The “similar audiences” feature has been around for some time now, but this type of audience list is still heavily underused. So, what are similar audiences? Google uses your existing remarketing lists to characterise the types of users within them and then makes its own audience lists based on users that have the same interests and characteristics.
This list is made up only of people who have been browsing the Google Display Network (GDN) within the last 30 days. You can then use these lists to show both text and image ads to these audiences across the GDN.
How Can You Find Them?
You need to have a remarketing list with at least 500 cookies in order to qualify for similar audiences. Additionally, those users need to have enough shared interests/characteristics for Google to able to create a corresponding similar audience.
Thus, if your remarketing audiences are very broad at the moment, this might make it difficult for Google to create a similar audience list. If this is the case, try honing in on smarter remarketing lists using combinations, or based on people who visited a certain section of your site rather than your entire site. Once you get your targeting down, you should start to see similar audience lists appear.
These similar users lists can be found under the Display tab -> Ad Targeting -> Interests & Remarketing, or in your Shared Library under Audiences. They’ll fall directly under your existing remarketing lists, as shown below.
You may notice that although you have remarketing lists based on different time durations, you’ll always have the same number of people in your similar users list for matching remarketing audiences; this is because the similar users lists are just based on the last 30 days’ worth of data. With this in mind, if your audience composition changes over time, so too, will the composition of your similar user lists.
If you want to show just your similar user lists, you can do so using a filter as displayed below.
Why Might You Choose To Use Them?
Remarketing is great, as it encourages traffic to come back to your site — but these people, of course, already know your site. If one of your goals is to get new visitors to your site and expand reach, then similar audiences is a great place to start.
Rather than guessing who your audience target might be and creating a contextual or topic-led Google Display Network campaign, you could launch a new campaign targeting people that Google knows have similar interests to your existing customers! The reach won’t be as wide as with contextual targeting or even topic targeting, but you can be more certain that you’re reaching the right audience by starting to expand traffic in this way.
As with most GDN activity, you can’t expect people to purchase your products right away; this isn’t search activity where there is intent to purchase. Instead, set up some goals in Google Analytics for other interactions on your site such as newsletter download, quote request, contact-us, a key page view, etc., and keep an eye on results that way.
You should also use Google Analytics engagement metrics to compare the metrics from this campaign to your other campaigns to judge performance. The bounce rate can be a great indicator of targeting correctness, so don’t forget to add this column in AdWords.
The visitors coming from the similar audiences lists will be added to your remarketing list so you can follow them up again later down the line. You could even use Google Analytics remarketing to build a list of people that specifically came in through this campaign so you can follow them up with different messaging!
My Thoughts So Far
- Lists can be small in size depending on how niche your remarketing list is; therefore, don’t expect this is going to take up huge volumes of your budget initially (unless you are advertising a site with millions of visitors per-day)
- This works better for certain audiences over others; the most successful case I have come across to date has been for a jobs board recruitment website
- Text ads can work just as well or even better than image ads, so don’t discount them
- You can get away with paying a much lower bid than with search for gaining new, relevant visitors to your site
- Though it excludes the matching remarketing list audience, you might find it’s worth adding your entire site remarketing list as a negative to this similar users campaign to make sure you only bring in new visitors
- A remarketing list based on your converted users (confirmation page) could potentially be a great place to start if you have large enough converting audience
Improvements have been made to reach an audience composition over the last couple of years; so, if you tried this when it was first released, it might be worth testing it out again. I would recommend starting this activity in a separate campaign with low bids, using the similar audience list tied to your best performing remarketing list first. Obviously, those are the type of users you want to come to your site!
What’s On My Wish List
As with any Google feature, it’s good to think ahead about what one might like to see next! Here are some of the features I would like to see:
- Use with remarketing lists for search ads to bid more for people that might be a good match for your site.
- Use with custom combinations. At the moment this is off the cards, but it would be great to tie this in with something like contextual targeting so you could reach users that might be similar to those visiting your site at the moment when they’re looking at content matching your site’s products. Inevitably this would reduce the scope of your list quite a bit which is probably why this is not an option right now; but, it would definitely be interesting for the future!
My recommendation would be to give this targeting method a go, especially if you’re already running remarketing or GDN activity! Start small and work your way out into different lists based on performance; that way, you don’t need to allocate a huge budget for testing!
As mentioned before, it is key that you start from a strongly-performing remarketing list; trying to extract characteristics and interests from a small, badly-performing or hugely-diverse remarketing list will more often than not result in a badly-performing similar users list.
If you have any thoughts on this particular feature, please share them with me, and I can pass these on to the global product manager for similar users, who would no doubt be very interested in hearing any feedback.