For many retailers, remarketing is considered to be one of the most efficient channels available for Return on Ad Spend. If done correctly, it can also be a great channel for customer acquisition.

But for most advertisers, remarketing is often done wastefully, meaning that if your remarketing partners were more efficient you’d have more dollars to pursue new customers or more dollars to return to the bottom line. Let’s discuss how Lytics can help.

Remarketing starts from a "cookie pool," a nebulous bunch of individual cookies with varying degrees of recency of last visit, product/content affinities, and likelihood to transact. Yet, most remarketing solutions treat these cookies more or less than same. Lytics can help to quantify things like recency, depth of visit, likelihood to return, likelihood to transact, etc. And Lytics can combine these with discrete factors such as categories browsed, content affinities, and whether or not someone is a subscriber or customer. Using this data alone can enable most of our clients to pay back their investment in Lytics within days or weeks.

To get started, we recommend first focusing your remarketing spend on the users most likely to transact. Almost always, these are the cookies most-recently added to the cookie pool, and those who show the greatest depth of engagement. This is a great place to start for clients first setting out to segment their remarketing cookie pool. Here’s how:

Start by creating a new Segment under the "AUDIENCE" nav

Give your new segment a name and be sure to select “Enable API Access” and provide an ID which we’ll use later in Facebook.

Begin building the segment with a qualifier for depth of engagement.

This could be somethings like a count of pageviews:

Advertisers have had success with semi-arbitrary metrics like "3+ pageviews," but for advertisers we’ve tested have found greater success (up to 60% more efficient when we’ve tested) by using Lytics Data Science, which scores things like “quantity” of interactions relative to an advertisers entire cookie pool rather than having one hard metric for all cookies. We’ll use Lytics Quantity Score for the purposes of this demonstration:

We selected the number 5 because it mapped to a size of cookie pool of the top 10%. You should play with these, and make sure to select a number that makes sense for you.

Next, we added a qualifier for how recent the users were

Of course, you could do a fixed date range like last 7 days or last 3 days, and this usually works well. But Lytics data science, again, has outperformed this in the past since it looks at the browsing patterns of your entire cookie pool:

In this case, we chose 95 because it was again the top 10% of the cookie pool.

Finally, if you’re really looking to get the extra mile, consider removing users for whom you have an email address.

Consider that the vast majority of your conversions from remarketing are "view-through," meaning that the user didn’t actually convert after clicking your ad. How many of those conversions might have occurred anyway? Might some of them have come from people who received the next email marketing message or a browse abandon email campaign? For most advertisers we’ve looked at, usually over 90% of view-through conversions would have happened with an ad. Plus, ads are expensive. Facebook newsfeed CPMs can range from $7-$25 depending on the time of year. Why not save that money and get the conversions from your subscribers for zero incremental cost in your email campaigns? For the right advertisers we highly recommend this. Here’s how you’d do it:

Simple. Add a filter to exclude anyone whose email exists in your account and EXCLUDE them.

Next, we need to find this audience in Facebook.

First, follow our instructions to authorize your Facebook integration if you haven’t already. Assuming you’ve already linked your Facebook Ads Manager to Facebook, you can find your audience in the Audience Manager:

If you can’t find it, create the audience manually as follows:

1. Click "Create Audience" within Facebook > Audiences and select “Custom Audience”
2. Select "Website Traffic" from the menu that appears.
3. Be sure to select "Custom Combination" from the dropdown.
4. Then select "Event" under “Include”
5. Select "Lytics Audiences" under “Event”
6. Under the parameters that appear, select the audience ID you’ve created that we mentioned upon building the segment:
7. Enter "true" after “contains”
8. You can manually use Facebook’s cookie pool expiration, or skip that and give your audience a name. Now it will be available for targeting.

Next create an ad on Facebook as you normally would

Facebook documentation offers a ton of help here: https://www.facebook.com/business/support/topic/creating-ads

Our advertisers have had the most success in the Facebook Newsfeed and Instagram. Talk to your Client Success Manager for more advice on using Lytics to power ads on Facebook, Google, AdRoll and more!