Note: This was a fun experiment. For science :). No client accounts were used in this test. It’s powerful, but use discretion!
Facebook is a powerful way to give your content more exposure. But it can also drain your bank account faster than you can say, “Zuckerberg”!
Promoting content at scale is a challenge for many of our clients. So recently, I decided to run an experiment here in the UpWord innovation lab that aimed to drive down the cost of Facebook advertising.
Could we bend and break the FB ad platform to our will?
The results are in.
Facebook ads work as an auction, much like Google Adwords. This means that the laws of supply and demand are at play.
With Google Adwords, some keywords will cost more per click than others.
Where Facebook targets people, some people cost more than others to reach (feels a little weird to say, but it’s true).
It all comes down to value. Certain keywords deliver more sales to advertisers, which ramps up the bidding competition. Same thing applies to different audience segments on Facebook.
The goal was to get cheap engagement on Facebook to boost the visibility of our content.
I used one of my “experimental” websites to hedge against any risk. The content I promoted was a case study about web hosting.
It was a fresh angle on a tired topic, but not groundbreaking work by any stretch of the imagination.
Step 1: Choosing the Audience
First, I needed to find a market of high supply and in low demand. By supply, I mean a large target audience. By demand, I’m referring to a market where few advertisers are competing for placement.
Why is the demand low for these countries?
I can’t say with 100% certainty, but I imagine that it has something to do with low purchase intent.
These countries have a comparatively lower English speaking population. They also, on average, have comparatively less disposable income. So it stands to reason that advertisers care less about spending money to reach these people because the expected pay off is less.
Next, I selected interests that would be a good fit for the case study I was targeting. This included interests like “SEO” and “Online marketing”, but also included followers of various related thought leaders.
This helped us work with a more discrete segment of the target population, which came in handy later for the second experiment.
Step 2: Setting up the Ad.
Knowing that the majority of my audience was skewed male, I chose a lovely (tasteful) stock photo of an attractive female.
Hey, we’re optimizing for click-throughs here!
Here’s what the ad looked like:
Simple and to the point.
Step 3: Boost the post
With my audience set up and post published, it was time to test. I used the “boost post” feature for ease of use, but could have easily run a standard news feed ad to get the same effect.
The Results: Can anyone split a penny in half?
Within hours of boosting the post, two things stood out:
1. Yes, the audience was decidedly male
2. My economics professor wasn’t lying. High supply and low demand resulted in rock bottom prices.
After spending $6.77, the post received 939 direct engagements (clicks/likes), with a total of 1,137 likes. That means that we received at least an additional 198 likes from the resulting organic reach (free).
It’s not hard to run the math on this and see that we paid $0.005 per post engagement.
That’s half of a penny per engagement.
At this point, we reached the goal. We got cheap interaction and visibility on Facebook.
But to what end?
What am I supposed to do with thousands of likes from Indonesia? It’s not driving leads, sales or any perceptible business value.
That’s where part two comes in.
Experiment #2: Jedi mind-tricking a different audience
“Social proof” is a powerful psychological concept. It’s why we advise our clients to gather testimonials from real customers. It’s why we show off logos of well known brands who use our product or service.
People can’t resist jumping on a bandwagon.
I hypothesized that the 1,137 likes could give us an advantage when it came to running a legitimate campaign targeting a high demand audience.
Step 1: Identify the high value target audience
This time I wanted to target the high demand, high competition countries, while using the same interest targets.
I want to stress that labels like “high/low demand” are overgeneralized. There are exceptions to every rule. I’m just referring to how the math plays out in this particular auction, not making a judgment call on the value of these particular markets.
Step 2: Split test the boosted posts
We served two different versions of the post and boosted them each to the “high demand” audience.
The only difference between the two ads was the 1,137 likes that we picked up from the first experiment.
After two short days of running the second experiment, we had noticeable results.
Although the total sample size isn’t that large, the trend is hard to ignore. Even with a wide margin for error, the post with “pre-loaded” likes received far more engagement than the post that started from scratch.
Specifically, the post with likes saw:
- 285% more engagements
- 64% less cost per engagement
- 300% more clicks
- 920% increase click-through rate
Again, the only variable we tested was the pre-loaded likes on the post. Everything else was held constant: the post text, headline, image, and Facebook page name.
How else can we apply this “hack”?
In no way am I advocating for anyone to inflate their “likes” and call that success. You would be doing yourself or your clients a major disservice.
The better question is, how can you leverage social proof to boost the effectiveness of your marketing?
Here’s a few ways I see it playing out, but would love to hear your ideas in the comments:
Thanks to the open graph, these social signals follow us around the web. Here’s what the top of the blog post looks like now:
The high count on social sharing buttons is a helpful trust signal for any user that lands on the page. I’d also recommend placing those social share buttons in close proximity to any call-to-action you have on the page.
Bring Scale to FB Ad Campaigns
Our experiment achieved a 64% lower cost per engagement. Even half that figure would give a significant advantage over the competition. It’s like having a 2 hour head start at the Boston marathon.
In a competitive ad auction, “priming” your ads this way can give you an edge. Ultimately, this also lowers your cost per customer acquisition and could make your advertising efforts exponentially more profitable.
SEO Benefit (?)
A study by Eric Enge back in 2013 appears to have debunked the notion that Facebook likes contribute to SEO performance. Although the study is compelling, it was not entirely conclusive.
I’m not ready to throw away social signals as a ranking factor (yet). I think it’s safe to say that they aren’t a major factor, but I’m not convinced that they don’t still have at least some influence on search visibility (a topic for another post).
That said, we know that even indirectly, social proof could help contribute to the “linkability” of the content. That is, human readers consider the page more trustworthy based on its social sharing results, which influences their decision to link to it.
I have my low-cost test audiences saved and plan to use these “primer” campaigns in the future with more ad spend to see if the results do, in fact, scale.
To build that audience – be it a low-cost test audience or your exact target – use our customizable Facebook Targeting Guide as your starting point to save time and headaches.
Is this crazy? How else could this pay off in the context of a larger marketing campaign?
Like what you read? Subscribe!
UPDATE: Some of our readers have brought up a great point and I think it makes sense to add an extra word of advice before you try this out.
When targeting low-cost countries, it’s likely that you will pick up a number of likes for your page (even though it wasn’t the primary goal). When I ran the experiment above, we picked up about 50 page likes.
If you start to accumulate a lot of low quality likes on your page, it can hurt your reach, as these un-targeted folks (bots?) cannibalize the visibility of your future posts.
This is an easy fix. Go to your Facebook page and click Settings > Banned Users, then select “People Who Like This Page”.
From here, you can remove recent likes from your page. Took me about 60 seconds. Just make sure you do it right after the “primer” campaign is finished, so that you don’t have to guess at who came in from the experiment.