Programmatic I/O Takeaways: What’s old is new again…with a modern twist

October 21, 2019

The theme of last week’s Programmatic I/O conference in New York City was focused on the need for marketers and tech providers to collectively shift mindsets and adjust targeting and measurement strategies, largely due to the changes to data and privacy laws. (Pardon me, but duh!)

I often compare our industry to bell-bottom jeans – fundamentally at its core, not much has changed. But with each decade, there seems to be a seismic shift or two to help modernize and bring advertising into the new era, make it “cool” again and bespoke for the current generation; usually the result of tech and measurement evolutions.

Throughout the conference, there were healthy doses of discussions pertaining to data-informed SPO and the fragmentation of tech and talent. The resounding buzzwords during the event were antitrust, privacy, contextual, incrementality, enterprise solutions, cloud, brand safety, triopoly, machine learning, future-proofing and a new one — data clean room. Let’s dig into these a bit.

Day one of Programmatic I/O Day began by acknowledging there are more questions than answers for executing cross-device measurement while also ensuring legal compliance, particularly as states are still determining their focus and legislature. What was abundantly clear, however, is that in order to prepare for a world devoid of cookies in 2020, there are six clear shifts that will occur.

  1. From precise, 1:1 targeting → aggregated or proxy audiences. Identifying alternatives to user-level targeting was not only the literal topic of one of the sessions, but a theme throughout the two days. It is anticipated that we will see an increased adoption of machine learning and modeling investment as the method to better predict and anticipate customer behaviors. Targeting…check! But what about measurement? It didn’t appear that anyone is willing to put a stake in the ground here, and I get it. With all these new unified IDs, which are filled with rich, high-quality data, the complicated piece will be tying them together to understand overlap and frequency management. Cue…conversation about enterprise cloud solutions and CDP adoption.
  2. From behavioral → contextual. I heard someone say “not your momma’s contextual,” and I think that rings true here. The modern twist is that we have more data than ever coupled with more sophisticated technology that’s often underpinned with machine learning. This enables us to go from what was historically based on keywords and semantics from one source toward using multiple signals and sources to more accurately predict and determine intent, sentiment and emotion. In short, in a cookieless world, we can still deliver an optimal experience to an end user with near precision.
  3. Programmatic shifts from open exchange → private. Bets were placed on shifts from open to private, and lower funnel to mid/upper. eMarketer even shared that it’s estimating 2020 will yield a tipping point in which private deals will surpass open market spend. Private deals still leverage automation, but with greater control and less competition. One topic I was surprised didn’t actually get touched upon was how complex it can be to manage and optimize these types of deals in aggregate.
  4. From MTA → …? While there are certainly unknowns as it pertains to how cross-device/domains measurement will net out (unified IDs…check! But how do we recognize the same ID across devices? Feels like a Jenga of sorts, no? TBD) however, there are some clear steps for a good-enough solution. As we still await future legislation, investing in enterprise cloud technology to unify your taxonomy and truly centralize your data, will all be critical. I observed that Data clean rooms had a healthy dose of introduction this past week along with chatter around declining use cases for DMPs alongside the rise of use cases for CDPs.
  5. From fragmentation and silos → consolidation There are two lanes here.
    1. With myriad products and solutions available, there is something for everything and everyone. The challenge is that the majority offer point solutions. While this is great to solve specific and/or individual use cases, these solutions often require another tech partner to help string them together in order to provide a unified and more holistic solve, which is why consolidation is always being referenced as a need.
    2. The more nuanced fragmentation topic, which usually gets discussed behind the curtain but received a bit more spotlight at Programmatic I/O, was talent and organizational structure. As the industry has evolved, the need for specialized talent has greatly increased. This has led to new teams being created but has also led to some fragmentation, including myriad agencies solving business objectives and different buying teams by channel. Just as our industry is undergoing partner consolidation, there is an apparent need to look at talent consolidation and evolution to better support customers and clients.
  6. From checking all the supplier boxes → SPO. This is critical to help not only identify and reduce tech taxes, but to better understand what you’re truly buying, how it impacts your goals and ultimately, how to leverage it to effectively optimize your campaigns. It’s no longer acceptable to simply check or uncheck the supply boxes in a DSP. Ensuring buyers have a complete understanding of these changes and implications should be table stakes. Education is key here. Understanding the variables being passed (or those that will be passed) within the bid stream — such as ads.txt, sellers.json and RTB supply chain object — are all data points to leverage.

In short, bye bye cookies, hello privacy. Digital marketers have a lot of other great data, and the capabilities exist to create new data sources, too. So let’s focus on unifying and organizing it better so we can apply and capture it in a smarter, more effective manner.

Knowledge sharing has become more prominent, predicated on the fact that we’re all eagerly awaiting the state-level legislature, or to see if a coalition will be formed to have this remain on the national level.

Some are predicting a complete reversal of the marketing funnel; others just believe the two extreme ends may be cut off, leaving mid-funnel strategies and tactics at the core.

Personally, I was a little surprised that more emphasis wasn’t placed on the measurement we do see in mobile app and CTV, and are starting to learn with Programmatic Audio. Those have always been cookie-less environments with plenty of investment and learnings from which to geek out on and learn from.

I was also surprised that the future-looking sentiment still seems a bit fear-based rather than putting the customer experience and satisfaction back at the forefront. It’s an opportunity we can lean into it; a shift from profiling a person toward profiling an experience.

However, there was (finally!) a collective understanding that investing in enterprise solutions to help manage data better than ever before will be table stakes. [Cue Google’s Ads Data Hub solution….] Machine learning and modeling also offers the ability to learn from data we’ve already collected to help predict future behaviors as we move away from hyper-precision buys toward a more aggregated application.

Ultimately, just like you plan, activate and measure a campaign and use those learnings to iterate, we can apply the same methodology to 2020 and beyond. We can learn from the past to better predict the future, and keep an open mind for a degree of additional change we may not see coming.