Why Cadent Is Embracing Automation and Digital Tools for Addressable TV

Addressable TV is rapidly evolving, and Cadent is evolving with it. Read a Q&A with Cadent VP of Platform Analytics, Rachel Herbstman.

How has the Cadent Platform Analytics team evolved in the past year?

Rachel: We know we need to stay up-to-date  to handle changes in the rapidly evolving media landscape. We’ve made key hires in analytics, focusing on building a team with varied backgrounds – both media and non-media. Our newest team members have data science and computer science backgrounds so pairing those skill sets with more traditional media backgrounds is beneficial as we all have different perspectives and offer different contexts. As an analytics team, we’ve increased our operational efficiencies and it’s fun to collaborate on different insights that some of us that have worked in media for years may not be accustomed to.

With varied knowledge bases and complementing roles, we’re able to surface the performance data quickly by means of standardized scripts which inform our software and analysis. This allows for automation, freeing up our time so we can interpret the results and offer actionable insights. Ultimately, our entire team, from analytics to sales, works together to do the heavy lifting to provide the agency or advertiser digestible and sound data from which they can understand the value of their media spend.

At the conclusion of campaign measurement, our goal is to ensure we’re using our campaign data in the most effective way. Our continued client retention allows us to have a deeper understanding of audiences and viewing environments, seasonality. With more data and campaign experience, we’re able to better understand expected outcomes and therefore plan smarter garnering stronger ROAS for each client.

Over the past year, what’s different? What have you learned and where are you going?

Rachel: Everything’s different… by the day! As the landscape evolves, we’re tasked with keeping up with the challenges and opportunities so we’ve actually transformed our organizational structure to make sure we’re as efficient and current as possible. Our software approach to automation allow us to quickly slice and dice the performance data by various dimensions – audience segments, viewing environments, creative lengths or even multiple dimensions like audience segment by viewing environment. 

There’s a ton to learn from every campaign, but more data does not always equate to more learnings.

For our team, it’s about understanding the overall context of the brand and campaign initiatives and then dissecting the data points that are most relevant. With every campaign, whether it is viewed as extremely successful or not, there are actionable takeaways so campaigns that are deemed less successful by the numbers may actually provide the most insightful learnings for the advertiser. In some cases, we may learn that the proposed target audience is not as responsive as a client initially thought, or we may see that a longer creative length initially thought to be the heavy lifter isn’t actually necessary to drive an efficient response rate. While we obviously always strive to achieve incremental results, sometimes it is the campaigns that prove the opposite that teach us the most and allow the clients to rethink their entire media strategy based on real data.

Back to what we learn, one of my favorite cuts of data is the frequency analysis. For most verticals, we’re able to see the optimal exposure levels against the target  group to drive conversion for each of those audience segments. Then, since we’re executing across various supply partners, we can layer on the complexity of different viewing environments. With that understanding, our strategic recommendation is stronger for the next campaign. Frequency cuts become a pertinent piece of information that allows for campaign optimization, whether that means revised targeting criteria, shifting audience budget allocations, revised unit length mix, etc. and ultimately, stronger ROAS. We’ve had returning clients that have seen ROAS ratios increase from $1: $1.15 to a third or fourth campaign equating to $1:$5 based on those data points.

We’ve also learned that it’s not just about lower funnel metrics that data-driven tactics are generally known for. We’ve also seen other areas of success beyond just closed-loop measurement. Since we all know that audiences are fragmenting, it’s becoming more difficult to find incremental audiences so using addressable as a tactic to complement a linear television campaign could show the value in precision targeting for incremental reach. For a recent campaign, we measured the incremental cost per reach point compared to a linear schedule based on efficiency and were able to understand the point at which reach plateaus. Based on that analysis, we can determine when it is the right time to turn on a targeted OTT and/or addressable approach to increase holistic campaign reach.

How are you thinking about the frequency needed to reach marketers’ goals?

Rachel: We’re continuing to learn and validate that not all impressions are created equal. Based on our data warehouse, we can analyze the data in different ways based on different dimensional cuts. We can look at different audiences based on viewing environment or a combination. For example: We just measured a campaign for a client looking at frequency by audience, and we saw a current customer needed less frequency to convert than a prospect. That current customer likely needed a quick reminder to purchase. We’ve seen other campaigns where linear environments require more exposure. But VOD, a more lean-in space, requires half the frequency to get the same outcome. As another example, for an auto campaign promoting a new vehicle model, we saw an opposite effect – a traditional linear viewer took less frequency to drive a conversion. This was likely because they were exposed in other media content versus a VOD-viewer that may be a light TV viewer or transformational viewer, therefore needing additional brand awareness and addressable exposures.

How are you using data science and predictive business outcomes?

From a more holistic approach, a key component of Cadent’s growth over the past year has been the internal alignment with different groups within our company – not only with our addressable sales and client service team but also data engineering, data science and product teams. Again, with different skill sets across the groups, we are able to leverage historical learnings across a wide swath of advertisers and campaigns to better inform scientific models and forecast particular business outcomes.

We believe advanced TV should embrace the best of digital marketing tools and analytics to deliver valuable insights to clients. We use data science to enable data-driven decisioning on premium addressable video content, including providing insights on business outcomes.

On a campaign, advertiser and category level, we want to make sure we’re using the data that we have in the smartest ways possible. We’re constantly aligning with data science to create a variety of different performance models that will help forecast expected business outcomes. There are definitely nuances to consider like various custom audience segments, campaign initiatives (for instance: conquesting competitive shoppers vs. increasing sales among current shoppers) but with every campaign and additional data points, our models are growing stronger.

How has your understanding of viewing environments evolved?

Rachel: Cadent provides agencies and advertisers an understanding of their audiences and an understanding of how those audiences are consuming content. Based on that content, we ask, how should we be targeting them? We work closely with the advertiser or agency to understand how their current targets are set up so we can use addressable media, whether that’s OTT, VOD or linear to act as a complement and not a standalone tactic.

While our ultimate goal is generally to precisely target the core audience and measure the outcome of the media, as viewing behavior becomes increasingly fragmented, it’s critical to understand the viewing environment in which the content (or ad) is consumed. And then within those various viewing environments, we can leverage historical data to understand the frequencies with strongest conversion rates.

How is Cadent working toward standardizing metrics across the marketplace?

Rachel: This is definitely still a challenge for the industry. As landscape fragmentation increases, there will be more hurdles for standardization. There are industry groups in deep discussions about how to create measurement and currency standards across the advanced TV landscape but as of now, we’re seeing more walled gardens and measurement difficulties. That said, having executed hundreds of addressable campaigns, Cadent has the data to navigate through some of these challenges. For example, at the simplest form – the definition of an impression varies across the landscape. One operator may count an impression if the ad airs for 6+ seconds, whereas another operator may count an impression after the first frame of the ad is served. And there may be three other definitions for various providers, vMVPDs and OTT partners. One of Cadent’s value props is understanding the value of each impression because they’re definitely not all created equal. Understanding the frequency analyses based on MVPD, viewing environment, and audience allows us to normalize the impression delivery and better understand the true value of that impression to influence conversion.

Read more about Cadent’s vision for the addressable marketplace.