Facebook oCPM bidding works [or does it]


The way I see it, the practise of digital analytics has one and only purpose:

Optimization.

As digital analysts, we are constantly trying to optimize our digital activities, with 3 key goals:

  • Get more customers.
  • As cheap as possible.
  • Get more out of each customer.

Looking at our potential advocate’s journey with our product, we can break it down to certain parts:

  • Acquisition. How users learn about us and visit our digital platform.
  • Conversion. How do they behave & convert once they “land”.
  • Retention. How do they keep using our product once they’ve converted.

So basically, we’re trying to do everything we can to optimize our digital activities and touchpoints or channels (SEO, Social, PPC, Remarketing, Email etc), to improve and beat the competition on our 3 key goals. Simple as that.

Moving on to the subject of this post.

In my opinion, the main promise a bidding tactic is looking to deliver is getting us to “acquire” more users, for less.

Facebook ads oCPM bidding looks to do just that, by asking you what you want to achieve, how much you’re willing to pay for it, and then:

  • Serving your campaign ads to audience most likely to perform the desired action.
  • Automatically optimizing your bidding accordingly.

As opposed to other channels which are (mainly) focused on a certain part of the potential advocate’s journey, like SEO for acquisition, or email for retention, social advertising can serve different parts, based on the campaign.

To put oCPM bidding to the test, we tried it out for our mobile install campaign ads(Acquisition focused), which were using mainly CPA & CPC bids.

Our initial plan was to split test & watch differences, but the improvement was so clear that we just shifted the entire campaign to oCPM ads in the course of 3 weeks, and have been using them since with similar results.

Here is a snapshot of the results, broken down by week and bidding strategy:

(I’ve added a small factor to the absolute metrics so not to include business data, but that of course doesn’t affect the insights we get)

As you can see, there’s a very big difference in the CPI metric (Cost per Mobile App install). We used to pay 2,87€/install, were now with the new oCPM ads we’re paying around 1,2€/install.

That’s a 208% improvement, which is pretty major.

To put this in context, we used to spend 1400€ for 500 installs, were now we’re paying 700€ for 613 installs. More users, cheaper


Note: Although it’s looking like a big improvement in the first 2 components of a digital analysts’ goals, meaning getting more users for less, what about the third one?

Getting as much as possible out of each user ties in with Customer Lifetime Value and Lead Quality.

How can we be sure that facebook serves the ads to the most qualified leads for our product, and not just to people who are most likely to perform the action- installing an app at this example?

In a perfect world, we would set up our data to show us the LTV metric, broken down by the different bidding methods. Then we would be able to answer whether those people were valuable enough for our service, and if the improvement in the ROI is true.

The mobile analytics ecosystems though is pretty far from being able to give us that kind of insight just yet though, so we’ll continue with this short scope that extends as far as the install, or go as far as try to notice a change in our mobile app engagement data (e.g. an increase in the app removals).

Contact: jspanom@gmail.com