Retargeting is a staple in nearly all PPC accounts at this point. Everyone has bought into the idea that just because someone didn’t purchase the first time they came to the site, doesn’t mean they aren’t going to purchase at some point and it’s important to stay in their mind space.
But despite the advancements we’re seeing from the channels in the ways we’re able to segment and target our audience, most accounts still have one campaign labeled “Retargeting” that has a single user list in it with the date range set to the longest it can be.
While this might be getting you the results you want, online marketing is no longer just about getting returns. If you’re hoping to protect the longevity of the brand you’re advertising, it’s important to keep the customer base happy. And sending them ads in perpetuity because they clicked on your ad over 9 months ago just doesn’t scream, “I care what you think about us.”
Now, I’m not saying you shouldn’t target people until the last possible day you can before the cookie expires. I’m saying it might be in your (and their) best interest not to. And we can back that hypothesis up with data. Let’s jump into what building a funnel looks like.
The Idea of a Time Funnel
The phrase itself might sound like a tornado with a bunch of clocks in it, but I promise that’s not it (read: it’s not that cool looking). The idea of a time funnel is to utilize time intervals to mimic the stages we put together in a buyer cycle. Here’s a common sales cycle example:
The top of funnel actions are those furthest away from the conversion action and as you move down, each stage gets closer to a potential sale.
For our purposes here, we’re going to use a similar structure, but with a couple slight adjustments.
In our time funnel, each time grouping has its own stage like the buyer funnel, but each time the date range gets longer, the person moves further up the funnel, further from the conversion. Although this isn’t a perfect science, we as marketers know that recency of visit is an important indicator as to whether someone will convert or not. So each time someone falls into a further time range, it stands to reason they’re further from purchase.
Each Cycle will be Unique to Your Company
The example I gave above is simply an outline that fits a number of clients I work closely with. But that doesn’t mean it’s necessarily the right timescale series for your business or your client’s businesses.
Whether you’re e-commerce, lead gen, have a long or short sales cycle, luxury product or household good, etc., all of these things can play a role in what your stages should be. Here are a couple examples of different retargeting time funnels that might make sense for your company.
- 7, 14, 21, and 30 Days
- 1, 3, 7, and 14 Days
- 30, 60, 90, and 180 Days
There’s no right or wrong answer, just make sure your date ranges make sense for your business. Don’t worry, we have a process for that here in just a minute. But first, let’s talk about the benefits of this strategy.
The Benefits of a Retargeting Time Funnel
The name of the game here is optimization. Think about an audience with 10,000 people in it that’s been accruing for the past 90 days. You have this one, large audience with widely different times for their last interaction with your site. Some visited yesterday, some visited 2 weeks ago, some visited 3 months ago. When you treat each of these users the same way, which you are if you have just this one audience, they’re not all going to respond in the same manner.
If your call to action is a hard sell, the users from the yesterday might be enticed back, but those from a couple weeks ago might be less excited and those from 3 months ago might be struggling to remember who you are.
If you’ve got a softer call to action like a content download, you might be able to entice the users from 2 weeks and 3 months ago to come back for some more information. But you’re most likely underselling the user from yesterday, not getting the full value from their interactions that you could.
Building a time funnel with your retargeting audiences can be a simple way to begin breaking up that large retargeting audience into smaller portions to allow you to speak to those unique user groups differently. I used the idea of differing degrees of calls to action, but that’s just one way you can differentiate the messaging between these groups. Overall, it becomes easier to optimize that 10,000 user list when you have it broken down into groups of similar users, being targeting in different audiences, so you can craft new messaging, adjust bid modifiers, test new calls to action, and leverage all of the other optimization levers we have at our fingertips.
Determining Your List Time Ranges
As I mentioned above, each company will most likely benefit from having a series of time ranges that make sense for their customer base. So how do you come up with those ranges? There are a couple different ways:
Google Analytics Time Lag Report
If you’re familiar with Google Analytics, you’re most likely aware of the Time Lag report. If you’re not familiar with it, then hold tight, because you’re about to be.
In Google Analytics, in the Conversions report section under Multi-Channel Funnels, you’ll see the Time Lag Report.
Once there, you’ll see a chart with a number of blue lines broken down into different groupings based on the days it took for someone to convert after their first visit to your site.
Odds are your performance will invariably look something like the chart below, with the first time grouping having the largest shares of conversions and conversion value and each time range taking up less of the share as you move down the chart until you get to the 12-30 day range, where it will jump back up a bit.
There are a couple things I like to adjust on this chart before making any decisions about how users interact with our client sites.
First, I adjust the time range to the longest it allows: 90 days.
