Dan Jaye, CEO of Aqfer, looks at how TV viewership has soared amid quarantine and how the rise of automatic content recognition (ACR) technology in Smart TVs means that the marketers are now swimming in more viewership data than they ever expected.
TV viewership has soared as people continue to social distance, work remotely and keeps their families at home. Where TVs once sat off throughout the day, they’re now constantly turned on, tuned to cable news or their streaming platforms of choice. Media buyers are scrambling to adjust to this dramatic change and find their audiences, desperate for reliable and timely data and insights. One emerging source is automatic content recognition (ACR) data, which uses technology inside smart TVs to match pixels of what’s playing on the screen to a database of programming.
This is, of course, a game changer as it opens up the opportunity to see the entirety of the 12ft experience vs the 12 inch experience, even from previously opaque solutions like broadcast TV, DVRs, analog and others.
The value of this data is clear, as it reveals what a viewer watched and when, and it’s incredibly important as viewers turn to TV amid the pandemic, whether that’s broadcast, cable, premium, addressable set-top-box, or over-the-top (OTT) streaming options.
The trouble is that ACR data, in its raw, collected form, is dirty, and if it’s not properly cleaned and processed, it provides little value while also increasing the risk of violating privacy regulations.
This is a challenge even in normal times, but increased TV viewing and the ACR data that comes with it are creating more issues around time, storage, processing and resources to manage.
As TV viewership bifurcates, ACR lets advertisers finally measure how many people actually watched their ads, and if those ads reached the right audience.
It provides more accurate viewership data than panels based on small sample sizes. And, perhaps most intriguing, it represents one the best ways to measure streaming viewership, which is incredibly valuable at this moment.
Simple math says that more viewership equals more ACR data.
If there is a double-digit increase in streaming on platforms like Netflix and Amazon, it would stand to reason that there is a double digit increase in the amount of ACR data available about those platforms (as an aside, we are hearing from some agencies that the amount of ACR data has grown more than 60% in the past month).
Several vendors provide access to the data and insights, ranging from established players like Nielsen, to new startups, down to the TV manufacturers themselves.
All of these vendors, along with the agencies and brands making use of this data, are now standing in front of a firehose, trying to figure out what’s useful without getting overwhelmed by the deluge.
While modern advertising is based on data-derived audience targeting, the existing data processing infrastructure is not set up to handle tidal waves on new data.
Monthly cloud costs for processing data can be exorbitant even in the best of times.
To suddenly have to deal with an influx in data while also watching ad spend slow is a major hurdle for the organizations that are leveraging ACR, now forced to process and sift through more data amid potentially smaller revenue and smaller workforces.
Now, more than ever, the need to analyze more relevant data to gain insight is critical to brands planning their next step for consumer engagement.
These insights remain valuable for understanding viewership, making targeted media plans, and using cross-device tracking, but getting from raw data to those insights is much more complicated and costlier.
For example, two of the most compelling use cases for ACR have the potential to dramatically drive up the cost of conventional solutions: cross-device attribution and content attribution.
Matching television ads to second screen responses introduces a variety of challenges, both in terms of identity resolution and correlation, especially since second screen engagement is often far greater than traditional response rates.
Content attribution, or the ability to measure the influence of cross-promotion of shows/content, breaks many attribution solutions that are designed for matching exceedingly rare events to exposures and aren’t suited to solutions where the “outcomes” are comparable in scale to the promotions.
Time becomes a factor as well.
While ACR isn’t real-time data, it is faster than traditional TV measurement models, most of which are trailing indexes that are built on 30 days’ worth of data.
Models that look that far back aren’t necessarily helpful, because consumer viewing habits are changing so rapidly amid social distancing.
Those dramatic changes are forcing brands to react right now, but the amount of ACR flooding in could slow down the speed with which it becomes available.
Brands, agencies and vendors need to digest data sources in different ways than they did a month ago, in order to understand completely different behaviors of completely different audiences.
It remains to be seen if brands and agencies have the power to gather and crunch all of this data in a timely fashion.
Taking the time to clean, process and manage the data correctly is vital, because any misuse of ACR data, accidental or intentional, could lead to a violation of local privacy laws or regulations.
It’s critical to make sure that the data is gathered with a clear opt-in and complies with GDPR and CCPA. Amid the rush to build a new media plan, the last thing that advertisers want is a privacy violation that lingers long after their ads have run.
Social distancing has rapidly increased the pace of evolution for TV buying, and it may actually be outpacing the ability of the ad industry to process the change itself.
For all the promise that ACR holds, the challenge is that advertisers and agencies need to be able to handle the sudden massive influx of available data, manage it, and turn it into usable, targetable audience data that complies with local privacy regulations.
ACR data may help many advertisers navigate this uncertain time period, but it remains to be seen if the industry can put together the resources so that this data can be put to use.
