Machine Learning Transforms Local TV into a Modern Advertising Channel

Share this:

CTV is one of digital advertising’s hottest channels. Conversation on CTV tends to focus on the ability to pair TV’s emotional storytelling power with granular targeting

Kalyan Lanka, VP of product management at Ampersand, argues that the same pairing of creative and targeting potential is hitting local TV. Here’s how marketers can take advantage of the opportunity.

Why is local TV news inventory a good choice for marketers concerned with personalization and breaking through?

Local TV inventory in general provides a great opportunity for marketers who are trying to engage with the right consumer at the right moment and within a highly engaged context. It is in fact less about local TV news but more about local TV inventory. 

If aggregated properly, local TV inventory can show a combination of viewership insights to help marketers find their strategic audiences, delivering personalized messages across various DMAs, granular geographical regions, specific network dayparts, or premium programming.

How is machine learning revitalizing TV advertising?

One of the key ingredients of any machine learning model is the underlying data. While access to rich viewership insights was limited in the past, companies are now able to aggregate viewership insights across millions of workloads in a privacy-compliant manner to power unique capabilities by applying predictive models and machine learning techniques. 

The speed and scale of machine learning helps identify any inefficiencies in TV spend, ultimately maximizing the effectiveness of marketing investments.

How is machine learning making local TV advertising more effective?

One of the most important aspects of media buying is the ability to forecast available impressions and deliver against forecasts. In local TV, this becomes an extremely complex problem due to the number of markets, network/dayparts, and audience segments involved. Most importantly, lack of enough sample sizes from panel-based currencies at a local market level often results in inaccurate forecasts and several zeros. 

Machine learning has the ability to deliver forecasting models using viewership insights to predict available impressions against a strategic audience segment beyond age and gender. These models help identify the exact network/dayparts and weeks that maximize advertisers’ ability to reach their strategic audiences across screens. They are also designed to be adaptive, increasing in accuracy over time as they ingest the most current data.

I imagine many advertisers would fear local TV will be unmanageable due to the number of markets and a lack of technical sophistication. What do you say to advertisers with those concerns? 

While the number of markets and networks might seem daunting from the outside, the scale of local TV and its ability to reach desired consumers locally is powerful. It can be easy to navigate with the three of the right ingredients: scale of inventory, technology, and data.

Joe Zappa is the Managing Editor of Street Fight. He has spearheaded the newsroom's editorial operations since 2018. Joe is an ad/martech veteran who has covered the space since 2015. You can contact him at [email protected]