"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Peter Vandre, senior vice president and digital analytics practice leader at Merkle.
I have a confession. As a marketing analytics professional, I’ve bashed the likes of Facebook and Google for creating walled gardens, which bring all kinds of problems for cross-channel attribution modeling.
But here is where I owe Facebook an apology. With its push toward people-based targeting and measurement, Facebook has opened a seam in the publisher landscape that I believe will ultimately lead to a new type of digital analytics – one that looks much more like version 2.0 of direct marketing analytics.
This appeals to me because, at heart, I’m really a direct marketing analyst – you know, traditional DM and email campaigns. I largely abandoned the trade several years ago to ride the digital analytics wave. Like others, I traded in my individual-level logistic response models, dedicated 800 numbers and rigorous hold-out testing for the much more squishy, cookie-based DSP optimization, clone modeling and placebo testing. For digital analysts, learning can now happen much faster, and we have a load of rich contextual data that we never had before.
But there’s a trade-off: much less precious measurement accuracy.
Although it’s still not perfect, this approach at least gives us a fighting chance at reducing spending waste due to ad fraud, accurately attributing to mobile advertising, figuring out incremental campaign impact and, ultimately, giving digital marketers the evidence they need to claim their fair share of the overall marketing budget. As a direct-marketer-turned-digital-marketer, this has me dusting off my fancy sample size estimator and refreshing my direct marketing modeling techniques.
So, what will this new age of digital analytics look like? I see some big things on the horizon over the next 12 to 18 months:
Better Cross-Channel Measurement
When you start the targeting process with individuals whose data includes PII, you gain the ability to go beyond tracking just online conversions to also tracking offline interactions. This is possible through matching targeted individuals back to brick-and-mortar store sales, call center transactions and B2B sales. Finally, marketers will be able to quantify how many of the contacted individuals conducted research online and completed transactions offline.
Real Incremental Testing
Once we can determine which individuals we want to talk to, we can create a random group from the same population to hold out from marketing communications as a control group. The result is real incremental measurement and real marketing program ROI.
This has plagued digital marketers from the beginning, in spite of millions of dollars in industry investment on advanced attribution. Even though people-based marketing will take a while to reach scale, I expect to see it used sooner as another proof point for broader digital marketing ROI.
DMP: A Bridge Between People-Based And Cookie-Based Targeting
Although, for privacy reasons, PII is disassociated from the data-management platform (DMP), data onboarders, such as recently acquired Datalogix and Liveramp, make it possible to connect first-party data to cookies.
I’m seeing the DMP, which grew up primarily as a third-party data audience targeting tool, being used more to activate first-party data to inform media targeting and inbound website experiences. Digital analysts are left to sort out audience measurement in this limbo state between cookie and people data, and DMP providers are racing to enable better people-based tracking and targeting tools.
Traditional Campaign Management Tools Used For Tracking Digital Marketing
As people-based marketing grows up, marketers need the ability to keep track of each people-based outbound communication. So, along with direct mail, short message service and email promotion history, expect to add Facebook, Twitter, eBay and many other programs to an expanded contact history. Once a conversion occurs, we need to then connect this data to traditional cookie-based log file histories in privacy-friendly ways to enable a much richer view of how digital customer engagement leads to offline purchase. These same advancements will also lead to more accurate fractional attribution algorithms.
Although I’m not naïve enough to believe these changes will all sweep in overnight, I’m excited by the direction in which the industry seems to be headed. An evolution in people-based digital analytics feels long overdue.