads to Facebook Custom Audiences lifted the response rates to coordinated email campaigns. We have found that display ads are typically more of a reinforcement channel than a conversion channel. Assuming the same is true for Facebook ads, the impact may well have been significantly larger. In addition, the $25,000 doesn’t include Facebook retargeting ads deployed by CARE (i.e., ads on Facebook targeting people who visited their donation pages and other key website pages but didn’t donate — check out our recent article on retargeting for more on this emerging strategy here).

And how about the return on investment?
CARE saw a three times immediate return on its investment from its Custom Audience and Lookalike model ads in the two weeks following Typhoon Haiyan. M+R indicates that some variations of the emergency campaign around Typhoon Haiyan had more than 1,000 percent immediate ROI in the first 24 hours of the campaign. In addition, many of the donations came from new or reactivated donors so the long-term ROI will likely be much greater.

The ROI for the end-of-year ads was a bit more spotty during this initial test year but still promising. Two of M+R’s six clients achieved an immediately positive ROI, while the other four came close, according to Gerber. This suggests that these campaigns may have positive ROI over time — and might have a positive ROI sooner with a little refinement.

Top takeaways and recommendations
1. Facebook ad type: Desktop news feed “promoted posts” typically perform best for donations and other asks targeting conversions. Mobile news feed ads and Facebook “right rail” ads are secondary but perform well in selected situations.

2. Targeting: Getting your targeting right is key — and probably requires testing. CARE used Custom Audiences to target recent emergency donors during the Typhoon Haiyan campaign, and they performed extremely well. Overall, Custom Audiences — targeting existing supporters on your database — performed better than Lookalike model audiences, although some Lookalike segments performed well, too.

3. Events: Haiyan disaster ads had a significantly higher ROI than the end-of-year ads, albeit to different audiences. The obvious conclusion is that disasters work well through Facebook as well as just about everywhere else. Could this reflect that Facebook donors are more event-triggered than institutional? We are guessing this is the case — especially based on the episodic manner in which most people use Facebook (or maybe that’s just us). It’s worth testing to see if other types of events or programmatic content can prove as successful.

4. Timing: For the end-of-year campaign, the last 24 hours had much stronger performance than earlier in the month. Facebook donors may respond best at the last minute.

5. Messaging: CARE did an interesting test of a more emotional ad (first image) vs. a more institutional impact-oriented ad (second image). Contrary to expectations, the impact-oriented ad did far better with an immediate return more than twice that of the emotional ad.

6. Test and refine: As with everything in direct marketing, some ad units performed much better than others. There’s an opportunity to grow future volumes and returns by testing and building on the learnings to date.

7. Who should try this: The short answer at this point is, “We don’t know.” We’re still in the very early stages. However, the national nonprofits with global footprints, big brands and strong social-media presences seem best-positioned to make use of these new tools. Other charities may benefit as well but may not be as good a near-term fit. Time and testing will tell.

Want to start a Facebook Custom Audience or Lookalike model ad campaign? Try checking with online-marketing agencies, direct-marketing agencies or list brokers — or do it in-house via Facebook.

Thoughts? Drop us a note. We would love to hear about your results from your recent Facebook advertising testing.

(The authors want to acknowledge the time and insight shared by some industry experts including Adam Gerber of M+R and Kiera Stein of CARE. Many thanks to you!)

Karin Kirchoff is vice president at MINDset Direct. Reach her at kkirchoff@mindsetdirect.com. Jeff Regen is vice president of development at WETA TV 26 and Classical WETA 90.9 FM. reach him at jregen@weta.org