Forward3D used a sample of generic keywords and applied a remarketing list for users that had previously been on-site. We then used AdWords campaign experiments in order to split traffic evenly. 50% of traffic ran at a control bid, whilst the other half ran at an increased bid of 30%.
This combination of user segmentation and A/B bid testing resulted in a fair comparison of the impact of remarketing for search, against regular activity.
We proved that although users who have previously been on-site are more likely to click and convert, they are not as loyal as regular users. As the user is using generics to search, they are still heavily influenced by the top positioned ads. The rule of thumb applies; the lower down the ad space you are, the less likely the user is to click. However, returning users display a propensity to click in both higher and lower positions.
The fact that returning visitors are more engaged both through CTR and CR is not surprising. The real challenge lies in knowing how to balance investment between new and returning visitors. We know that improved ad positions result in more sales, but are those sales cost effective without the use of RLSA? The answer is simple; No.
Had all impressions been served from the higher position, regardless of being a new or returning user, our overall cost would have increased by 43% for a mere 25% upturn in sales. It’s clear that targeting all users, indiscriminate of their past engagement, will result in a less efficient campaign.
However, by tactically pushing bids for the returning user segment, whilst leaving ads lower down the ad space for new visitors, we would have seen a 12% decrease in overall sales, with a cost reduction of 40%. Slightly less sales, but a significantly better ROI.
As is often the case when assessing the success of PPC activity, the aim of the campaign is key.
Remarketing for search is still very much a new feature, and its full potential will not be realised for a while to come. However, it is already clear that value can be added when targeting new and returning visitors.
This experiment is not yet finished, and there will be further refinements before we detect the ‘sweet spot’ for our bid modifiers by using RLSA audiences. There are also other factors to consider, such as altering ad copy, and assessing the impact that might have.
When considering ROI alone, RLSA will definitely help improve the efficiency of a campaign. It is a commonly known fact that returning users have a higher level of engagement, and result in favourable conversion rates.
Moving forward, a key challenge will be maximising user retention whilst maintaining a regular stream of new users which are essential for growth.