Search Retargeting: Supercharge Your Search Ads with RLSAs and Google Analytics
Over the last few years, the search landscape has been evolving. It’s moved far beyond keywords, and now uses many audience based marketing features. The first example of this audience based marketing within search ads came with the release of Remarketing Lists for Search Ads (RLSAs) back in 2013. Since their release, we’ve seen RLSAs go from strength to strength, with many new functionalities and features. One of the biggest updates to the functionality of RLSAs has been their integration with the Google Analytics remarketing code. This update was relatively unpublicised, but the impact it has was huge. It meant that we were no longer stuck with creating lists based only on page views, and instead could use almost all the data in our Google Analytics accounts to create lists. Today RLSAs are one of the most powerful, and also one of the most underused AdWords search ads features. In this post I’ll be sharing the presentation I gave at SMX London on the 19th of May 2016, and providing some ideas for powerful ways to use RLSAs and Google Analytics data, as well as some set-up tips and tricks.
You can find my full presentation on Slideshare here, and in the post below I’ll be talking about each slide in detail.
Using Remarketing Lists for Search Ads (RLSAs) with Google Analytics Remarketing Code
Using RLSAs with the Google Analytics means that you can use almost all of the data in your Google Analytics account within your RLSA lists. This includes:
The AdWords Remarketing Tag vs the Google Analytics Remarketing Tag
There are many other benefits for using the Google Analytics Remarketing tag as opposed to the AdWords Remarketing tag.
Using the AdWords tag means that you need to obviously place this additional code on your website, whereas the Google Analytics remarketing tag simply works using the standard Google Analytics code, so there’s no need for any additional code to be added to your website.
The AdWords tag limited advertisers to only being able to create remarketing lists based on page views, unless Custom Parameters were added to the code. Analytics lets you create lists using everything that was available with Custom Parameters plus lots more!
You can keep both the Analytics and AdWords remarketing codes on your site without any interference. If you use AdWords to run dynamic remarketing you will need to keep the AdWords remarketing code on your site as this code is what allows the dynamic remarketing to function.
If you haven’t already activated your Google Analytics remarketing tag it’s very straight forward. Simply go to the Admin section of your Google Analytics account and then find the tracking info section, and finally within the Data Collection section you’ll see a switch under the heading saying Remarketing which you simply need to toggle to ‘on’.
Tactics for using RLSAs and Google Analytics Data
Increase bids for users who have visited your website the optimum number of times before purchase:
For example, a holiday retailer such as Virgin Holidays might find that their average website visitor visits their site at least 5 times before they make a purchase.
They could find out this information within their Path Length report in Google Analytics:
In the example above you can see that most users visit the site 5 times before making a purchase.
Based on this data, you can create an RLSA list in Google Analytics under the Audience Definitions section within the Admin tab. Here you would specify that the user must have had at least 5 sessions but not yet completed a transaction:
The next step is to apply this list to your existing ad groups in your AdWords account. You can do this in AdWords Editor or the AdWords interface, under the Audiences tab:
Apply this list using the Bid Only setting. Using the Bid Only setting means that anyone can see your ads if they are searching with your keywords, including anyone in your RLSA audience, however it gives you the option to see bid adjustments if the person searching is a member of your RLSA list. You can learn more about this setting here.
Now you’ve applied your audience, you need to choose the amount you’ll set your bid adjustment as. You can do this within the Audiences tab where you’ll see the different audiences applied to your ad groups and then a column called ‘Bid adj’ where you can type in a plus or minus percentage.
My advice would be to apply the remarketing list to your ad group and let it gather data against the ad group before setting the bid adjustment. You can then make sure that the audience is converting at a higher rate than your standard ad group performance therefore decide whether it justifies a bid increase. If you’re wondering what exactly to set your bid adjustment at, Periscopix have written a handy blog post about it here.
Use Facebook ads to drive cost effective new audience traffic to your website, and then reach these users again when they are searching with broader research style terms on Google.
For example, Toni&Guy could run awareness ads on Facebook designed to get users thinking about trying a new hairstyle. They could then use RLSAs to reach these users when they are searching for new hairstyle ideas on Google.
