Similar Audiences for Search are now available in AdWords

After Remarketing Lists for Search Ads (RLSA), Customer Match and demographic targeting, AdWords has now launched Similar Audiences for Search (SAS), which are a powerful way to reach new and qualified users who have a shared interest with one of our existing audiences.

Similar Audiences for Search help to identify users with similar search behaviour as people listed in RLSA audiences that haven't yet been to your site. Initial results have shown similar conversion rates (CVR) as RLSA campaigns, however, Similar Audiences for Search are expected to reach an audience pool seven times larger than that of RLSA. So, bigger and more qualified… Sounds like we should all give it a go, right?

How does it work?

Similar Audiences for Search are automatically created through Google's machine-learning capabilities based on historical data from past visitors to the website - the only requirement is that at least 1,000 cookies need to have been collected in a RLSA. Both lists are constantly updated and, once one user visits a site, they're immediately removed from the Similar Audience list.

SAS how does it work

This means that from now on, we'll be able to detect people who have similar behaviour to our most valuable customers and identify people with higher intent to connect - and, more importantly, more likely to buy or convert. And, if we apply this in a smart way, we can push performance across all campaigns.

How to optimise?

Lists: Define core audiences and have a relevant list of RLSAs, maintain them and keep them updated with all the correct website URL's; the amount of Similar Audiences for Search lists that are created by Google will be dependent on the amount of RLSA lists that you have.

Bids: These are key to campaign success! It's fundamental nowadays to optimise based on your audience's behaviour, as understanding this will ultimately get you the best results.

It's recommended to overlay the relevant Similar Audiences for Search across all your campaigns with a 0% bid adjustment from the start, and then collect data for about 2-4 weeks (depending on the amount of traffic to your site). Once you have enough insights to compare results between all audiences, it's time to start modifying your bids - for example, by increasing bids for the audiences that are more likely to convert.

This logic can also be applied to brand and generic campaigns; however, we can be more strategic in Shopping or generic campaigns as there's typically more room to drive improvements for these.

Let's imagine that we work for a fashion company that's particularly interested in generating brand awareness and new customers through generic campaigns, in a bid to ultimately increase brand traffic and sales. We can, for example, split generic campaigns into three categories in order to understand user behaviour based on keyword affinity and similarity to existing visitors, allowing us to push the best combination:

SAS best combination

Given that lower funnel lists (converters, basket viewers, etc) usually perform better, this strategy means that every list will have a different bid modifier for each campaign, and as such will probably will follow the structure below:


SAS bid modifiers


As always, each brand will get different results, but it's important to monitor it closely in order to adapt your strategy for each account.

What can you expect to achieve then?

Similar Audiences for Search allow you to reach new customers, scale prospecting activity and improve performance, which, in turn, should positively affect metrics like CVR, CTR and CPA - so it's time to get involved.

Filter posts

Powered by

Back to Top

© Copyright 2018 Greenlight. All Rights Reserved Terms & Conditions