This article originally appeared in StreetFight
By: James Moore
Let’s talk about burritos—specifically, selling them. Imagine a nationwide QSR wants to reach new customers. Its advertising partner suggests targeting “regular fast food customers” across the country. Sounds easy—maybe too easy.
Different cities have different burrito chains. Wouldn’t it be valuable to target folks who have visited one of the top competitors in the market, such as their local TexMex eatery? If the restaurant chain overlooks the nuances of local markets, it overlooks an opportunity to target more precisely, and with more personalized and meaningful creative. The proof is in the pico de gallo.
For a national brand, researching 50 different markets to create custom targeting strategies isn’t realistic. That is part of the appeal of structured data and prepackaged audience segments: You can target an audience quickly and at scale. However, in exchange for that speed and ease of use, you sacrifice precision.
Audience segments are built by grouping together a bunch of behaviors, demographics, and other variables and slapping on a label. These segments contain data that is good and bad, new and old, and precise and vague. You will learn something from serving large volumes of impressions, and your technology should optimize your campaign based on factors beyond the audience segment, such as the time of day and creative version. But you won’t be able to account for the differences of a burrito eater in Texas and a foodie in Maine.
With unstructured data, marketers can build custom segments from scratch by handpicking the relevant behaviors and variables—an alternative to buying off-the-shelf audience groups. Sometimes, the trade-off is scale, but it doesn’t have to be if you have the right strategy and partners. Plus, if you target effectively, you could drive performance with fewer impressions than you’d need if targeting more generally.
With the new wave of technology tools on the market, marketers can access detailed reporting to identify specific data points that are performing, or not performing, and then optimize their campaigns accordingly to improve results and curb wasted expenditure. The latest evolution in the use of unstructured data is real-time audience localization, or data signal identification, in which you use technology to identify signals and behaviors impacting performance in specific locales. Sometimes, these signals have no bearing on other markets, but by identifying them and using them when relevant, marketers can improve their advertising at the local level to drive overall performance. Here are three scenarios in which local nuances matter, inspired by actual campaigns.
1. What’s rock got to do with it?
A national real estate firm uses unstructured data to target audiences programmatically throughout its major U.S. markets. The technology identifies that in a particular city, people who search for the word “rock” convert at a high rate. It optimizes the campaign to serve more of these impressions. The performance is impressive—and confusing. Turns out, an independent real estate firm in that area called Rock Reality has been advertising heavily. The national advertiser didn’t know this, but the audience localization technology picked up on the trend where it mattered, while ignoring it in other markets where “rock” would have no bearing on performance, or even have a negative impact.
2. As different as a Tundra and a Prius
In the past, automotive advertisers commonly targeted a broad group of “auto intenders.” Now, more specific audience segments are available, like “Toyota intenders.” But that is not niche enough. A Toyota intender in Fort Worth, Texas is basically the opposite of a prospect in Palo Alto. In Texas, dealers mostly sell pickup trucks. Buyers care about towing capacity and tire quality. Serve an ad for a gas-guzzling pickup to a Toyota intender in Palo Alto, and see what happens. Nothing. The lots there are selling Priuses, and the buyers care about fuel efficiency.
Even markets that are similar have different dealership names, or even different legislation. By reflecting local nuances and buying behaviors in their targeting and creative strategies, auto advertisers will improve their campaigns.
3. You call that pizza?
The national media buyer for Domino’s probably knows what Pizza Hut is up to. Keeping tabs on leading local parlors around its roughly 2,800 U.S. locations is more challenging. Plus, there are regional trends in how people like their pizza. Certain cities prefer thin crust; others like deep-dish. Areas have popular toppings and prevailing flavor profiles, or favorite places to source local ingredients.
There is something else that varies from city to city—language. New Yorkers order a pie. In other places, pie’s a dessert. Or think of soft drinks (a coke, pop, or soda, depending on where you are) and sub sandwiches (hero, grinder, wedge, or hoagie, and that is a non-exhaustive list). These monikers encapsulate small regional differences that can have a big impact on your advertising’s relevance.
Painting with a broad bush has benefits, but it might lead to serving more impressions over a longer period than you would have if you had used a fine tip. By leveraging unstructured data for audience localization, marketers can target and optimize based on local nuances, and even extract learnings they might not have discovered otherwise. This can make a tremendous difference in advertising, whether a company is selling vehicles or burritos.