For this tactic Toni&Guy would need first create their Facebook ads, and use UTM tagging on the website links:
The UTM tagging criteria is then used to create the RLSA list in Google Analytics. This list would be created under the traffic sources audience builder tab and would use the name of the Facebook campaign, source and medium as written in the UTM tags.
Create a new AdWords campaign and ad groups containing your new ‘top of funnel’ research term keywords. It’s important to create a new campaign for this so you can easily control your budgets, because if you create these ad groups in an existing campaign and they are alongside other high spending ad groups, the high spending ad groups might use the campaigns daily budget before your lower spending RLSA ad groups get chance to show.
In the example of Toni&Guy, these keywords might be things like ‘hair cut ideas’ ‘new hairstyle ideas’ or ‘hairstyle inspiration’. The idea is to select keywords that you probably don’t already bid on, and that users who are further up the path to purchase might use to search when conducting research.
The ad text you create in these particular RLSA ad groups are only ever going to be seen by your RLSA audience because you’re going to use the Target and Bid setting (rather than the Bid Only setting), so you can create this ad text bespoke for that audience. Consider that the searcher is in the early stages of the path to purchase, so your ad text should reflect this. Don’t go with too much of a hard sell or mention prices, but instead encourage them to find out more or offer them something free / low risk to draw them back to the site. In this Toni&Guy example the ad text could encourage the user to book a free consultation to help them decide on a new hairstyle:
Remember to apply this list to the new ad groups you’ve set up in your new campaign, using the Target and Bid setting. The Target and Bid setting means your ads will show only if the user searching is on your remarketing list. If the user is searching using these keywords but is not on your remarketing list, none ads of yours will show.
Use RLSAs and Dynamic Search Ads together to show ads for your entire product inventory if the user searching is someone who regularly makes purchases from your website.
Some customers are creatures of habit. Even if they can find the same product cheaper elsewhere, they’ll choose to buy it from your website because they’ve bought from you before and they know and trust your website and business. You can capitalise on this behaviour by combining RLSAs and Dynamic Search ads to always show ads for every product in your product inventory that the user searches for, even if you don’t normally show ads for them, because you know the searcher is a loyal customer who is likely to make further repeat purchases from you.
The benefit of this tactic is huge, because it means you can have ads showing for anything in your product inventory if the user searching is a regular shopper, without having to spend time building those ad groups upfront.
Amazon would be a company who would find this tactic useful. I’m usually a savvy online shopper, but since buying Amazon Prime, I’ve ended up buying all sorts of extra things from Amazon that I would normally wait and buy in my weekly grocery shop, like moisturiser or cleaning products. Amazon could be using dynamic search ads and RLSAs so that as long as the product I’m searching for exists in their product inventory, they could show text ads for it to me. They know I’m a regular customer, so even if they don’t normally advertise on these products because they maybe aren’t competitive on price, it’s worth bidding on them when regular shoppers like me are searching because I’m still likely to buy it from them.
This RLSA list would be created in Google Analytics under the conditions tab, where you would specify that the number of transactions per user needs to be at least 5 for example, and that the list duration however long the period that you want those transactions to have occurred is. For example if the list membership duration is 4 weeks, then the user will have had to have made 5 purchases in four weeks to be part of the remarketing list.
Apply the list to the Dynamic Search Ads campaign you have created, using the Target and Bid setting, so only those on your RLSA list will be able to trigger these ads.
Use Google Smart Lists
Very rarely would I recommend letting Google do that hard work for you, but Smart Lists can be very successful as RLSAs if used correctly.
Smart Lists are lists which Google create to include users who Google have identified as being close to making a purchase in your industry. Smart Lists work best on ecommerce accounts which tend to have a huge wealth of data. You need at least 500 transactions per month and 10,000 daily page views for Google to create Smart Lists based on your data. If you don’t meet these thresholds, Google will still create Smart Lists for you but they’ll be based on data from other similar websites and so they don’t seem as effective. The data Google use to determine what makes audiences similar to your existing remarketing lists includes the users location, device, browser, referrer and session duration.