Second, I adjust the chart to only show the specific conversion actions I’m going to be offering in my retargeting campaigns. By default, this report will include every goal you’ve set up for your Google Analytics account, meaning that some of these might not make sense to analyze for this purpose.
You can adjust the conversions included by clicking on the “All” dropdown under Conversion, just next to the time range, then unchecking the box next to all conversion actions not suited to this analysis.
Now when I look at the Time Lag report, my findings look much different.
The overall trend still applies, but the breakdowns of time ranges have changed a bit.
From this analysis here, I might determine that it’s in my best interest to break down our retargeting user list into 4 groups: 1-14 days, 15-30 days, 30-60 days, and 60-90 days.
Talk to a Sales Rep/Database Analysis
The approach above can be really powerful when the user you are trying to understand performance for is a single decision maker and utilizes the same device each time they come to your site. But for non-linear sales cycles, it can be tough to get this type of analysis and feel confident in it. Even the analysis above I have a hard time believing, but that’s because this is a non-linear sales cycle, not because I think the data is faulty.
For these non-linear sales cycles, it can be important to talk to your clients, their sales teams, or analyze their database to understand how prospects move their way from first interaction to sale. B2B software, for example, might include any number of decision makers, each with their own interaction flows with the website and conversion performance. At times, these sales cycles can take months or even years to complete depending on the size of the purchase, so although the Google Analytics report shows quite a bit of data for the 90 days, it might only be scratching the surface for non-e-commerce business models. Be sure you’re checking in with these other knowledge centers before breaking your lists apart.
Guess and Check
Sometimes, it’s better to get something done than to have it done perfectly. If you’re not able to find any definitive answers in a reasonable amount of time through the two processes above, use your instincts and create buckets that make sense to you and test them.
You can then keep an eye on performance within the advertising platforms, or you could hop back into Google Analytics if you’re using AdWords and review the Interests and Remarketing report in the Display Targeting report under the AdWords Acquisition section.
Keep Audience Size in Mind
While building these lists, it’s important to think about your business model and how users interact with your business. But it’s also important to think about audience size as well.
Each time you look to segment your audience further, the actual number of users goes down. Each channel has a minimum threshold you must hit to be able to advertise to an audience, so be sure you’re keeping that in mind when you build these lists. Segmentation is wonderful for being able to optimize on a more granular level, but it becomes worthless if your audiences are too small to target in the first place.
Building Your Retargeting Lists
When it comes time to actually build the lists for this strategy, it’s actually fairly simple.
All we need to do is create a number of lists with differing cookie lengths. Here’s how you would do that in AdWords for one of the examples I gave above.
First is your list of 7 Days:
Then 14 days:
Then 30 days:
And finally, 60 days:
Yes, it’s seriously as simple as that for audience creation.
But utilizing those audiences takes a little ingenuity to execute.
Campaign Structure for a Retargeting Time Funnel
The last part you need to consider is your campaign structure. It’s one thing to build these lists, but it’s another to make sure they’re being used correctly in your accounts.
It’s important to consider the relationship between the audiences we’ve just created. They’re all effectively nesting audiences with smaller portions of the 60-day audience. If you were to show their relationship visually, it would look like this:
For our campaign structure, this means that if we were to simply add all of our audiences into an account without any exclusions, there are some people we would be targeting in multiple ad groups.
So we need to add in a number of exclusions to make sure we’re only targeting people in stages.
Effectively, we want to target people in this sequence:
To accomplish that, we need to exclude the smaller time ranges from the longer date range ad groups. Following our example, that means that the 60-day audience has the 30-day audience as a negative. This allows for that ad group to only target the users from day 31 to day 60.
For the 30 day audience, we’ll exclude the 14-day audience so that the 30-day ad group is technically targeting users only from day 15 to 30.
The same for the 14-day ad group. It will have the 7-day audience as an exclusion and target users only from day 8 to 14.
Lastly, the 7-day audience doesn’t need any exclusions since it’s the smallest group already.
If you need to see that laid out a bit differently, here’s how those exclusions should be made:
With these exclusions, we’ve now created the segmentation that we were aiming for in the section above. Each audience member works their way from one ad set to another as they continue to move through the time funnel after engaging with your business.
Upping your retargeting game doesn’t require you to reinvent the wheel. Even starting out with this fairly basic strategy can give you a leg up to better communicating with your audience based on where they’re at in the buyer cycle and allow you to optimize on a more granular level. Now that you have this funnel built, there are all kinds of fun ways you can leverage it…but that’s a post for another day.
Do you utilize time funnels like this in your retargeting strategies? How have you seen them work? What are your experiences with building a phased audience approach like this one? Share with us in the comments!