Match Rates = The New Click Rate in a Mobile-First Age

by Chuck Moxley, CMO "audience targeting", cookies, "match rate", "mobile measurement"
Match Rates = The New Click Rate in a Mobile-First Age

In a recent Ad Exchanger article, author Chris O’Hara wrote about the challenge of good match rates: “The first, most important issue to solve before any ‘matching’ takes place is a vendor’s ability to match people to the devices and browsers associated with them,” and then to match up that user ID with location data – where they are, where they go.

That’s a pretty good summary.  And a pretty big challenge.  

The problem so far has been that most companies were attempting to achieve the Holy Grail – delivering your message to the right person in the right place at the right time on the right device – by using antiquated web techniques, such as cookies, and measuring by equally antiquated KPI’s such as click-through rates (CTRs).  What’s wrong with cookies and CTRs?

  • Cookies crumble in a mobile-first world, as eloquently explained in this Marketing Land article. Short story: matching users by their browser cookies was a barely adequate method in the desktop world, and effectively useless in a mobile device world.
  • Measuring by clicks (or taps on mobile devices) doesn’t identify resulting sales lift, nor does it reliably ensure that you have matched your message to the targeted audience.
And so, matching has become a critical foundational factor in mobile ad delivery, and even more critical if you’re doing a cross-platform ad campaign. A vendor’s ability to deliver a high percentage of matches (matching the target audience to devices) has become a key factor in a mobile vendor’s ability to provide cross-device identity mapping.


A Brief Primer on Match Rates
How well do you understand the concept of match rates? Here’s a primer on the key points and key methods.  (Or download a more detailed guide here)

Match rates are about the accuracy of identifying a person with their devices, both off-line and online devices, because only then can you be sure of delivering the right ad to the right person at the right time, and in the right place.

Marketers are often sitting on loads of first-party data, but that data only becomes valuable – actionable – when it can be matched to the right devices. Whether you want to deliver an ad to desktop or a mobile device, the trick is to know who your reaching and then tie it back to your first-party data. 

Then the trickier part is this: finding a vendor who can match your customer knowledge to their devices, as this single-slide aptly and overwhelmingly demonstrates. 
 
In the effort to narrow down the delivery of your ad to your target audience, there are primarily three methods used today to create this all-important match-up:
  • 1st Party Data (authenticated users)
  • Location signals
  • Probabilistic modeling techniques to narrow your target
Here’s how each method works, plus their strengths and weaknesses.

First party data is a precise silo of data, often acquired by user authentication, such as a login at Facebook, AOL, Google, etc.  it’s big advantage: There’s no guessing who the user is and which devices they use. But this method only works within the confines of the “walled garden” – those who self-identify, and thus scale is limited for everyone but Facebook and Google.

Using location signals to build audiences is a method vendors such as xAd or PlaceIQ use in which the device owner is not known but they can make educated guesses based on where the device is – like assuming the device owner is a mom because it’s at a school. Such inferences are inherently capable of inaccuracy because it is essentially guesswork – still, a useful tool when you have no other way to identify your target ad audience and usually without the scale limitations of Walled Gardens.

Probabilistic modeling, a method used by onboarders such as Neustar and Datalogix as well as cross-screen vendors such as TapAd and Drawbridge, achieves cross-device matching by algorithmically analyzing anonymous data to create statistically likely matches between devices and owners. But as the method’s name implies, it is inherently inexact, with the vendors themselves admitting up to 30% inaccuracy.

As O’Hara’s Ad Exchanger article implores: “Marketers should ask vendors how and what they are matching.” The comparative information above explains why that’s important. And it is also what makes 4INFO wholly unique – our unique, patented method gives marketers the accuracy of deterministic but without the reach limitations of Walled Gardens.  What’s more, because our method doesn’t involve cookies or email addresses, we are able to achieve an unparalleled match rate when connecting household purchase data to device owners. This is the primary reason why more than 200 national brands trust 4INFO for the mobile campaigns, including 6 of the top 10 retailers, 8 of the top 10 CPG companies, and all 5 of the largest auto manufacturers.
We provide extraordinary reach (95% of all US smartphone users, 300 million-plus devices, 165 million users) and then tie that to purchases and other key data (via deep partnerships with industry-leading data suppliers) to offer nearly unlimited options for targeting and measurement. We’re able to use this inter-mapping of devices, people, and data to deliver large scale campaigns to your target audience.  If this appeals to you, contact us for more info.


back