A month ago I offered my initial thoughts on the University of Calgary's experiment with placing social ads within Facebook. Here are the results of our second round of advertising, where we switched the content of the ads to see if students seemed to be responding more to the type of ad or the content of the ad.
As a quick refresher, we were trying to generate interest in our subscriptions to ebrary and Refworks. Check the first post for all the background on the two types of ads and the bidding process.
Based on my experience with the first round, I set my initial bids at $1.00 per CPM and CPC and left them there the entire run (Dec 1-12). We ended up spending a little more money this time, $40.61 vs $30.66 last time.
Once again, we paid about 3 times more for the ads on the CPM (cost per thousand impressions) ad type than we did on the CPC (cost per click) ad type. This time we got more clicks on the CPC ad than we did on the CPM ad, which is opposite what happened the first time around. This suggests to me that students may have been responding more to the message of the ad, which is a good thing. They seemed to like the ebrary message more than the Refworks message.
Average cost per click was significantly higher on the CPM campaign - it cost us $3.51 for each click, vs $0.60 for each click we got while displaying the CPC ad. In both our campaigns, the CPC ad type was more cost effective. However, we got in front of more eyeballs with the CPM ad type - 80,000 vs 55,000. Interesting to note that these numbers are almost identical to the numbers we got when running the first ad campaign, even though with that one I had started out much lower on the bid amounts.
I don't seem to be able to access the fancy graphs now that the campaign has been over for a few weeks, which is too bad. I see no way to back up the timeline for that type of report. I made my own in Excel, but they're not quite as pretty. This is the number of clicks received on the Refworks CPM campaign:
And here's the number of clicks received on the CPC campaign:
You may recall the reason for that big spike on Dec 5th is that I actually forgot to set the initial bid amount to $1.00 until that day, so we received a big jump in the number of times the ad displayed (we'd gotten 1,600 displays on the first 4 days of the campaign, when I inadvertently left the CPC bid at $0.10 and then 20,000 displays on the 5th, when I set it at $1.00, where I'd originally intended.)
You can see that the clicks do tend to tail off over time when the bid amount remains constant. It seems that in order to ensure your ad continues to display often over the entire course of your campaign (and thus generate the clickthroughs) you may need to up your bid every 4 days or so. There's a definite correlation between how often the ad is displayed and how many clicks were generated, as one might generally expect.
Here are all the numbers in one place: (click through for bigger size)
So would we do it again? Sure, why not? Grand total spending on about a month's worth of advertising for our roughly 10,000 students on Facebook was just over $71. For that money we displayed the two ads over 270,000 times, and generated a total of 55 clickthroughs. Certainly not a huge success ratio, and more than $1.00 per click, but we got our message in front of the online students we wanted to reach. I haven't yet compared stats on the two products to see if general usage went up during the campaign periods.
What do you think, was it worth the money?