Approximately $70 billion is forecasted for spend on TV advertising in 2019, and advertisers are rightly demanding advanced measurement of their investments to understand exactly how TV strategies are impacting their business.
However, one of the biggest measurement challenges has been quantifying the true incremental lift of TV advertising on real-world outcomes such as website visits, online purchases, or location visits. Unlike digital advertising where randomized or targeted control groups are possible, the primary roadblock in measuring linear TV advertising was the complexity involved in isolating the correct unexposed control groups for TV audiences. TV ad evaluation requires alternative methods to be able to conclude causal effects.
Existing methods, which simply compare unexposed TV households to exposed TV households, cannot support causality because one cannot isolate the effects of TV ad exposure; the differences between exposed and unexposed could be due to other factors, such as audience differences and tune-in propensity. This means that advertisers are making strategic decisions based on incomplete data and, as a result, building waste into their TV media plans.
But with the evolution of Automatic Content Recognition technology and access to large, diverse datasets, Samba’s Causal TV Attribution using synthetic control groups can now be applied to TV measurement in order to understand the causality between TV commercial exposure and incremental lift.
Being able to separate correlation and causality is critical to TV performance measurement. Budgets should be allocated where TV advertising is truly moving the needle—not where it might be moving the needle—on incremental lift.
The foundation of measuring Causal TV Attribution is the application of synthetic control groups built on the diversity of Samba TV’s proprietary dataset and expertise of our data science team. Samba’s dataset is powered by the industry’s largest and most-representative multi-source TV dataset from fourteen global smart TV manufacturers.
Having access to a massive and diverse TV dataset allowed Samba’s data science team to create a proprietary technology designed to identify households that were not exposed to a specific TV commercial, but that look and behave very similar to households that were exposed—a control group for TV commercials.
It’s imperative that households included in a synthetic control group are those already targeted by the TV media plan (same network, timeframe, and DMA) to ensure the highest degree of measurement accuracy. This is attainable only because of the diversity, scale and proprietary data science of Samba TV’s dataset.
So what does this mean for marketers? Measuring the causal effects of TV, rather than simply measuring conversion rates, ensures that advertisers have the most accurate and complete information on where their TV investments drive incremental lift. This helps advertisers spend wisely, instead of wasting TV ad spend on audiences that were likely already going to convert without ever seeing the ad.
This methodology can also inform current and future TV strategies. Samba TV’s Measurement Sciences team has developed a methodology to generate insights for which networks, dayparts, and the days of the week impact incremental lift within a TV media plan.
TV Network 1 has the highest conversion rate, but the lowest incremental lift. Causal TV Attribution shows that TV Network 3 actually has the high incremental lift after viewers have seen an ad.
Additionally, Samba’s Causal TV Attribution should be used in tandem with digital campaign measurement—which has long used control groups as a standard—in order to better understand which components of your campaign are driving the best performance.
With $70 billion on the table, TV advertisers should leverage the most sophisticated measurement capabilities available to ensure they are making informed media decisions based on accurate data and insights.
Causal TV Attribution is available now from Samba TV for all national campaigns.