Despite all the excitement around streaming platforms, traditional broadcast television still attracts the largest share of advertising spending. This year, companies are expected to spend about $139 billion on “linear” TV—where viewers watch scheduled programmes—compared with roughly $33 billion on streaming or connected TV. Even in a rapidly changing media landscape, broadcast television continues to play a dominant role in how brands reach large audiences.

Yet, after decades of television advertising, measuring its true impact remains surprisingly difficult. Systems like Nielsen track both viewing and purchasing behaviour, but only for a relatively small sample of around 42,000 households. This limited scope makes it challenging for advertisers to understand whether their campaigns are actually driving sales. As a result, many decisions about TV advertising are still based on estimates rather than precise evidence.

This uncertainty echoes a well-known quote often cited in marketing. As explained by Rex Du from the University of Texas at Austin, retailer John Wanamaker once remarked that half of his advertising budget was wasted—he did not know which half. In contrast, digital advertising offers clearer insights by linking clicks directly to online purchases. This ability to track results more accurately has encouraged many advertisers to shift more of their budgets towards digital platforms.

In recent research, Du and his colleagues developed a new way to measure the real impact of television advertising using digital data. Working with LG, they analysed viewing behaviour from millions of households that had agreed to share data through their smart TVs. Because these devices are connected to the internet, researchers could track exactly what people were watching and when, across broadcast networks such as NBC and ABC. However, the study did not include streaming services like Amazon or Hulu.

By linking TV exposure data with purchasing behaviour from a food delivery service, the researchers were able to study advertising effects in detail. They found that traditional methods had significantly overestimated the effectiveness of TV ads by as much as 55%. However, the analysis also revealed useful insights. Promotions such as coupons were highly effective, increasing the likelihood that first-time customers would make another purchase. Timing also mattered: viewers were more responsive to ads shortly after making a purchase, suggesting that well-timed advertising can reinforce consumer habits.

The type of programming viewers watched also influenced their behaviour. People who watched more news were less likely to order food delivery, possibly because they tend to be older or less engaged with such services. In contrast, frequent sports viewers—often younger audiences—were more likely to respond positively to ads. The study also showed that advertising is most effective when customers are already somewhat familiar with a brand. After two to four purchases, consumers are at a critical point where targeted ads can help turn occasional buyers into loyal customers.

Overall, this research demonstrates that better data can significantly improve how television advertising is measured and used. By connecting ad exposure directly to household purchasing behaviour, advertisers can make more informed decisions and reduce wasted spending. At the same time, viewers may benefit from more relevant and less repetitive advertising. As television continues to evolve, combining traditional media with digital insights may offer a more accurate and efficient way forward for both advertisers and audiences.

More information: Tsung-Yiou Hsieh et al, Leveraging Large-Scale Granular Single-Source Data for TV Advertising: An Identification Strategy, Marketing Science. DOI: 10.1287/mksc.2023.0582

Journal information: Marketing Science Provided by University of Texas at Austin