How much is a Facebook ‘like’ worth? + other social media ad experiments 9.26.11 0
by Sean Blanda in Experiments

Scaling an online properly is challenging and sometimes, you need a little help.

Lately we’ve been experimenting with ways to attract readers to our sites beyond just word of mouth and content partnerships. For the first time ever, we purchased advertisements.

However, our budget is rather limited so we decided to test the advertising platforms of what are arguably the three most important social networks on the web: Twitter, Facebook and LinkedIn.

Below is the result of a tests conducted over the past 45 days for our technology news site Technically Philly. This test was not scientific, but we hope our experience can help educate your future advertising buys.

1. LinkedIn

Target: Those with “Information Technology” or “Entrepreneur” in job titles.
Geography: Greater Philadelphia Area

Our LinkedIn ad was a simple square that pointed directly to our homepage (see above). Above is the exact report from LinkedIn (click to enlarge). Once you’re done laughing at our .031 percent Click Through Rate, you’ll see that the ad buy was a complete disaster.

Not only did the ad not convert well, there is no LinkedIn “fan page” where we can keep people consuming our content. A $3.00 CTR is way too high for Technically Philly,  a site where we are not selling a direct product.

In other words, if we were selling widgets for $100 we could perhaps justify this ad buy as the occasional $3.00 click would result in $100 in revenue. But Technically Philly needs repeated engagement for it remain profitable.

2. Twitter

We started to explore Twitter’s promoted tweets product, but it turned out that you needed $5,000 minimum. No thanks.

3. Facebook

In its first two years, Technically Philly focused much of its efforts on cultivating a community on Twitter. We’re a tech news site and most of our readers spend their days on Twitter. However after checking our Facebook page and discovering a laughable number of likes earlier in the year, our egos required us to do something.

We implemented Facebook comments and experimented with posting different types of content to our Facebook page in the early part of the summer. After seeing a respectful growth in likes, we decided to explore an advertising buy.

There are many variables to a Facebook ad. Firstly, there are two types of ads:

1. “Sponsored Story” notifies you when your friend interacts with Technically Philly “your friend has liked Technically Philly”

2. Straight ad. You can link to your Facebook page or your website. We chose our Facebook page, as our primary goal here was to increase our “likes.”

You can also have an infinite number of target demographics which is a nightmare for testing. We tested in two rounds.

The first round:

Ad # 1- Sponsored Story

• who live in the United States
• age 18 and older
• who are not already connected to Technically Philly
• whose friends are already connected to Technically Philly

Ad #2 – Straight advertisement

• who live in the United States
• who live in Delaware, New Jersey, New York, Pennsylvania or California
• age 18 and older
• who are not already connected to Technically Philly

If you thought the LinkedIn ad buy performed badly, the Facebook straight ad put it to shame with its .009 percent CTR. The sponsored story resulted in more connections, though I suspect that the 42 number is deceiving, as our likes rose by much more than that during the run of the ad.

After seeing the results of the two ads, we placed a third buy focusing on the success of the sponsored story ad buy with a bit more geographical focus:

Ad #3 – Sponsored Story

• who live in the United State
• who live within 25 miles of Philadelphia, Pa
• between the ages of 18 and 30 inclusive
• whose friends are already connected to Technically Philly

As you can see, the ad was seen an average of 18.7 times (!) by each targeted user. However connections doubled with the price of our buy suggesting that each like cost us roughly $1.86.

So, did the ads work?

You’ll see our monthly active users spike when we placed ads on Aug 8th and Sept 11th.

The amazing part is that the interaction held up even after the ad expired which suggests that the effect is cumulative.

Above is a Google Analytics graph that measures weekly Facebook referrals. Google anaytics traffic shows small but sustained spikes when we placed the Faacebook ads, and those spikes do not coincide with general traffic, suggesting that our ads had an effect.

So what to make of all this?

Our takeaways:

• Facebook ads promoting our Facebook page merit additional testing as early data suggest that Facebook advertisements have a cumulative effect that outlast the advertising buy.
• Since the ad buys, our Facebook page has received 289 additional likes.
• Highly-targeted sponsored stories worked best for us on Facebook.
• Twitter is expensive. Though if you have the budget, you should try it. Or, you could always do it grassroots-style.
• LinkedIn’s returns were so bad we stopped the advertising run early. Perhaps our ad could use some improvement, but we likely won’t be using LinkedIn for ad buys.