Is DV360’s Auto-bidding the AI Solution We’ve Been Waiting for?

Read Time: 4 mins

When it comes to your display and video campaigns, managing your ad budget and optimising your bids is a tricky beast. With the sheer amount of information available on users – their demographics, their searching habits, and where you’re likely to find them – it’s incredibly difficult to manually find the perfect bid to reach the right person at the right time. Auto-bidding takes away all the hassle, enabling you to optimise your bids using countless data points that would be impossible for a human to group together.

What is Auto-bidding?

Auto-bidding uses AI to dynamically optimise bids and ad delivery to best achieve your goals. By learning from both failed and successful bids, the algorithms are able to predict the behaviour of your target audience. 

To put this into context, let’s use the example of a prospecting line item targeting 18-to-24-year-olds in the UK with an interest in football. Our goal is to drive these users to site at the lowest cost. 

In a fixed bidding scenario, you’ll use the same bid for each individual user in this audience. Auto-bidding, on the other hand, layers a number of more refined data points to create a unique bidding strategy. This takes into account a number of behavioural trends relevant to this audience group, such as:

  • Users in major cities have a higher CTR and better CPCs compared to those in the countryside, despite having a higher eCPM.
  • 19-year-olds engage with the ads at the same rate as 24-year-olds, but tend to be a cheaper audience to target.
  • Engagement rates are best between 6 pm and 8 pm.
  • Contextually, advertising alongside football news content garners higher engagement than alongside technology review sites.
  • Showing 20-year-olds an ad 5 times a week leads to improved overall performance.

The Benefits

As you leave the AI to do its job, you’re able to focus your attention on other things, ultimately giving you more time to develop strategy or work on creative outputs. Advertisers no longer need to worry about analysing engagement and conversion data to determine the best way to adjust the fixed bids – instead, looking at the overall picture of performance. 

Also, the algorithms are able to interpret these swathes of data in a way which a human simply cannot. This unlocks a huge potential for performance to improve which would not be possible without AI. 

Finally, as Google completes the roll-out of the first-price auction to Google Ads Manager, understanding how auto-bidding works is a necessity if you want to succeed. While first-price auction buying improves the bidding across all exchanges – as the auctions take place simultaneously – it will inevitably lead to increases in CPM. Complacency in diversifying your bid strategy will likely lead to poor performance as you are left behind the pack. 

Bear in Mind

However, as with any new technology, there are a couple of things to remember as you get started. 

First things first, you’ll need to make sure you have enough conversion data available for the algorithms to work from. The exact amount is tricky to pinpoint, as the more data available the more effective the algorithms can be. You’ll need to find a balance between the value of your conversion goal, and the amount of data this conversion goal can give your algorithms. Also, be mindful that in the testing and learning phase it can take anywhere between a week and a month before you start to see improvements, so patience will be crucial. 

When choosing the type of strategy to employ, it’s important to note that auto-bidding strategies designed to minimise your CPCs will often bring in low-quality traffic or clicks that fail to convert. Instead, set up your auto-bidding strategies to minimise your CPA with the conversion goal as a landing page visit. By setting up your account in this way, you are able to take potential drop-off between clicks and landing page views into consideration. 

Another feature available through auto-bidding is Insertion Order-level budget allocation. Here, there is an extra layer of AI which evaluates the performance of your line items before allocating your budget accordingly. From experience, however, we recommend analysing your results, taking into consideration the CPAs and amount of inventory available (cookie pool size), and then manually allocating the budget.

So, is DV360’s auto-bidding the AI solution we’ve been waiting for? 

From the results we have seen, we definitely recommend implementing Autobidding – but there are a couple of caveats. If your conversions are low, or you’re retargeting a very small audience, you won’t be able to generate enough machine learning data for the algorithms to optimise from. However, if you have plenty of data and the right strategy, auto-bidding is a great way to unlock the potential of your display and video campaigns.