RLSA: A Practical Use Case
Through the holiday season, I began using Remarketing Lists for Search Ads (RLSA, formerly known as RFSA or Remarketing For Search Ads), and now it’s a regular part of my SEM repertoire. The tests I ran were on Holiday/Christmas products for an e-commerce stationery site. This selection of products is not the most ideal for RLSA as they are relatively quick purchase decisions—you see them, you like them, you buy them—and there is not any extended period of time for consideration, planning, and price checking involved in these purchases compared to other products on the site.
RLSA is best suited for products for which there is a long consideration process, involving multiple players, but this was a just a test to see how well it worked.
A couple of things to be aware of in breaking out RLSA:
- They only work with old remarketing pixels.
- You cannot target users with messaging that suggests that you know they have performed any action (this very intelligently shuts down the common remarketing creepiness factor).
- You can alter landing pages to your heart’s desire.
- You cannot use your YouTube remarketing lists for RLSA (yet).
The basic methodology for my test was simple: I wanted to target the people who had already been to my site with higher bids to drive my ads into higher positions and ultimately drive return customers to the site at a higher rate than new customers.
The hypothesis being that return customers would be more likely to convert than those who had never been to the site. This test ran for 25 days in Holiday, and 17 days for Christmas at the end of the season. This is a very basic level test, but my learning proved to be very valuable and has continued to be part of my process for other product categories.
You will have to be approved for the RLSA beta by a Google rep. Once you are approved, you will be able to assign audiences to your search campaigns as you would for any GDN campaign, except the only targeting that will work is remarketing lists. Select the audiences tab, and then select the Interests Remarketing button. After this you will need to click the green “Add interests remarketing” button and select an ad group; this will display your targeting options.
In your targeting options, you want to click on remarketing lists and select the needed lists; in my example, I selected the Holiday—30 tag, which is all visitors to holiday pages in the last 30 days.
The bad news is that you have to apply audiences at the ad group level for RLSA. The good news is that it is now supported by AdWords Editor, and you can add retargeting audiences to RLSA groups the same way you can with regular GDN retargeting. This is a small change on the Google side that makes scaling these campaigns much, much easier.
Results of My Test
I found that in some cases I was able to more than double my conversion rates. And in cases where I had otherwise struggled to keep my ads in good positions due to CPA constraints, I was able to keep my ads in much higher positions for return users. My only regret with RLSA is that I didn’t start earlier, or test more.
For my general holiday campaigns, I found that I was able to more than double my conversion rate on head terms (highlighted in blue) and improve position by 1.3 spots; this moved my ads into the coveted top 3 consistently, which increased my CTR by 71 percent, all while maintaining a CPA that was 38 percent lower than with new users.
For my regular exact match campaign (highlighted in red), I saw an even bigger increase in conversion rate, going from 2.23 percent to 8.26 percent, a 270 percent increase, despite a less effective position increase from 2.7 to 2.2 which led to only marginal increases in my CTR.
The bottom line with the exact match campaign was that the users who had already been to the site were much more likely to convert.
For the Christmas head terms, I was able to move from position 4.8 to 3.7, putting me in the top 3 much more often. This increased my CTR by 278 percent while maintaining a CPA of 25.4 percent lower than for new users. The conversion rate difference here was not as great as with the holiday head terms, but it did allow us to compete in a position that we wouldn’t have been able to reach if new and returning users were to be lumped together.
Now that I’ve explained all the reasons that this beta is so great, I need to mention that during our test the returning users campaigns only made up about 10 percent of my total impressions, but that is a limitation of the product behavior, not of the beta itself.
The only limiting factor for RLSA is the size of your lists; return users should be of higher value than new users in almost every case. This was a preliminary test of the effectiveness of this beta, and I have a lot of new ideas I plan to roll out in Q1 using this beta (I will update you on them as they roll out).
A couple of RLSA test ideas for you to think about for Q1:
1) Run broader keywords in the RLSA campaigns for higher-value users.
- If you normally bid on +red +shoes, you should test advertising on broader search terms like +shoes for those who have been to the site in the last 7 days.
2) Start an RLSA campaign for abandoned carts, or people lost in a conversion/lead funnel.
3) If a user’s value is very high one day following a visit, try running dynamic search ads on a RLSA campaign with a one-day remarketing tag applied to it.
- Dynamic search ads tend to be overzealous in what they serve ads for, but if the user is very high value within 24 hours of visiting your site, it might make sense to be very aggressive in the terms you show up for.
- Note that this is a bit of a risky test as the dynamic search ads product is very aggressive in the targeting of users.
My best advice: If you haven’t started at least testing RLSA, start now. I’d love to hear your own experiences with the beta in the comments.