Traditional TV Ads Are Half as Effective as Believed

New research from the University of Notre Dame indicates that traditional television advertisements for food delivery services are significantly less effective than previously believed, with traditional measurement methods overestimating their impact by 55 percent.

Despite the growth of streaming services, linear television continues to command the majority of advertising budgets. In 2026, advertisers are spending $139 billion on linear ads, compared to $33 billion allocated to streaming and connected TV.

Challenges in Measuring Ad Effectiveness

Determining the actual return on investment for traditional TV ads has historically been difficult because there was no reliable way to track the individual behavior of viewers. Most measurement strategies relied on aggregate market data and ratings, which often conflated a viewer’s pre-existing habits with the actual effectiveness of an advertisement.

To address this, Shijie Lu, an associate professor of marketing at Notre Dame’s Mendoza College of Business, conducted a study published in Marketing Science. Lu and coauthors Tsung-Yiou Hsieh of Oklahoma State University and Rex Yuxing Du of the University of Texas at Austin utilized digital data to isolate the true impact of these advertisements.

“We show TV ads are only about half as effective as we thought. When corrected, the real sales impact is much lower, which has important implications for how advertisers evaluate performance and allocate spending.”

Shijie Lu

The Role of Real-Time Viewership Data

The research team employed a natural experiment approach by leveraging second-by-second household viewing data from LG smart TVs. The study analyzed millions of people who opted into sharing their viewing data over a four-month period, focusing specifically on broadcast networks such as ABC and NBC. The researchers did not track streaming applications like Amazon or Hulu.

By combining this viewing data with the actual food delivery app usage of those households, the researchers could determine if a specific ad led to a purchase. For example, if a household watched only a portion of a live game, they might see a food delivery ad if it aired during their viewing window but miss it if it aired earlier or later.

This timing difference allowed the team to separate the ad’s impact from other factors, such as whether a household was already more likely to order food regardless of the advertisement. Lu described the ability to link precise viewing data with real purchase history as a game-changer for measuring credibility in advertising.

Consumer Behavior and Targeting Insights

Beyond the overestimation of general effectiveness, the study revealed specific behavioral patterns regarding how consumers respond to food delivery promotions. The data suggests that brands are often wasting money on ineffective placements by failing to target based on purchase history.

The researchers identified several key trends in consumer responsiveness:

• Promotions specifically designed for first-time buyers were found to increase customer retention.
• The peak responsiveness to advertisements occurs within two days of a consumer purchasing food on a delivery app.
• The highest engagement rates were observed among customers who had previously ordered food two to four times.
• Younger, tech-savvy sports fans were identified as better prospects for these ads than older viewers of news programming.

Lu noted that traditional tracking methods fail because they do not fully account for who is more likely to see ads and who is more likely to buy.

Shift Toward Precision Targeting

The findings suggest a need for a shift in how companies allocate their advertising spend. Rather than relying on broad demographics such as age or gender, the study advocates for targeting based on actual purchase behavior.

By applying digital-level precision to traditional TV, the research provides a tool for more accurately measuring the return on investment. This approach allows advertisers to identify which specific customers are most likely to respond to a prompt based on their history and current habits, reducing the reliance on inflated aggregate data.