Marketers are eager to get the attention of fast-growing US Hispanic audiences. But measurement can get fuzzy fast.
This measurement challenge also extends to other marginalized groups, but the Hispanic market is the most talked about by marketers because it’s the biggest audience, said Adriana Waterston, an EVP at custom primary market research company Horowitz Research.
In July 2021, the Census Bureau ballparked the US Hispanic and Latinx community at about 62.6 million, which comprises approximately 19% of the population, making it the second-largest racial or ethnic group in the US after non-Hispanic whites.
US Hispanics are also amassing economic clout. A September Latino Data Collaborative report found that US Latino consumer purchasing power grew to $3.4 trillion in 2021 from $2.8 trillion in 2020.
Despite this growth, the multicultural/Hispanic marketing budget for many companies has for decades been but a fraction of the budget for the general market, Waterston said. But now, as each generation becomes more diverse than the last, the writing is on the wall.
Companies are adapting their budgets and strategies to better understand and reach Hispanic audiences.
But although precisely measuring Hispanic audiences is in the best interest of marketers, there are many ways measurement can get messy, from the data sources used to quantify who is Hispanic to the shortcomings of measurement technology itself.
The limits of census data
Measurement companies, research companies and media buyers rely heavily on US census data to plan and strategize marketing efforts. One of the main reasons measurement gets murky is the fallibility of this data.
Even in the best circumstances, census estimates involve generalizing. But, in 2020, the last time a census form was sent out, that sample wasn’t even representative.
For starters, confusion swirled around how to respond to the survey’s race and ethnicity questions. A post-census report card revealed that Hispanics were undercounted by nearly 5%.
Adding fuel to the fire, the Trump administration led a concerted effort to quash the participation of immigrant communities in the 2020 census, leading to major data discrepancies, Waterston said. “I have zero confidence in the 2020 census data,” she said.
If the 2020 data can’t be trusted, the most reliable data comes from the previous decennial census in 2010. Although the census conducts the annual American Community Survey, it’s a much smaller body of work and now well over a decade old.
“If we’re not getting accurate census data and we’re not sampling these folks accurately, then it’s the fruit of the poisoned tree,” Waterston said.
Checking a box on a census form also doesn’t tell you anything about someone’s heritage or what they connect to culturally, Waterston said. A respondent raised by an Afro-Cuban father and a Puerto Rican mother of Taína and Spanish descent, for instance, might identify as Afro-Latina and feel strongly attached to Afro-Caribbean culture.
Yet another respondent with the same profile might identify very differently. And because identity is malleable, the same person might check different boxes on the census over time.
A matter of distinction
The Hispanic and Latinx market is characterized by differences upon differences – and capturing its complexity and diversity means accounting for linguistic and cultural nuances. For instance, not all US Hispanic and Latinx communities speak Spanish.
“Hispanic [households] speak three languages: Spanish, English and Spanglish,” said Ana Ruedaquintero, associate director of media strategy at digital marketing agency Wpromote.
But, she added, language shouldn’t be the “default attribute” to determine if an audience falls within a Hispanic segment. “It’s not necessarily about their country of origin or even their language,” Ruedaquintero said. “It’s more about cultural affinity.”
Roughly one-third of US Hispanics have lived in the country for multiple generations, which mirrors the general market in many ways, including not speaking any Spanish and preferring to see ads in English. Some households made up of recent and older immigrants, however, may be primarily Spanish-speaking.
But though the breakdown is different, the same problem exists, said Roberto Ruiz, EVP of research, insights and analytics at TelevisaUnivision.
“‘Hispanic’ is just too broad a term,” he said. Two people classified as Hispanic might have vastly different consumption habits.
Sometimes, even the same multigenerational household might contain people with varying levels of Spanish-language ability and affiliation with a Hispanic or Latinx identity. It’s a puzzle that TV measurement providers have yet to solve.
Not that the industry hasn’t tried.
For example, back in the day, under the Nielsen regime, a people meter device would monitor the TV viewing habits of household panel participants. Each household member (with their age and sex recorded) would press a viewing button to identify themselves when they watched content. Guests could also input their age, sex and viewing status through additional buttons.
In 2005, Nielsen rolled out active/passive meters that could measure viewership in digital environments. And, in 2007, the company introduced portable people meters, wearable devices that automatically pick up what respondents are watching.
Although this self-reporting system has its flaws, it makes it feasible to identify who viewers are, Ruiz said.
But with the rise of CTV, people might be watching on multiple devices, moving not only between their iPad, phone and TV, but also across multiple platforms.
“The moment you move from people to devices, you have a challenge,” Ruiz said.
This challenge applies to all audiences, but it’s more acute when trying to measure Hispanic and Latinx audiences.
Pixel me this
For brands, accurately measuring a US Hispanic audience requires some of the same tactics as measuring any audience, according to Ruedaquintero, including using site pixels to gather data about their location, browsing history, device and operating system.
Brands can also provide options for users to disclose first-party data. At minimum, brands should ask for first and last name, ZIP code and language via site forms, but they can also try to collect race and ethnicity, language of choice and other information. From there, brands can incorporate these data points into their strategy by, for instance, making a segment for users whose preferred language is Spanish.
But even without much personally identifiable information, brands can target a Hispanic audience using other signals, such as looking for soccer fans in the 10 US states with the highest concentration of Hispanics.
Measuring Hispanic audiences correctly also requires understanding the structures of whatever panel or database is used for measurement, Ruiz said, such as the type of households included, their geographic distribution and how many self-identify as Hispanic.
“You have to peel the onion to understand: Who are these Hispanics, and how were they recruited [for panels]?” he said.
Put your money where your mouth is
But traditional panels are a bit passé.
Using attention metrics has become an increasingly popular way to measure the success of creative versions featuring different languages served via programmatic CTV and display, according to Jess Nachtigall, EVP of analytics and optimization at Brandtech Group-owned performance agency Jellyfish.
For example, a brand might put out creative in English and Spanish and use CTV panels to measure the attention the different versions receive.
Ideally, marketers would have data for specific segments within the Hispanic/Latinx bucket. After all, crafting compelling messaging should look different depending on whether marketers are targeting, say, Colombian consumers versus Mexican, Puerto Rican or Dominican.
Yet many research and measurement outfits take a “pan-Hispanic perspective,” Waterston said, which obscures the many distinct Hispanic cultures in the US.
Whereas a pan-Hispanic approach finds commonalities in language and culture across Hispanic communities in the US, Ruiz said, campaigns focused on audiences from a specific country of origin are “absolutely powerful,” he said.
These more precisely targeted campaigns, however, are not always possible, because of audience measurement and segmentation challenges.
But ensuring robust sample sizes for measurement and research projects goes a long way toward mitigating measurement challenges, Waterston said. With healthy sample sizes, it’s possible to analyze data across different subgroups.
Still, an ongoing commitment to prioritizing Hispanic measurement takes a hefty – and ongoing – financial investment.
Corporate America needs to understand that “multicultural America is America,” Waterston said, “and there is no total market that’s not a diverse, multicultural market.”
“The dollars need to flow where the numbers and the people are flowing,” she said.