“Managing the Data" is a column about customer and audience data strategy written by longtime AdExchanger contributor Chris O'Hara.
The promise of “data is the new oil” has slowly come to fruition over the last five years.
Connected devices are producing data at a Moore’s-law-like rate, and companies are building the artificial intelligence systems to mine that data into fuel to power our ascension into a new paradigm we can’t yet understand. Whether you are in the Stephen Hawking camp (“The development of full artificial intelligence (AI) could spell the end of the human race”) or the Larry Page camp (“Artificial intelligence [is] the ultimate version of Google”), we can all agree that data is the currency in the AI future.
We are witnessing an incredible synthesis of fast-moving, data-driven advertising technology coming rapidly together with the slower – yet still data-driven – world of marketing technology. Gartner’s Marty Kihn thinks the only way these two worlds tie the knot for the long term will be with data management platforms (DMPs). I think he’s right, but I also think today’s DMP will evolve quickly as the data it manages grows and its applications evolve alongside it.
I think the most immediate changes in this ongoing evolution will be the ways in which data – the lifeblood of modern marketing – will be piped between data owners and those who want to use it. That’s because the way we have been doing it for the past 20 years in incredibly flawed, and second- and third-party data owners are getting the short end of the stick.
Unless you are Google, Facebook, Amazon or the US government, you will never have enough data as a marketer. Big CPG companies have collected data for years in the form of rewards programs and the like, but the tens or even hundreds of millions of addressable IDs they have gathered often pales in comparison to the billions of people who interact with their brands every day across the globe. To fill the gaps, they turn to second- and third-party data sources for segmentation, targeting and analytics.
That deal included offering their data as a mechanism of insights and discovery for marketers and agencies. Ad tech companies would showcase their data in various ways or use it as an input for lookalike modeling. In the end, data owner would infrequently be rewarded if the data found its way into a delivered advertising impression.
The real use of the data was sometimes unknown. Many cookies got hijacked for use in other sometimes competitive systems, and there was little transparency into what was happening with the underlying data asset. But, the checks still came every month. The approach worked when the best data owners – quality publishers – had a thriving direct sales channel.
The game has changed considerably. More than half of enterprise marketers own a DMP, and even smaller mid-market advertisers are starting to license data technology. Data is being valued as a true financial asset and differentiator. On the publisher's side, manual sales continue to plummet as programmatic evolves and header bidding supercharges the direct model with big data technology. Marketers need more quality data to feed the machines they are building to compete, and publishers are getting better and more granular control over their data.
More importantly, data owners are beginning to organize around a core principle: Any system that uses my data for insights that doesn’t result in a purchase of that data is theft.
Theft is a strong word but, if we truly value data and agree that it’s a big differentiator, it’s hard to argue with. For years, data owners have accepted a system that allowed wide access to their data for modeling and analytics in return for the occasional check. For every cookie targeted in programmatic that was activated to create revenue, a million more were churned to power analytics in another system.
From the data owner’s perspective: If you are going to use my data for analytics and activation but only pay me for activation, that’s going to be a problem.
To fix this, the systems of the future will need to help data owners provision their data in more granular ways. Data owners need complete control of the following:
How is the data being used? Is it for activation, lookalike modeling, analytics in a data warehouse, user matching, cross-device purposes or another use case? Data owners need to approve the exact modalities in which the data are leveraged by their partners.
What is the business model? Is this a trade deal, paid usage, fixed price or CPM? How long is the term – a single campaign or a year’s worth of modeling? Data owners should be able to set their own price directly with the buyer, with full transparency into all fees associated with piping data to a partner.
What is being shared? What attributes or traits are being shared? Is it just user IDs or IDs loaded with valuable attributes, such as a device graph that links individuals to all the devices they use? Data owners need to control data at the attribute level and decide how much of their data they are willing to share and at what price.
Outside of big data and blockchain conversations, the phrase “data provisioning” is rarely heard, but it’s about to be a big part of the advertising ecosystem. However, it is those very security concerns that have kept data sharing at scale from becoming a reality.
Data owners need complete control and transparency, and they must stay ahead of nuances in law, such as the new GDPR requirements. As ad tech and marketing tech continue to come together and systems evolve in parallel with their ability to make the best use of data, data security is needed before data innovation can truly happen.
Data may be the new oil, but will it be run by ad tech wildcatters, or will the rules be governed by data owners?