In late 2013, Google introduced cross-device conversions as part of the launch of Estimated Total Conversions. I’ve always thought it was a huge leap forward for the industry. These were conversions that were already happening, only we hadn’t been able to attribute them properly. Since the announcement, marketers could measure and act on all of those interactions that start as a click on one device or browser and then finish as a conversion on another (and it’ll be coming soon to apps as well).
Many advertisers have bought into cross-device conversions and seen great results (as research recently published over on the AdWords blog proves). We also understand that some folks are still skeptical about cross-device conversions.
In light of that, I wanted to talk about three things:
The model looks at your clicks/conversions from users who have previously signed into Google services. People aren’t always signed in to sites, especially across devices and browsers, so this calculation doesn’t account for all of the possible cross-device conversions you could have received. As a result, we extrapolate what the total conversions would have been if everyone were signed in, customized for some stuff that’s unique to you and your account (things like country, conversion type, date and device), and use that calculation to scale your conversions to an estimated total.
The key part of this is you’ll only see cross-device numbers if the data meets a strict and conservative 95% confidence interval. It’s a rigorous number to hit; the term “estimated” gives some the wrong impression.
Google actually uses the same confidence interval for cross-device conversions that the Department of Health recommends for public health. If a new drug shows promise during a clinical trial with a group of testers, the Department of Health wants to be really confident that the general population will have similar results. It’s the same deal with cross-device conversions: We’ll only extrapolate conversions if we’re really confident in presently observed behavior.
One quick note here: These estimates are based on a last ad click from AdWords model. Any traditional conversions that you’re already counting won’t appear as a cross-device conversion as well.
Here’s the takeaway: The estimates have a high confidence threshold to pass, and they aren’t duplicated within AdWords. If you don’t include cross-device conversions, you’re missing an integral integral insight into how your advertising dollars are working for you.
There are plenty of industries that see a lot of cross-device behavior. And remember, these stats aren’t conversions that we’re calculating out of thin air — they’re conversions that were already happening that you couldn’t attribute correctly. Understanding typical cross-device behavior enables you to see the value that you could be measuring yourself.
There are some clear industries that benefit from users’ cross-device behavior. These percentages are the uplift in reported conversions after including cross-device.
Look at the travel industry:
|Japan Display||Japan Search||U.S. Search||Great Britain Search|
Retail is similarly strong:
|Japan Display||Japan Search||U.S. Search||Great Britain Search|
I don’t want to get into all of the numbers (you can check out Inside AdWords for that), but they’re promising pretty much across the board. There are clearly trends across industries, networks and geographies, and by understanding those trends you can take advantage yourself.
For example, if you’re advertising in mobile-heavy markets like Japan I think there’s a lot of reason for optimism. You, too, could finally track the 16% increase of your conversions that you haven’t been able to see before. (And I should note that these don’t yet include in-app cross-device conversions, so the uplift is most likely higher.)
So what do you need to do to actually take advantage of these stats?
Although I think the mechanics behind cross-device can be a heady concept, the best practices that you can adopt to get more of these conversions are pretty straightforward.
1. Count as many conversions as possible.
Like I’ve said a couple of times already, these aren’t new conversions. They’re instead conversions that you didn’t know to attribute to AdWords spend. Because cross-device conversions are trying to take a wider look at performance, you’ll know more about user behavior (and ROAS) by tracking more conversions.
How can you do that? Longer conversion windows are a start. Cross-account conversion tracking can also be really insightful if you use multiple accounts to drive traffic to the same site. You’ll also want to steer clear of conditional firing on your conversion tags. While that route can provide you with more siloed numbers, cross-device behavior isn’t supposed to be siloed. The more conversions you capture, the better the system is at allocating credit where it’s due.
2. Create a system to report on cross-device activity.
Add the “est. cross-device conv.” column to your reports so you know what’s going on in your different campaigns. Also, I should note that you’ll want to do this at the campaign (not the account) level. Because we’re so strict on statistical significance of our estimates, some campaigns may not see any cross-device estimates at all. Focus on those campaigns that are seeing activity, and assume that those that don’t see any cross-device activity don’t have enough volume or confidence to provide stats.
You should also create custom columns that cater to what you care about most. Here’s one that I find particularly useful:
There are a couple of limitations that you should remember when reporting. First, we don’t report them on search partners. You may want to create custom columns for Google.com conversions if you’re bidding on search partners. Second, cross-device conversions won’t show up for goals imported from Google Analytics or imported conversions — those are the types of campaigns that won’t see any cross-device activity whatsoever, so be sure to exclude them from any analysis.
3. Get more aggressive where you’re seeing lots of cross-device stuff.
As we mentioned in the recent AdWords livestream, AdWords automated bidding will factor in cross-device conversions later in the year. In the meantime, you’ll want to adjust your base bids according to the desktop cross-device uplift you see, then take another step to adjust your mobile bid modifiers based on the performance ratio of mobile to desktop. Segment your reports by device, and if you see lots of mobile cross-device activity bid to reflect that.
Cross-device conversions are an important way to understand how your users navigate your site. And they’re not simply bonus conversions on top of your normal performance — they are your normal performance only you’re finally able to track it. This behavior is becoming more and more commonplace. It’s best to develop a cross-device strategy sooner rather than later, as your customers have already evolved.
The post What Google’s New Cross-Device Benchmarks Actually Mean For Advertisers appeared first on Search Engine Land.