With the shift from DAP to DAT, travel brands are now able to reach a highly engaged audience. For groups like IHG, this has led to a cost per booking decrease of 20 percent with a 50 percent increase in campaign scale. With its 1.8 billion monthly users, Facebook has the unique ability to reach customers in a native environment without interruption. In order to maximize campaign performance, it is important to have a fundamental understanding of the auction environment.  The DAT system has been designed to combine the needs of both advertisers and customers. This approach improves user engagement while providing much-needed cost controls.

How does the Facebook Auction Work?

Facebook used an economic theory called VCG (Vickrey-Clarke-Groves), to create a second-price auction environment. Essentially, the winning bid does not pay exactly what was bid, but rather a bid slightly higher than the second highest bidder.  The primary benefit of this model is that it mitigates the risk of sharp jumps in spend. Ultimately, it is a win for advertisers as it creates a stable auction environment.

So now that we have established the theory, we can break down the elements that determine ad delivery. Facebook uses a Total Value Equation to determine ad delivery. The equation takes into account both customer and advertiser generated signals.

Each component of the Total Value Equation plays a role in determining delivery.  The ad that has the highest total value is the one that will be delivered. This means that advertisers need to consider multiple elements when optimizing campaigns.

Advertiser Bid

There are a couple of different ways to approach bidding, but before we jump in, we need to cover campaign objectives. Simply put, a campaign objective is the goal of the campaign. In other words, the consumer action the campaign is trying to drive. There are currently eleven campaign objectives supported by Facebook.

Once a campaign objective is selected, it is possible to select a bidding method. There are a couple of different ways to approach bidding.

Option 1: Automatic Bidding

Automatic bidding allows Facebook to actively alter bids on the advertiser’s behalf.  The system focuses on maximizing the selected campaign objective. So for the average travel ads campaign, this would maximize product catalog sales. Essentially, Facebook will bid whatever it takes to drive the most bookings. The obvious drawback is that there is not a way to directly control acquisition costs. It is possible to set a weekly budget, which gives some control over total daily spend. Automatic bidding is a great option when the campaign is not held to strict acquisition cost targets. If the campaign is held to a tight efficiency metrics, then it typically makes sense to consider a manual option.

Option 2: Manual Bidding

Manual bidding allows advertisers to leverage more control over the bid and ultimately the campaign CPA. With manual bidding, there are two options to choose from:

1. Manual-Maximum bidding  – Allows advertisers to set a maximum bid, creating a cap that Facebook’s system will not exceed.
2. Manual-Average bidding –  Allows advertisers to set a target average acquisition cost. Facebook will actively manipulate bids to work towards that target.

The potential drawback for both of these bid types is that they can potentially limit conversions. If Facebook determines it can’t achieve the campaign objective at the specified limit, the ad delivery will slow down or could stop completely.

Estimated Action Rates

Facebook prioritizes delivery based on the probability that individual users will complete the campaign objective. For example, let’s say a hotel group creates an ad targeted at users with travel intent and the goal of the campaign is to drive bookings. Instead of delivering the ad to every member of the target audience equally, Facebook will prioritize delivery based on who is the most likely to actually book through facebook. With this method, ads are shown to the most relevant audience.

There are multiple factors that go into determining the Estimated Action Rate. One major driver is recent ad activity which takes into consideration engagement signals such as likes, comments, shares, and clicks. Another element, user attributes, which takes into account basic information and past activity with similar ads.  Facebook also lists page, app, and ad account history as important data points. All of these indicators are used to create an estimated action rate.

Relevance and Quality

After an ad has been served more than 500 times, it will receive a relevancy score from 1-10.  Relevance is determined based on the predicted positive or negative interactions.  Positive engagement activities (likes, comments, shares) improve ad relevancy. Negative engagement indicators (Users hiding ads or requesting to no longer be served ads) decrease relevancy score.

Cost is largely impacted by relevance score. The lower the relevance score, the higher the bid needed in order to compete. This means that a highly relevant campaign can be a cost saver. Continual testing of ad content can improve relevance score. In turn, this positively impacts performance as it results in both an increase in conversion and a decrease in cost. For more information on relevance, be sure to check out Facebook’s in-depth page on the subject.

In Closing

Understanding how Facebook’s auction works is integral to designing and optimizing a strong DAT campaign. This information will get you started, but experience in the auction environment is important.