In the summer of 2013, Google officially released a paid search advertising platform out of beta called “Remarketing Lists for Search Ads (RLSA)”. This ad platform gave us, as marketers, a new opportunity to use our audiences within our remarketing lists. Instead of just using remarketing lists on the display network, RLSA allows us to reach out to potential customers within the search network. The great benefit of this is that our ads will be serving to consumers who have already been to a website based off the specific parameters that we decide within our audiences in AdWords. By hyper-segmenting the audiences, we can add creative that speaks specifically to whomever we target within the search results page.
AdWords users have already seen promising results from RLSA. For example, an online tire store saw a 161 percent increase in conversion rate in addition to a 22 percent increase in overall sales.
The paid search team at Parallel Path has been testing this ad platform since mid-November across a variety of clients. Our audiences have ranged from visitors to a specific page on a client’s site, to those users that abandoned a form fill or a shopping cart. With RLSA, our expectation based off the responses from our AdWords reps was that traffic would be much more limited on account of traffic being based off our audiences, not keywords. Our AdWords reps also highlighted the fact that we could be more aggressive with our keyword choice since only users that fit our specific audiences would be seeing our ad copy. For one of the very few times, generic, broad terms set to broad match had the opportunity to perform well! That being said, using broad terms typically leads to increased costs per click, costs that most likely will be higher than what you are typically seeing if you are not as aggressive with your keyword choices as you will be with RLSA.
In an initial test with generic, broad terms in an account that we are targeting abandoned shopping cart users, average cost-per-clicks were coming in 462% higher than the rest of that account during the same time period. Though we saw click-through-rates that were 182% higher, the CPC discrepancy was too high to be profitable.
At this juncture, there were key action items we needed to take in order to immediately improve performance:
- Run negative query mines during the initial test to improve the traffic quality. Since we were dealing with very generic, broad match terms, it was key to improve traffic quality even though we were dealing with very refined audiences.
- We also cut CPCs to just remain on the first page instead of at the top of the page. The revenue performance did not override the high cost-per-clicks, thus we needed to make adjustments. Since we are dealing with users that have already been to the site, not controlling the positions 1-3 is not as important.
- Adjust keywords to “broad modified”, “exact” and “phrase” match. We may lose traffic, but generic, broad match terms cast such a wide net that it can be difficult to control the quality coming in on a daily basis. By refining the searches on our end, we can immediately remove some of the poor queries.
With these tips, I hope you can get your tests launched with confidence and eliminate the trial and error that we often endure when launching a new paid search test. Whether you are targeting B2B or E-Commerce, the potential for success using RLSA is there. Now it’s time to test and remarket to those that have already shown interest in you!