Michael Calleia is the Senior Director of Marketing at Voltari, a company that develops predictive analytical solutions for the mobile space. Voltari optimizes mobile advertising campaigns from the start, learns and improves automatically in real-time, and provides insights. Michael comes to Voltari with a background that encompasses advertising, publishing, branding and marketing on both the agency and client sides.
The media industry celebrates and thanks digital for getting us quicker on our feet. Digital always has allowed us to measure, track, learn, optimize, and improve upon our own best efforts, well beyond what traditional ever could. We’ve long appreciated this contrast, with the improved capacity for learning being key. But, considering the advancements within digital itself—the growth of programmatic and the move toward machine learning—even typical digital methods are starting to feel “traditional.” Thanks to more powerful computing systems and science, there is so much more you can learn from a predictive programmatic campaign than you can from a traditional digital campaign. But, what’s the difference?
Old Habits—Scrutinizing and Validating Publisher Placements
The life of a digital campaign used to go something like this: armed with client objectives, a media planner conducted classic media research to determine best publisher placements based on demographics and potential for reach; sites were chosen and IOs executed; trackable creative was trafficked; the campaign was monitored based on delivery of contracted impressions, click volume and click through rate.
On a monthly basis, sites were evaluated for fulfilling their contracted (IO) obligation and sites received more or less of the future budget, based on their click volume and rate performance. These “performance” conclusions were drawn in isolation—with little if any consideration for back-end performance metrics such as conversions and sales volume, no cross-referencing of data sources, and no real involvement of the creative team. The media placement was expected to carry the weight.
In this situation, what appeared to work was increased. What appeared to falter was essentially optimized off the buy. We relied on cookie-derived media stats, publisher data, and then appended some siloed back-end data to build our story. We learned little in the process, and the cycle continued.
Mobile Raises the Bar and Hobbles the Cookie
Looking toward present day, as Mobile and cross-platform digital have become more commonplace; we have started to realize the limitations of our former ways. This realization includes grasping the limits of our media campaign process and the cookie itself.
Targeting browser behavior is no longer considered a sufficient means to understand who our audience really is. Nor can we expect to understand how they spend their days across platforms and devices, or effectively target and optimize their engagement. In light of this realization, the industry has gradually started to deal with the relationship between placement, media, and creative. And we have starting using data to better optimize that relationship. We are getting better at this.
Even five years ago, we finally were starting to see full account or marketing teams sit down at the start of a campaign to plan media and creative, with data and consumer insights in mind. Getting beyond the usual figures and specs, they became accustomed to using consumer behavioral insights to map out how they were going to launch, test, analyze, and iterate. The imperative to operate and learn across platforms in order to optimize, ultimately involved more of the right people in the process. And, it is under the weight of this cross-platform imperative that the cookie started to show strain. We have known for some time that we needed more.
Many of the current systems and platforms available use data and technology to buy media more efficiently on RTB platforms, but they are still using old traditional methodology, just applied on an ad call by ad call basis, applying manually selected targeting filters evaluating the publication to optimize the bid price.
Predictive Programmatic—Where Strategy and Machine Unite
Enter programmatic and audience based planning and buying. The growth of programmatic technology, systems and inventory options has yielded a greater proving ground for audience science. As more and more media is bought and sold on this basis, the data science behind it has expanded and produced options for learning that are beyond what we could achieve even with traditional digital media.
Because we are able to develop personalization models and target across content, time of day, device, and location—and integrate multiple data sources to do so—true audience or people based targeting is the new norm. Now, we have a richer level of audience intelligence that can be used to greater and greater return. This moves us quickly beyond the realm of basic media research through which we simply validated our own target audience assumptions and media commitments—and into something far more useful and scalable. We can run self-optimizing campaigns based on who is engaging with our creative and see in real-time what this audience looks like. Targeting is now audience-centric, rather than publisher-centric, allowing for deep insights into who we is actually consuming our message.
When strategy folks, media and creative are all armed with models and deeper audience data, the entire team has a closer view of the real consumer and can tweak their part of the mix to drive performance. Budgets can be re-allocated, visuals and messaging modified, audience targets re-thought or expanded, based on what we discover in real time. Thanks to predictive programmatic, gone are the days of simply checking the box, validating the IO, or extrapolating behavior based on over-simplified publisher based categories. The learning is detailed and based on reality—allowing you to serve targeted ads to real people and truly scale your efforts, feeling way more confident in the process.