A good example of a retailer who might be able to successfully use Smart Lists could be Marks and Spencer because they will have a huge amount of conversion data. They could create a Smart List and then apply it to their existing ad groups and see if this list converts at a higher rate to their standard audiences, and increase their bids for this audience if so.
When you click the create new audience button under the Admin tab in Google Analytics, you’ll see Smart Lis appear in the options for ready-made lists. Simply click that option to create a Smart List. You can then apply this in the usual way to your existing ad groups and chose the bid only setting.
Show users who have engaged with your email marketing specific ad text when they search on Google.
For example Fabletics run lots of email marketing and could tag the links in their emails so that if a user lands on their site via their email marketing, they will be added to an RLSA list which would then be applied to new search ad groups which are duplicates of their existing campaigns, so they can customise the ad text in these campaigns for these specific users. The ad text would then reinforce any promotions of messages that were used exclusively within the email marketing activity.
This list would be created in a similar way to the Facebook ads example above, where you will specify the source, medium and campaign of the traffic:
Turn high value Black Friday shoppers into repeat customers.
Black Friday sees shoppers go a little bit crazy for any good deals online. Many will purchase from new retailers for the first time during Black Friday because the offers are too good to refuse. Using RLSAs you can increase your bids for these high value Black Friday shoppers when they are searching after Black Friday in an attempt to bring them back to your site for repeat purchases.
For example John Lewis usually have fantastic deals around Black Friday, and so they could increase their bids for all users who spent more than £500 in a single transaction on Black Friday, when they are searching later on in the year.
To create this list you would go to the Conditions tab in the Analytics audience creator and state that the session date is the 27th November 2015 (Black Friday last year) and the revenue per session is at least £500.00.
Increase bids for users who clicked a promotional ad but didn’t convert.
For example Boots have an offer on Marc Jacobs Daisy perfume, which was promoted via their text ads:
If a user clicks this offer and goers through to the website, it’s likely that after they’ve done their comparison shopping they’ll come back to the site and make a purchase because Boots know that the price they’ve offered is very competitive. Using RLSAs Boots could increase their bids when the user is next searching so they have the best possible visibility and impression share for that audience because they are likely to convert on their next visit.
To create this RLSA list simply state that the ad content contains whatever the code that was used in the ad was, and that conversions are zero. I’ve also layered on the criteria that the session duration needs to be at least 60 seconds, to weed out the users who clicked the ad and visited the site and then realised this was not in face the perfume they were looking for. The list would be applied as bid only to existing ad groups:
Increase bids if the user searching has generated a high conversion value for you historically.
For example, Waitrose might have some shoppers who spend more than their other shoppers on their monthly or weekly shop. They could use RLSAs to increase their bids of the user searching for terms like ‘groceries online’ or ‘M&S groceries’ is one of their high value customers. This tactic is particularly effective if your consumers aren’t very loyal and often flick between different retailers, because you can make sure that if your high value customers are searching for other supermarkets you have a good high bid to increase your visibility at those times.
TO create this list simply use the Ecommerce section within the Google Analytics audience builder and state that the revenue per session needs to be more than £500.00 for example. You could then apply this to your existing ad groups as a Bid Only list so you can then increase your bids for these users. If you don’t normally bid on competitor names, you could create a campaign containing competitor names but apply this list as a Target and Bid list, so your ads would only be triggered if the searcher is on your RLSA list.
Adjust your bids if the person searching is similar to your existing customers.
Similar Audiences are created by Google automatically, based on the characteristics of your existing remarketing lists. Your existing remembering list needs to have been running for a fair amount of time and have lots of members. The larger the original remarketing list the better, because Google have more data to draw conclusions about what makes these users similar from their browsing activity in the last 30 days and therefore create more accurate Similar Audiences lists.
For example, Feel Unique could use Google’s similar audience list based off of their existing remarketing list of customers who recently completed a purchase. The list could be applied to their existing campaigns as a Bid Only list, so they could increase their bids of the Similar Audience list tends to convert at a higher rate than average.
