The Holy War Of Cross-Device Tracking

One of the biggest challenges for advertisers today is figuring out how to reach consumers across their multiple devices, from smartphones to tablets to personal laptops and work desktops.

The data required to understand consumers’ behaviors as they move across devices has been called the holy grail of data in digital advertising. But this singular vessel is comprised of two different types of data: probabilistic and deterministic. 

Have the challenges of reaching consumers across devices really been resolved -- or as we sometimes say in ad tech, have they been product marketized

Enter @AdTechPOTUS, a parody Twitter account that rephrases real tweets from Donald Trump to focus on ad-tech news, trends and often, industry issues. I started following the account after LUMA Partners' Terry Kawaja referenced a tweet in a recent presentation.

@AdTechPOTUS often says what we are all thinking in the industry:

     "Mobile device-graphs are imploding. It's a disaster and 2017 will be the worst year for probabilistic, by far! Telecoms will save adtech!"

As the collective IQ of the industry increases, it’s becoming apparent how flawed the guesswork of cross-device targeting can be. Without having a direct-to-consumer relationship or brand-owned data (data partnerships and co-ops need not apply), advertisers are holding themselves back from having any meaningful cross-device engagement with audiences.

To use one example of how cross-device matching can swerve out of control, an IP address is still a popular method used for geo-targeting and understanding households.

At best, you’ll be able to link a user to his or her exact home address. At worst, you have the Maxmind debacle, as Kashmir Hill reported in Fusion, when a remote farm in Potwin, Kansas -- the centroid of the U.S. -- was mapped to more than 600 million IP addresses by default when the company couldn’t find their exact location.

In addition to these bloated numbers, there are also looming privacy concerns being raised by consumer groups and the Federal Communications Commission about tracking all the data points necessary to ensure that probabilistic methods are relatively accurate.

@AdTechPOTUS swoops in again for a reality check:

  "The so-called probabilistic 'device-graph' vendors selling tech into agencies, are actually ad nets with FLAWED applied data science! Sad!"

When you target a consumer using probabilistic data, the hope is that you’ve probably reached that right person. In effect, it’s a gamble.

Today’s average consumers aren't tethered to a desktop computer when online, and they don’t shop exclusively in-store or even on one device. That's why there's a real need in our industry to be able to reach consumers on whatever device they prefer at the time, and then continue the conversation from one device to the next flawlessly.

Here’s one final shout out from @AdTechPOUTUS on the value of deterministic:

   "Device-Graphs that ingest deterministic data are HUGE! But those ad tech pundits saying COOKIES are dead! Lies!"

There’s an attempt to balance scale with precision under the probabilistic methodology. Powered by anonymous data points, and cookies, which are not particularly accurate over time, this method can be useful to reach consumers -- but only for a moment.

In simple terms, when you target a consumer using deterministic data, you’ve determined exactly who that person is, using first-party data.  It’s not a guess, and it’s not temporary. There is some debate about whether there is enough scaled deterministic data for marketers to act on, but the reality is that there are options out there and more will emerge.

The impact that first-party, deterministic data can have on campaign performance is indisputable. And advertisers are only granted this data by having direct and trusted relationships with their users. If you want your campaigns to make impact, start there -- not by guessing.

Next story loading loading..