What is automated bidding?

In this blog, we will explore what automated bidding is and how it can improve your online strategy.

Should you use automated bidding instead of manual, optimize for conversions instead of reach, or maybe try traffic if your audience is small? If you have been working on Facebook Business Manager for a while these questions have probably popped into your head more than once. With the increase of algorithm-based bidding options in the different engines this is now relevant for every performance marketing expert, but what does automated bidding means and how does it work?

The basics

Automated bidding is a bid strategy ideated to maximize results within a given budget. By using it, you are letting the engine set the bidding for you based on the likelihood of the ad generating a specific action.

The bid will change depending on each user and how similar is him to other audiences that have already done that action, but also on his historical behaviour.

What does this mean?

In practice it means that even before a campaign starts delivering, the platform is refining the audience according to the objective you have set up. It also means that the traditional way in which marketers set up different campaign for different objectives along the funnel may not be the correct strategy.

As Google enters into automated bidding, the traditional barrier between audience-based targeting (display advertising) and interest-based targeting (search engine advertising) gets more and more blurred. Due to this, I believe it may be useful to enter on how that audience targeting works.

The audience-based bidding difference

Obviously there are plenty of differences when we are creating audiences for a display campaign, however I will not focus on them in this article. My main focus is to point out the difference in terms of competition.

Whenever you bid for a keyword, you will be bidding against competitors that have that keyword in common. That means that when you bid for “guitar strings” or for “business insurance” you will be competing with other companies that are trying to sell the same product as you. Potentially, they will have similar enough margins and similar targets in terms of CPM (cost-per-thousand).

However, when you bid on an audience, every user enters different categories:

You are competing for impressions with companies from different verticals and that will value them totally different. In the example above, a guitar shop will be interested in the three users, however, they are unlikely to show their ad to Emily as the others, as they will be competing with companies that are willing to pay much more to serve an impression to that user.

Trying to simplify this concept, we can say that every user on a display network has a minimum CPM. That CPM depends on the information the display network holds, and on how valuable is the user for the highest bidders out there. 

Whenever we opt-in for an algorithm-based bid optimisation, the platform will try to deliver to your cheaper audiences first, increasing the bid as we increase the budget, and avoiding the expensive audiences if it can spend your full budget on the cheap ones.

Behaviours and objectives

So far, we have spoken about the intrinsic values of users based on demographic audiences. The platforms have much more information than that, and when behaviours and marketing strategies enter into play, we get a more complex picture.

Your audience is composed of several individuals with different historical online behaviours. For example:

  • People that tend to click on ads
  • People who avoid clicking on ads
  • People who buy online
  • People who would rather buy in-store or over the phone

When we try to visualize this, the picture may not be so far from:

If we try to overlap them and rename those behaviours, we arrive at the following:

They don’t differ so much from what we would consider objectives for a campaign. And they are (admittedly in purpose) some objectives setups we can find on Facebook.

Let’s then consider the competition in all these objectives. Would you agree with me if I say that I expect more competition (and willing to expend more) on campaigns which objective is conversions, then traffic, and finally awareness?

Looking back on what we said about how the algorithm works with different values per user and focusses on the cheaper audiences first. If we apply that to the example above:

  • An awareness/reach campaign will deliver first and most of its budgets in users that don’t click or buy online
  • A traffic campaign will deliver first on audiences that click but don’t buy online
  • A conversions campaign optimising for all conversions will deliver first to people that purchase online, but that avoid clicking on ads

A conversion campaign optimising for post-click conversions will optimise towards the users that you may track in cross-channel attribution tools, but will not be as efficient on delivering overall ROI for the company.

Strategy, planning and the marketer’s job

Traditionally, when looking at the funnel, the marketer tends to find proxies to determine the value of the activity based on the key metrics that are expected. As an example, when looking at acquiring new customers, the marketer may look at traffic as a proxy of the final conversion objective.

However, when we transform that proxy on the objective of the campaign and combine it with automated bidding, we may be harnessing the success of the campaign. The original assumption of “The users that clicks will enter into remarketing audiences and convert” may be wrong, as the audience itself may be focused on users that click and don’t convert.

Therefore when opting for an optimised cost-per-thousand (oCPM), there are two possible ways to follow:

  1. The first one is to focus on the traditional proxies if these are still relevant as secondary objectives of the campaign. Eventually the budgets will grow over the original behaviour boundaries and start delivering to users that are not so good for the proxy but that work for the overall campaign.
  2. The second option is to challenge those original proxies, continually test objectives and measure the results as best as we can on the final success metrics.

Our role as marketers moves away from how well we deliver specific KPIs on certain channels. Instead, the focus is on understanding how useful are the existing KPIs, how are they going to impact the algorithms, and how can we create new ones that get us closer to the final business objective.

To find out more, get in touch.

by Croud
13 May 2020



Related posts