Similar Audiences lists don’t actually need to be created, as Google automatically create them so they simply appear within your Analytics Audiences tab once your original remarketing lists are large enough for them to be built:
Upsell on products you wouldn’t normally bid on, because you know the user has just purchased something related.
For example Pandora might not usually bid on silver cleaning kits because they are a low value product that doesn’t usually generate a good conversion rate for them. However, if they know that the person searching has recently purchased a silver bracelet from them, then the likelihood of the user choosing to purchase the cleaning kit from them is much higher, so they can afford to bid on these keywords.
To create this list, state that the product purchased by the user must have contained the term ‘silver bracelet’. You’d then create a new campaign containing your silver cleaning kit words and apply this list using the Target and Bid setting, so only those on your remarketing list can trigger your ad
Increase bids if the user searching has purchased similar brands from you in the past.
For example ASOS sell lots of different brands of petite clothing. They could increase their bid if the user searching is searching for any petite brand they stock if that user has previously bought another petite brand from them in the past, because they are more likely to purchase petite clothing from them again.
Like all of the RLSA list examples I’ve given, this list can be applied to both your search and shopping campaigns.
Simply create the list to state that the user must have purchased one of your petite brands in the past, using the ‘or’ functionality instead of the ‘and’ functionality between each of the different petite brands you sell:
Increase bids if the user searching has indicated a strong interest by completing a micro-goal on your website, but not yet a macro-goal.
For example, Halifax could increase their bids and visibility if the user searching has used their mortgage calculator on their website (micro-gaol), but not yet used their contact form to request a meeting with a mortgage advisor (macro goal). If you don’t have any micro goals on your website, then you could create one out of specific session duration criteria. For example, you could set up a goal to track anyone who spends more than 3 minutes on your services pages because that is an indication of strong interest.
This list is very simple to create, you simply select goal one (the macro goal) is zero, but goal two (the micro goal) is at least one. In the image below you’ll notice that I have a lot of goals set up so goal 20 is my macro goal and goal 16 is my micro goal:
Tips and tricks for using RLSAs
Remarketing Lists for Search Ads and Automated Bidding Strategies
RLSAs work differently with different types of automated bidding strategies. If you apply RLSAs to ad groups using Conversion Optimiser, Target CPA, or ROAS, the bidding strategy will automatically override any manual RLSA bid adjustments you’ve made and use the audiences predicted performance to dynamically set bid adjustments during the auction. If you’re using Enhanced CPC on the other hand, then you’ll find that the bidding strategy makes bid adjustments on top of any that you have set manually.
Google advise that you avoid any audience overlap if you have more than one RLSA list on an ad group which is using automated bidding strategies. This means members of one audience should not also be members of the second audience on the ad group if it uses automated bidding strategies. Having audience overlap can cause conflicts for automated bidding strategies, so if in doubt just stick with one RLSA list per ad group.
Bid Adjustments Stack
Don’t forget that like all other bid adjustments, RLSAs stack on top of each other. For example if you have several audience bid adjustments on an ad group, Google will apply one, then apply the next one to the total of the last one.
It’s important to test your audience list durations as these can have a big effect on the success of your RLSAs. The maximum audience list duration for RLSAs is 180 days, which is less than display remarketing lists of 540 days. If you use a remarketing list with a membership duration longer than 180 days, you’ll still be able to apply it to your search ads but the users will stop being valid within search after 180 days.
Other handy tips for RLSAs:
- Lists need at least 1,000 members before they can be used with RLSAs, unlike display remarketing which only requires 100 members
- You can have up to 2,000 remarketing lists per Google Analytics account, so there’s plenty of room for testing new lists
- In Market Segments and Days Since Last Visit are not applicable for RLSAs, and neither are demographics such as age and gender
If you’re not already using RLSAs with your Google Analytics data, I really recommend giving it a try. I’ve seen incredibly strong conversion rates from them. Even if you aren’t yet ready to actively set bid adjustments or set up new campaigns for them, just create the lists and let them gather data attached to all of your ad groups. You can then analyse this and it will show you the kind of potential performance you could expect with RLSAs.
Have you used RLSAs with Google Analytics? If you have any more handy tips or tactics please comment below and share them.