“TV Attribution.” The first time I heard the term, I was curious. Then, I thought about it, and dismissed it as a dumbed-down explanation of TV measurement for those digital-minded marketers who are used to thinking in terms of clicks. Then, I heard the term repeated by other players in the industry. Then, I had an influential industry analyst ask me about the term. That’s when I knew the term “TV attribution” had gone mainstream. "TV attribution" is the worst kind of buzzword; it is a term designed to over-simplify a complex business challenge so that it fits within the business model of attribution-centric service providers.
Let’s look at our old friend, Wikipedia, to define attribution:
Attribution is the process of identifying a set of user actions (“events” or "touchpoints") that contribute in some manner to a desired outcome, and then assigning a value to each of these events. Marketing attribution provides a level of understanding of what combination of events in what particular order influence individuals to engage in a desired behavior, typically referred to as a conversion.
What this means in a real-world example is: You saw an online ad for a cool handbag, you clicked on the ad, and navigated to the site selling the handbag. You purchased the handbag. The vendor’s ad network notes that you clicked on a specific ad to get to the site and ultimately made the purchase. Therefore, the sale is attributed to the ad that you saw that led you to make a purchase.
“Last click attribution” is attribution in its simplest form. You may see many advertisements, but the last click (be it search, banner ad, video or website) before a user makes a purchase is given credit for the purchase.
Digital attribution has come a long way since last-click. We now have cross-channel attribution where algorithmic and game theory methodologies assign conversion credit along the consumer touchpoints. More recently, integrated attribution and marketing mix modeling, or as it has been termed by Forrester Research - UMIA (Unified Marketing Impact Analytics), is regarded as the state of the art solution for marketing attribution. Done right, an integrated approach accounts for non-addressable media and holistically measures incrementality. A UMIA approach not only measures clicks, but also all other elements of the marketing mix including offline activities and promotions and mass marketing like TV, print, PR, outdoor, sponsorships, and word of mouth. A UMIA approach provides marketers with fuller picture of what elements of their full marketing spend is driving sales.
The distinctive feature of digital attribution is that it is calculated at the individual consumer level. Through cookies, widgets, and other online tracking tools, we know to a high degree of certainty what an individual was exposed to on the Web before making a purchase or transaction. In contrast, TV viewership is tracked at the household level. TV is a broadcast medium, so there is no mechanism (like a cookie) to track what advertisement an individual saw on TV. Furthermore, there is no way to tie specific sales to specific ad exposures. Therefore, there is a fundamental flaw with the concept of “TV attribution.” You cannot have attribution because TV does not (yet) provide a feedback mechanism to explicitly “attribute” a sale to specific advertising exposures.
The promoters of “TV attribution” build upon their fiction by claiming that TV advertising can be optimized through attribution the same way that digital marketing can be optimized through attribution. There are two fundamental flaws with this logic:
- TV just doesn’t work like digital: TV advertising has an impact on businesses in both the short and the long term. TV is frequently regarded as an upper-funnel activity which occurs earlier in the sales cycle. TV builds pricing power for brands; it drives word of mouth; it also drives folks to do research online. Measuring TV through its call-to-action is assuming all TV acts like direct response TV, and that simply isn’t the way it works.
- The data for “TV attribution” isn’t there – YET: The term “TV attribution” suggests that it tracks consumers along their journey. For instance, online you can track the individual consumer who searched for a hotel room in NYC, clicked on a banner ad, and followed up with a reservation. For TV, that data isn’t shared at the user level, and where it can be shared, it isn’t comprehensive enough to be representative. Furthermore, “TV attribution” is often calculated with “opportunity to see” data. “TV attribution” isn’t based on the household or consumer who actually saw the commercial, but the fact that the household or consumer had an opportunity to see the ad. Those are some big assumptions that make a mockery of the concept of attributing a sale to advertising views.
Promoting TV attribution as the functional equivalent of digital attribution does a real disservice to advertisers and to the industry. Typically, TV attribution is limited to analyzing website and search query activity after spots air. This may be useful in some industries (e.g. gaming or app downloads), but it is of very limited value in most industries. Think about the most widely viewed and highest priced TV advertising event in the US, the Superbowl, and think about your own behavior. Did you do a Google search or visit an advertiser website following the Superbowl based on viewing a Superbowl ad? Think about how you interact with brands after seeing a TV spot: How many TV commercials have motivated you to visit a website? TV attribution is a great concept, but it doesn’t look all that useful upon scrutiny. TV attribution, as it is practiced today, doesn’t provide the holistic picture or context that is required for true TV measurement, and therefore, it doesn’t live up to the name.
We are fortunate to be living at a time when the robustness and granularity of data is continuously improving to provide us with more opportunities to generate insights. As a marketing analytics professional focused on data-driven value creation, I believe it is irresponsible to term TV’s immediate impact on search and website activity as “attribution.” That only serves to confuse the industry and provides ammunition for the skeptics of data-driven insights and decision making.
While I am optimistic that the data will evolve and more analyses of the cascading and user-level impacts will be possible, the appropriate approach to capturing TV’s true value is an integrated and holistic measurement. TV advertising’s value is well beyond the attributed impacts on search and website visits, methodologies should also incorporate the direct, indirect and synergistic impacts as well as longer term branding effects like improved pricing power and brand equity.