Our last post introduced Remarketing as a powerful tool for driving conversions and building relationships by targeting ads to users who’ve already visited your website. In this post we’ll go over the best practices and a few advanced tactics for building awesome Remarketing lists.
To refresh, users tagged with your Remarketing cookies fall into a master audiences list. This includes all your visitors.
Out of this master list, you can create custom lists by applying conditions for eligibility. For example, you can create a list that only includes users who visited a specific product page. You can also provide exclusionary conditions such as eliminating users who’ve already completed their purchase by excluding the purchase confirmation page from the list.
What’s more, if you’re generating lists via Google Analytics, you can take it a number of steps further and apply time-based and sequential conditions for limiting serving to only those folks who stayed on your site for more than five minutes, or who viewed a specific series of pages. There’s really no limit to what you can do with Google Analytics custom lists and it’s for this reason that GA is now the preferred list building option for many advertisers.
One great approach to building effective Remarketing lists is to base them on user intent. You should always consider what a user’s behavior says about them. For example, it’s quite reasonable to assign greater value to a user who visited a shopping cart page (thus demonstrating the addition of a product to their cart) than a to user who bounced from your homepage. Similarly, a user who viewed multiple product pages relating to low value products would likely be less valuable than a user who viewed multiple high value product pages.
Why you should build granular Remarketing lists:
- Enables audience lists based on intent / value
- More bidding control
- Customized messaging
- Better organization
- Elimination of low quality traffic
Defining Your Goals:
Hopefully, you’re measuring more than one type of conversion action. You may have a purchase-based macro-conversion as well as a number of engagement or action-related micro-conversions a time on site goal or e-book download, for example. The Remarketing lists you create should be designed specifically to execute these goals and every list should ideally contain a user segment whose future behavior you can reasonably predict. For example, you should assign a higher degree of purchase probability to the shopping abandoner than to a user who merely visited a product page. Granular lists mean granular bidding, so you can set a custom bid for these lists based on your assumption of their greater purchase intent. This works for micro-conversions too. A user who downloaded your e-book, or viewed over ten pages on your site should be prioritized because their actions demonstrate a heightened interest in your product or service. This should also be reflected in your bidding.
You can even get really awesome with this if you’re assigning values (and I hope you are) to your micro-conversions, and understand the ratio of micro-conversion completion rates to macro-conversion completion rates. Armed with this information, you can establish an ideal cost-per-click for your e-book download audience based on the statistical likelihood that they’ll complete a macro-conversion once they’re back on your site.
Isolating Your Lists
One issue you’ll quickly come across when building your lists is that users often fall into multiple list buckets. For example, if you intend to target users who placed an item in a shopping cart, it’s likely that a great deal of them also belong to your “purchasers” list — users who reached your purchase confirmation page.
However, you just want to target those who added an item to their cart, but did not complete the purchase. This requires a yet another custom list that includes only users who viewed a cart page, but excludes users who reached the purchase confirmation page. This can be easily set up in either the Google audiences library or in Google Analytics, depending on which product you’re using to build your lists.
This is a common issue due to the frequency with which users qualify for multiple lists as they explore your site. In general, the best approach is to identify a hierarchy of user intent through the conversion funnel and add each defining action as an exclusionary condition in descending order.
For example, your cart abandoners are probably your most important group in terms of their likeliness to convert. You want to isolate targeting for that group to non-purchasers only so add the purchaser list as a negative condition. Next, you might want to exclude your cart abandoners for your product page viewers if you want customized messaging for each. Lastly, you could add product and category pages as negative lists to your homepage viewers to ensure both that product and category viewers are shown customized messaging, as well as to have more bidding control for each group.
Remember, many of your cart / product page / and category page viewers entered your site through the homepage, but then took additional action that made them more valuable to you. You want to take special care of these users, not show them the same messaging or set the same bids as for those users who reached your homepage and simply left. The more you can isolate your audience buckets, the better.
Think of lists as buckets, with each relating to degree of conversion likelihood.
- Time on site
- Number of pages viewed
- Viewing of specific product pages
- Viewing of specific category pages
- Adding products to a shopping cart
These actions all provide important information about your users. By segmenting each group into a separate audience list, you’re giving yourself the flexibility to bid separately for each, based on what each action says about the user’s likelihood of converting.
Hopefully this will give you some ideas for creating effective custom Remarketing lists based on your own unique business goals. Just remember:
- It’s hugely beneficial to take a granular approach to audience segmentation.
- Things aren’t always this clean in the real world as they are in theory. There are always surprises when it trying to predict user behavior based on past performance.
- Putting the time and care into creating smart Remarketing lists will definitely pay off in the long run.