Using Google’s Custom Bidding algorithm to boost digital advertising value


Much of the buying on programmatic auctions is still through off-the-shelf algorithms, these are provided as standard by the main demand side platforms (DSPs). Algorithms are the scripts that instruct the platform what is important to advertisers and their campaigns. They need to be generic to fit the goals of all advertisers on the platform, from FMCGs and automotive to retail.
From our experience in writing algorithms for DSP technology, there is a real opportunity for advertisers to get better value by using a custom algorithm that is designed specifically for what they need.
What's new?
Custom Bidding is an exciting BETA product on Google’s Display & Video 360 platform, and it is soon to be open to all buyers. It supports custom logic built on top of the standard algorithm. This allows more effective bidding and optimisation of programmatic campaigns towards custom goals that represent true value for businesses.
As a Google Marketing Partner we enrolled in the BETA programme last year. We wanted to understand how effective the tool is in driving better results when buying through the platform. We planned to test the functionality by setting up unique buying models for specific clients.
Does it work?
As a basic use case, we started by using Custom Bidding to optimise towards ROAS/ROI. We created a bespoke script to optimise towards a revenue goal for one of our retail clients. We were able to track the results of this by running an A/B test against the standard CPA algorithm and found the custom algorithm produced a 32% stronger ROI for the client.
To build on these results, we wanted to add additional real-time buying variables from our client’s website into the custom algorithm script. We did this by segmenting products into different levels of profitability, and giving more profitable products a higher priority within the script.
Again the results were very significant. The campaign bought by the custom algorithm produced a 39% higher ROI versus the standard CPA algorithm. It also saw a 129% increase in sales of the more profitable products.
What's next?
We are now experimenting with adding further value by introducing more complex variables, as well as additional advertiser industries and objectives. To do this we are integrating clients’ first party data with the algorithm scripts. This is something which has not been possible within the Google ecosystem until now, and we had to lean heavily on our experience as a Google Cloud Partner to automate it.
We want using Google’s Customer Bidding solution to be as easy as possible for our clients, so have built out a hub of proven templated custom scripts, which can be quickly imported into their Display & Video 360 seat.
With this successful test of Custom Bidding we are helping clients gain competitive advantage by customising Google's dominant stack to match specific marketing goals. Although this is just a single step, it highlights how advances in programmatic tech and an appreciation of how client data can be used, are increasingly central to pushing advertising forward.