After President Donald Trump’s surprise victory in the 2016 presidential election, political polling got a bad rap. And the 2020 race did little to change people’s minds after many political polls again badly missed the mark.
But there’s much more to the world of polling, especially in the private sector, than the firms that monitor approval ratings and conduct horse race polls for political races.
“Private companies use polling primarily for three reasons,” explains Mike Moschella, director at DKC Analytics. “One, for testing marketing or product concepts, two, for tracking key metrics like aided awareness and net promoter score (a percentage likely to recommend minus the percentage unlikely to do so), and three, establishing thought leadership or justifying a new product line.”
BerlinRosen MD Alex Navarro-McKay suggests surveys can be used for everything from gauging broad public attitudes about industries and issues, predicting consumer behavior and assisting the trustworthiness of messengers and spokespeople.
Yoni Gedan, Zeno Group’s EVP for research, analytics and measurement, sums it up as “ensuring that key decision makers have the pulse of key stakeholder audiences.”
Armed with stakeholder input, a decision maker is able to build this into their strategy and execution, “while minimizing the risk associated with potentially damaging public opinion ‘blind spots,’” he adds. “Polling, when done effectively, is a powerful mechanism for ensuring that all voices can be heard.”
Ultimately, polling can help teams “more effectively and more rapidly get to the ‘why behind the what’ that underlies and drives the analytics,” Gedan adds. “This allows [teams] to more swiftly and deftly optimize campaigns and maximize campaign impact.”
It’s become easier to gather this insight as the types of available polling have increased over time, from relying on phone surveys to the choice of phone, online or focus group. However, other types of polling are not as simple. Moschella notes that an interested company needs a vendor partner who is responsible for buying space within apps or sites that reach the targeted consumer.
“Then you need someone with enough of a stats background to run the counts and crosstabs and provide valid analysis if you plan on having any media scrutiny,” he says.
The shift away from telephone surveys has been accelerated by the fact that it’s harder to reach people by phone than before, notes Navarro-McKay.
“The cost of getting a representative sample by telephone has gone up because fewer people answer their phones when unknown numbers call, and it is more expensive to call cell phones because [Federal Communications Commission] regulations require that these be dialed manually. And the ability to reach people at lower cost online is growing,” he says.
Moschella adds that the growing domination of online polling is a result of the fact that “it has the broadest reach, fastest speed and lowest cost—three characteristics rarely found together.” But he warns that PR pros need to make sure the sample is not flawed by being mindful of demographics and delivery mechanisms.
Navarro-McKay also advises against an online-only strategy. Instead, he calls the “multi-modal” approach of using several methods to reach respondents the “gold standard” as it is more likely to produce a representative sample.
Moschella argues that “the market research world is Balkanized” because of the numerous polling methods available. Another reason: it’s a lot harder to understand who or what is a “winner” outside of the political arena.
“We can tell if you’re right in political polling because people vote,” he says. “We’ll never be able to verify a poll about whether people prefer sweet potatoes or mashed potatoes as a Thanksgiving side.”
There are other challenges to both private and political polling. Michael Maslansky, CEO of Maslansky+Partners, notes that turnout is important, but getting the sample right is crucial.
“You need to make sure that your sample represents not just the population but the population of people who will actually vote,” or in the case of private companies, buy, he notes.
There’s also the inherent bias of who will respond to a poll or survey. Moschella uses the example of a poll around video games.
“If something pops up on your screen about video games, then video game players will inherently be more likely to stop what they’re doing and fill out a survey,” he says. “As the researcher, you have to gut check what your sample looks like compared to other available data...For example, in running a survey of U.S. homeowners, we’ll triple-check to make sure the age and income breakdown mirrors known homeowner demographics.”
There’s another hurdle: declining social trust. Navarro-McKay cites political data expert David Shor, who has argued that in demographic groups with decreasing social trust, those individuals are less likely to participate in surveys, a fact that those conducting polls may struggle to account for.
“If Shor is right, survey research that under-represents voters who are less likely to trust other people and institutions could have implications for private sector research,” he says. “If brands and other organizations are missing a portion of this growing segment of the population, they may fail to develop effective marketing strategies and may underestimate opposition to a broad variety of campaigns.”
Gedan is more optimistic. He notes that real-world outcomes or outcome variables are readily and frequently available in private polling.
“Because we have access to such robust and rapid ‘feedback loops’ into our research and data models, we are continuously able to test, refine, optimize, augment and validate our research on a rapid basis,” he says.
What’s more, it’s all based on an extensive dataset. This is unlike in political polling, which faces the challenge of a number of variables and nuances, such as “varying voter segment participation rates, presence and impact of third-party candidate(s), variations in available methods of voter participations and even influencing factors such as weather conditions,” Gedan says.
Even though there are complexities for private polling, the ability to obtain “larger and more frequent ‘outcome variables’ allow us to continually refine and optimize our research and models,” he argues.
Despite the challenges to private polling, Moschella concedes that it remains easier than political polling.
“It doesn’t have to contend nearly as much with the problem of ‘who is the voter or consumer?’ Anyone can go to the store anytime and make themselves the consumer,” he says. “Comparatively, voter registration and then actually voting is a whole different process.”