Internet-connected televisions make it easier for marketers to measure broadcast ads against sales

Based on the research of Rex Du

Despite hype about streaming services, traditional broadcast television still dominates advertising dollars. This year, advertisers will spend $139 billion on “linear” TV — where viewers watch programs at scheduled times — compared with $33 billion on streaming or “connected” TV.

But 85 years into the broadcast era, it’s still hard to measure the return on all that spending: how effectively their ads get people to buy products or services. Nielsen tracks both viewing and purchasing for only 42,000 households.

“The cliché in marketing is a quote by Johnny Wanamaker,” says Rex Du, Shelby H. Carter Jr. and Patricia Carter Regents Professor in Global Business Marketing at the McCombs School of Business at The University of Texas at Austin. “He once famously said, ‘Half the money I spend on advertising is wasted; the trouble is, I don’t know which half.’”

Digital ads, by contrast, let an advertiser connect clicks directly to online purchases. That’s why more of ad budgets have moved to the internet, he says.

In new research, Du helps level the field for broadcast TV. He devises a more exact way to measure the causal impact of TV ad exposures on consumer purchases. Ironically, he does it with digital data.

Ads, IP Addresses, and Food

With Tsung-Yiou Hsieh of Oklahoma State University and Shijie Lu of the University of Notre Dame, Du used data provided by the smart TV company LG for their research. Because smart TVs were connected to the internet, LG had IP addresses for millions of households that had opted to share their TV viewing data.

The data showed what and when people were watching during four months on broadcast networks such as NBC and ABC. It didn’t cover what they saw on streaming services such as Amazon or Hulu.

LG matched IP addresses to exposures to programs and ads, second-by-second. It also linked them to day-by-day purchases from a food delivery service that was airing about 700 broadcast TV ads a week.

By using their new measurement method, Du discovered those ads were delivering less than marketers thought. Traditional measures overstated the effects of TV ads by 55%. He also gleaned several insights that could help the company better target its ads:

Promotions work. Promotions such as coupons aimed at first-time buyers increased their likelihood of buying again 58%.

Recency sells. Viewers responded more to ads during the first two days after making a purchase. Says Du: “The longer you wait after your purchase, the less likely you are going to buy again. Seeing another ad stops that recency trap.”

Programming matters. Heavier news watchers were less likely than average to order delivery, perhaps because they tend to be older and less tech-savvy. By contrast, heavy sports watchers, who tend to be younger, were more likely to order delivery.

Habits pay off. Viewers were most responsive to an ad when they had already ordered food from the company two to four times.

That point is a critical juncture, Du says. “They’ve made a couple of early purchases, and they’re on the fence. You need that ad to nudge the customer to become a loyal long-term customer sooner.”

Although the paper covers food delivery, the method can be applied broadly to brands, time periods, and markets. So says Hsieh, who originated the research when he was Du’s doctoral student.

It can benefit people on both sides of the screen, he adds. “For advertisers, better measurement helps them allocate budgets more efficiently and avoid waste. For viewers, more effective advertising can reduce the incentive to over-saturate programming with ads that are poorly targeted or less relevant.”

The research can help TV networks as well, says Du. “TV faces the measurement problem even more so than digital media. The new generation of data means they can connect individual households’ TV ad exposure to what they purchase.”

“Leveraging Large-Scale Granular Single-Source Data for TV Advertising: An Identification Strategy” is forthcoming in Marketing Science.

Story by Omar Gallaga