Data driven TV ad sales
The rise of the cord cutter has sold investors away from TV-related stocks, causing the media sector to lose billions in value. But is this perception accurate? According to a recent report released by Forrester, Young TV Cord-Nevers Have Arrived And Are Here To Stay, 76% of the US population still pay for TV while cord cutters have only risen a modest 6%. Forrester also predicts that number to stay below 15% through the end of this decade. However, cord cutters, especially millennials segmented into the 18-31 age range, represent a future audience that networks must pay attention to now as they plan for future media consumption behavior.[i] To do this, the media companies need to make TV as accountable as digital, and the enablement of data becomes the catalyst for this.
The placement of television ads using third-party and first-party data is limited mostly to addressable buys, which, although highly targetable, have limited scale. However, there is major industry movement in the ability to leverage first-party and third-party in large national buys, which make up the majority of television spend. For the first time, big data is coming to TV.
Historically, linear TV has been planned and bought in two ways – either Nielsen breaks are used or Nielsen data is fused to a vendor like MRI – and they both have limitations. If the buy only leverages Nielsen, the planner is limited in the type of audiences available. If fusion is used, sample size constraints present issues and there are methodological questions about fusion techniques. Compare this to an environment where a CPG advertiser could algorithmically optimize bidding strategies for moms in the market for diapers while they are at a specific retailer, based on a custom list of moms likely to switch to their brand. No wonder television doesn’t have the data-driven reputation of digital media. But this is all starting to change.
As set top box (STB) data has scaled and started to become accepted as a planning tool, new approaches are taking hold. STB data vendors can track several million households, as opposed to panel methods that track in the tens of thousands of households. Historically, this STB data was leveraged to create local ratings where the panel methods were not possible. However, the vendors are improving, collecting larger and cleaner data sets. With a sample of viewing behavior in the millions, several possibilities have opened. For example, it becomes possible to join third-party data to that household viewing behavior and create ratings where the third-party data is the common data set used in the advertisers’ industry.
Additionally, now advertisers can join their first-party lists from their CRM systems. For example, if a credit card issuer identifies a likely audience for its product (perhaps for sending direct mail), that credit card issuer can now see which television shows index higher against those same individuals and buy those shows. This is big step toward creating custom audience buys, which are now so common on Facebook. However, the network can also get in the game. By joining digital or CRM data into this process, networks can create audience extensions and look-alike models across their digital properties. Finally, advertisers, agencies, and publishers can better measure the impact of a campaign by loading post-campaign CRM data to see the lift in sales and leverage sophisticated attribution models that are common in digital.
Let’s take a hypothetical example. A luxury automotive maker wishes to migrate people from its entry-level car to a more expensive model. The automaker can combine its list of known entry-level owners with third-party data (such as financial data), model out likely buyers for the more expensive car, validate those models with actual sales data, plan TV buys around that audience, develop creative for that audience, and then see the impact in sales. It’s a new way to think about TV, but it has many similarities to digital. It’s reasonable to define programmatic linear TV as the programmatic infrastructure that will automate the buy flow process, data transfer, and analysis described above.
It’s worth sharing a couple of realities. First, the approaches described in this article are in their infancy. Second, these approaches can upend some business models, and present challenging questions. For advertisers, do media teams have access to CRM data? At agencies, can the agency fee structure support the extra work? At publishers, how do they load these buys into planning tools and respond to RFPs? All of these are still being worked out. Finally, how do you build these systems to respect the privacy of the viewer? It is critical that viewer data remains 100% anonymized and every approach I have seen does this. As systems evolve, privacy must remain at the center of each decision. Fortunately, privacy compliant approaches that support the business models of advertisers, agencies, and networks are a reality today.
Big data in television is good for everyone. Networks can better monetize their inventory and charge higher prices for better targeting. Advertisers get better targeting and thus less waste. Agencies demonstrate more value to their clients by providing data-driven strategy. Most importantly, it is good for viewers. Viewers will get more relevant ads and this data-driven approach will support the economics of the marketplace, allowing continued content creation and new innovative content that viewers crave.
[i] McQuivey, James, and Michelle Moorhead. "Young TV Cord-Nevers Have Arrived And Are Here To Stay." Www.Forrester.com. October 6, 2015. Accessed November 10, 2015. https://www.forrester.com/Young TV CordNevers Have Arrived And Are Here To Stay/fulltext/-/E-res121124.
Andy Fisher, Chief Analytics Officer at Merkle, Inc. and
Ryan Katz, Industry Executive, Travel, Media & Entertainment
Tel: 443-542-4000
Web: www.merkleinc.com
Email: marketing@merkleinc.com
Twitter: @merklecrm