When you’re setting up Facebook display ads, you have a choice about how you want to pay. You can pay based on impressions (CPM), optimized impressions (oCPM), and clicks (CPC).
If it’s engagement — post shares, comments, and likes — that you’re after, you should obviously pay on a CPC basis, right? Wrong.
The truth is, you need to experiment. It’s common for advertisers to think of CPCs as the most efficient advertising solution; after all, you’re paying for specified user actions, as opposed to eyeballs (the premise behind traditional CPM pricing models). But CPC isn’t always the answer to your biggest brand-building need: maximum engagement at the lowest possible cost.
AdParlor traffics ads for many large advertisers who are looking to drive efficient audience engagement at the lowest costs. They’re looking for cheaper CPCs, unique clicks, and shares. We frequently find that oCPM pricing models are the best way to achieve these goals.
The oCPM Algorithm: A Crash Course
Facebook’s oCPM algorithm is highly proprietary, so we don’t know what’s happening behind the scenes. What we do know, however, is that Facebook has a wealth of anonymous psychographic information at the user level. Facebook is able to determine the likelihood of a user taking action on a specific ad.
Pretty cool, huh?
That’s where the oCPM algorithm comes in — it actually uses this wealth of user data to calculate which audiences are likely to engage with an ad.
So what should advertisers optimize towards? The oCPM algorithm is designed to provide advertisers with lower cost, more efficient click actions. Brand marketers can use this strategy to generate post engagement actions at a lower cost. Here are some example user actions that marketers will optimize towards:
- Follow Question
- Claim Offer
- Answer Question
- Click Link
- Like Page
- View Photo
- Play Video
With oCPM, you’re paying per impression, but you’re giving Facebook the freedom to find more users within the price point that you’ve specified. You may spend more for some engagement metrics and less for others. On the aggregate, however, your costs level out.
When you pay a CPC, you spend exactly what you specify per click. If you pay for $4 a click, you won’t spend more than $4 for that one click. When you pay based on an oCPM, you aren’t exactly paying for a click, but the amount you’re paying can be backed out to a click. The cost you pay per click will be variable.
Which Is More Efficient?
We put this question to the test across a sample of four posts. In total, the sample generated 119K oCPM impressions, 119K CPC impressions, 4K oCPM clicks, and 3.5K CPC clicks. The four ads were photo posts, and the advertiser was a consumer products delivery service.
On aggregate, according to data from AdParlor clients, we found that our oCPM bids generated higher CTRs at a lower effective cost.
|oCPM -> Post Engagement bidding||118,880||4,005||3.37%||$0.21|
We also looked at a breakdown across four specific posts — each generating more than 20,000 impressions and 350 engagement actions.
Here is a breakdown of post engagement actions for each of the four posts. Note that the last columns in the table include all post engagement actions, not just the sums of the ones we’ve listed.
|Actions||% change in cost per action when oCPM bidding||% change in actions performed when oCPM bidding|
|Post Engagement (all FB engagement actions)||-3.81%||3.96%|
Here is what stood out to us:
- Virality: This metric represents the reach of an advertiser’s content. For three out of the four posts, share rates were higher when bidding on an oCPM for post engagement.
- Post Likes: This metric represents the popularity of an advertiser’s message. For oCPM ads, “like” counts were unanimously higher than the CPC units.
- Comments: This metric represents a brand’s ability to spark conversations. It’s interesting to note that, on aggregate, oCPM ads generated a lower comment rate. It is possible that commenting is one of the more involved forms of engagement, requiring thought and time and that the oCPM algorithms may lean more towards “faster” actions such as likes and shares when optimizing for post engagement. A great follow-up test would involve running oCPM, optimizing for comments, and comparing the results with the performance of CPC bids.
Choosing a bid type is far from a black and white process. If you’re not sure where to start, run a small test. If you’re totally sure where to start, you should still run a small test. Look at the details as well as the picture to determine the optimal bidding model for your brand and engagement goal.