Programmatic: Moving Beyond The Tipping Point

“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 Dina Zelikson, Digital Program Manager at Bluestem Brands.

The word “programmatic” is a relatively recent entry to the ad industry vernacular, quickly joining buzzwords like “big data,” “data-driven marketing” and myriad other terms that are, likewise, shrouded in ambiguity.

Despite its recent popularity, programmatic media has been around since the emergence of RTB. Its scope, however, is expanding beyond the realm of exchange-based display. There’s a growing sense that, however ultimately defined, programmatic is going to be huge.

In its most basic sense, “programmatic” refers to anything that is automated, with decision-making via the artificial intelligence of machines. This descriptor can be applied to multiple subsets of the ad industry, including media buying, ad monetization, content optimization, etc.

The term “programmatic marketing” encompasses all of these subsets. It leverages data and technology to deliver real-time, personalized brand interaction across paid and owned media channels. It’s the intentional intersection of media tactics, user relevance and the dynamic delivery of advertising across digital platforms. Thus, it requires a high degree of integration across historically disparate platforms. You can think of the pathways to this integration as pipes connecting previously manual and disjointed processes, enabling their automation.

On the supply side, programmatic marketing is linked to ad monetization and ad serving, driven by the goal of maximizing yield. The retiring of the static daisy chain has propelled publishers and the ad industry into the nascent phase of RTB and programmatic buying. The challenges of this early phase are all too familiar: publisher frustration with decreasing CPMs, paralleled by advertiser frustration with viewability and ad clutter.

As RTB-enabled technology developed, marketers began to ask for increasingly complex data ingestion and activation. They wanted to use impression-level data, third-party audience attributes and first-party attributes together to inform marketing strategy, and also to activate that data without being restricted to inventory sources.

Digital marketing no longer lives in an age of head-to-head, demand-side platform tests. It has, instead, arrived at an age in which marketers ask these questions:  How can you help me leverage my first-party data? How can you help me extend my proprietary audience segments into multiple digital channels while maintaining a unified brand experience? And the most troublesome question: How can you help me leverage these segments across the guaranteed “premium” inventory space?

New Attribution Model Needed

The advent of Data Management Platforms marks a departure point in programmatic marketing because it brings its full potential – for both the supply and demand sides – into focus. DMP technology makes it possible for both sides of the industry to benefit from audience-activation techniques that employ first- and third-party data.  Both sides, however, face obstacles to unlocking the full benefits of programmatic.

On the marketer side, one of the biggest challenges lies in attribution. Those who stick to last-click or even first-click attribution while engaging in programmatic marketing are essentially learning to swim with anchors strapped around their necks. The situation evokes the dystopian society from Kurt Vonnegut Jr.’s short story “Harrison Bergeron,” in which everyone is required to take on an artificial handicap to preserve equality. Maintaining legacy attribution models that favor traditional and click-based media in today’s programmatic marketing world is, arguably, equally misguided.

The problem is that the industry hasn’t reached a consensus on a clear attribution alternative. Changing an attribution model is a huge undertaking. It is particularly daunting to those who mistakenly believe that programmatic marketing only includes niche, display-based advertising, thus missing programmatic’s potential to engage users across multiple brand touch points, including email, social and on-site.

Meanwhile, this misperception increases the challenge for the rest of us. Before we can agree on a meaningful attribution model, we’ll need to first convince our peers that programmatic marketing is viable, relevant and here to stay.

That said, I do think it’s worthwhile to consider attribution models that  accommodate the programmatic approach and lay a strong foundation for long-term objectives. Ideally, in programmatic marketing, paid and owned media tactics play off each other so that the combination and weight of various media types change, depending on the advertiser’s specific stage in its customer conversion funnel. Thus, attribution should factor in the concept of view-through, multi-touch modeling.

My vision for an attribution model alternative is one that lets me decide how to effectively balance my paid and owned media, prioritizing owned media when it makes sense. The next phase of programmatic marketing will lead to increased CPMs but should also give marketers the tools to divert advertising spending to where it’s most effective.

Next Steps

Once attribution changes are put in place, existing planning and forecasting practices must be adjusted. In companies with traditional attribution models, media planning is conducted by channel, in silos. New attribution models will come with new operational challenges, which will require internal team restructuring.

That’s not all. For programmatic marketing to become scalable across channels, some platforms in the LUMAscape will need to be seamlessly integrated. To accomplish this, both the supply and demand sides of the industry need to take action.

Marketers need to centralize disparate database records and establish pipes that feed into various systems of data activation, including site-content optimization, media buying, email delivery, social media platforms, and more. API technology enables this integration, but coming up with the time and resources to get it done remains a challenge.

In the meantime, the supply side needs to increasingly automate its direct sales while also finding a way to increase its revenue from RTB. Advertisers want to buy audiences that are infused with their own data and would be willing to pay higher CPMs to reach those segments. At the moment, however, RTB inventory isn’t set up to compete against directly sold media on the basis of bid price. To achieve the full integration of publishers’ monetization stacks, publisher ad-serving platforms need to be linked to sell-side platforms. The idea is to centralize the ad-serving decision process, a concept referred to as holistic ad serving. Until widespread centralization occurs, publishers will struggle to grow significant revenue from RTB, while advertisers will struggle to find scale.

Last but not least is the challenge of bringing programmatic concepts to directly sold inventory, which still comprises roughly 80% of digital media. A number of companies including Centro and NextMark have begun to address this challenge, and innovation in this area continues to progress.

The potential of programmatic marketing is huge, and all parts of the ad industry will ultimately feel its impact. The next year will bring acceleration to the already dizzying pace of technological innovation and system integration. The rate of adoption will depend on the extent to which the supply and demand sides of the industry can solve their many challenges. And to sustain a steady pace of evolution, both sides will need to make this progress in a synchronized and coordinated way.

Follow Dina Zelikson (@DZelikson) and AdExchanger (@adexchanger) on Twitter.


Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Great article, Dina, you cover a lot of ground very clearly and cogently. Amen to your points about bringing automation to direct sales, fixing attribution, and taking advantage of marketer data. You could easily write three more posts expanding on those topics, you have a clear vision that I think many would find valuable.


  2. Dina –

    A great, well-written post. I wanted to comment on your statement, “My vision for an attribution model alternative is one that lets me decide how to effectively balance my paid and owned media, prioritizing owned media when it makes sense.” The great news is that your vision is now a reality. Companies such as the one I work for, Adometry, as well as others or now providing this capability for hundreds of brands at scale.

    This approach to “advanced” attribution accomplishes the task of collecting and consolidating all the marketing touch and response data across channels and dynamically models the results to create a “true” result for each marketing investment at the impression, creative, and channel level. The good news is you can now more effectively balance your priorities across channels, both paid and owned.

    In addition, we are now feeding these attribution results via APIs to programmatic buying platforms. You are correct in these integrations can greatly enhance the efficacy of these media buys by using more accurate performance data.

    It is exciting to be a part of and helping drive these types of innovations which makes marketing even more intelligent.