Programmatic TV is new and evolving quickly. While recognizing the traditional model of television has served audiences well for more than half a century, times are changing and automation is the key to better targeting and scale.
It’s all about audience modeling and the improved automation to make programmatic TV possible. Combining audience modeling with automation yields better analytics and more targeted, relevant messaging. Customers gain improved efficiencies and TV media owners can monetize under-utilized inventory. The current, highly manual processes necessitate automation.
There is a common misperception that TV is an outdated system and is in no rush to change. That’s an over-simplification. TV advertising has evolved over decades and continues to adapt to emerging technologies. TV relies on the supply of content and distribution channels. The challenge of programmatic TV advertising is to meet the industry where it is and improve ad serving and buying within its current structure while precipitating the next level of automation.
Audience-based buying is not new. TV ad buys are always directed to an audience that is expected to find the message relevant. But there’s so much more content now that the old, personal touch simply can’t keep up, much less realize the potential. This new world requires precise delivery into selected spots within a schedule and to specific geographic areas to match the viewing habits and purchasing tendencies of that audience. This results in insertion orders for dozens of networks, within hundreds of ad zones distributed nationally and generates thousands of line items to be scheduled, trafficked, delivered, tracked, reported, rated and billed. This requires an enormous amount of coordination. Automating this process will yield significant benefits to both media owners and marketers.
What Happens Today?
Let’s look at the nature of the current delivery infrastructure to consider ways we can advance TV automation:
i. Plan: Media planners use a variety of research sources to translate a campaign brief (product, audience, budget and timing) into a set of campaign goals and a desired schedule of programs or network/day part combinations. This step is largely manual.
ii. Buy: Media buyers place and negotiate insertion orders with one or more suppliers, often with audience expectations based upon Nielsen Gross Ratings Points (GRPs), which are two-dimensional (age vs. gender) audience estimates projected from limited samples of historical audiences for programs. These expectations grossly generalize the original audience targets (“Males, 35-54,” instead of “Married men with a college education, an income greater than $100K, and in the market for a new car”).
iii. Traffic: Supplier systems map insertion orders to available inventory, producing detailed insertion schedules. Thousands of network/delivery zone combinations are each scheduled independently based on loosely defined formats for local ad placement opportunities. At the end of the day (quite literally) a schedule that defines exactly what ads to insert in each pre-supposed ad break (plus or minutes a few minutes) is created.
iv. Deliver: Supplier automated insertion systems deliver ads based upon traffic schedules. The traffic system doesn’t have any of the necessary targeting information to resolve scheduling conflicts or to correct delivery problems; these issues have to be addressed with manual intervention.
v. Report: Well after the campaign has run, each Supplier provides an affidavit listing all the insertions that occurred and the associated audience measurement. The affidavit is typically formatted as a human-readable spreadsheet.
vi. Reconcile: Media buyers import the affidavits into their agency systems and reconcile them with the original orders.
This labor-intensive system has been in place for decades. It is difficult to measure and control. It seems to work, however, for the networks that are driving the ad revenue. It can be improved without throwing out what works. One fear of automation is that the networks that pay for the content will not be able to get a return on their investment. The goal of automation, however, is to let the evolution of that new capability increase support for the present model, and eventually meet the challenges at all levels.
How Do We Get There?
Others in the industry have rightly pointed out that audience modeling is an essential piece of the programmatic TV puzzle. With such modeling, the planning and buying stages can be merged: marketers can just buy the audiences they want. Audience modeling also allows the reporting and reconciliation steps to be combined and enhanced, since the units delivered (impressions) will correspond directly with what was purchased (impressions).
An automated process allows campaigns to be dynamically managed while in flight to meet marketer expectations. While it is true that supplier systems vary widely, making automation a challenge, such automation is not impossible (it just looks that way if you’re coming from the digital side, with no exposure to TV technologies). Once automation is in place, leveraging the resulting data into actionable analytics becomes a straightforward proposition.
Automate the Suppliers
Automation isn’t something new to media owners and distributors—they’ve been using sales, traffic and insertion systems for decades. And there are a wealth of standards available that would facilitate the level of automation needed for programmatic TV: SMPTE BXF, SCTE-118, SCTE-30/35, SCTE-130, etc. The real challenge lies in wrapping all of the supplier-specific interpretations and implementations in an abstraction that will support a consistent set of interactions across suppliers. There are several companies working to meet this challenge; clypd is, of course, one of them.
Automate the Buyers
Buy-side automation essentially boils down to developing a standard for programmatic transactions (such as inventory availability, order negotiation and measurement reporting) and implementing the standard. The TVonTap specification is the most relevant and current attempt to do just that.
Of course, all of this forward progress is dependent on support for cross-industry standardization such as EIDR, Ad-ID, a common service for measuring audience and other initiatives that allow for streamlined communication and workflow. So, let’s work together to come up with the standards to guide the emerging field of programmatic TV and use all the data to improve delivery, service and monetization for all parties.