Welcome to part two (of five) in our ongoing analysis of marketing takeaways from 2020’s record-shattering US elections. Yesterday, I wrote about how cultural identity drives behavior, and the power of common enemies. Today, we’ll pick up the series by diving into the opportunities and limitations of both positioning and segmentation.
In case you skipped yesterday’s post, a few caveat reminders:
- We’re analyzing the election here on SparkToro in five parts, one each night this week, because its size and impact are such excellent teaching points for marketing, market research, and audience understanding.
- This series is meant to be non-partisan, but your author (me!) is generally politically progressive.
- Comments are welcome and encouraged, as is disagreement! But please, be kind and respectful, or the comments will be removed.
And now, on with the analysis.
#3: Changing Entrenched Positioning May Be Impossible (in the short term)
One of the most surprising outcomes of the 2020 election was the degree to which voters in a number of diverse geographies supported progressive and left-leaning legislation on the ballot while simultaneously rejecting the candidates who shared, endorsed, and promised to represent that legislation.
Indeed, when Americans are asked about the policies they support, from health care to drug regulation, climate change to social services, taxation to gun safety, and beyond, there’s a surprising amount of consensus. As observers often note, the United States’ citizens are, on most issues, a center-left group. But, when it comes to electing officials, the inverse is true.
One compelling theory is that while progressive policies may be popular, the Democratic Party’s brand is not nearly as well-liked. They lose elections not because voters disagree with their legislative plans, but because they have a negative impression of the party’s labels: Democrat, Liberal, Progressive, The Left, etc.
Some accuse the last fifty years of negative branding from well-funded opponents for this mentality. Others say Democratic politicians haven’t delivered on their promises, or even sold out on their policies to donors, special interest groups, and to compromise with Republicans.
So why not rebrand? Position themselves around popular issues instead of less-popular labels? Wouldn’t Democrats then reap the political rewards?
Via The New Republic: What if Democrats’ Message Just Doesn’t Matter
The paragraph above neatly sums up this peculiar dichotomy, and points out a problem almost every smart marketer struggles against. Simple sells. But reality is complex.
If you want to accurately, comprehensively describe your product, or position your company, or, heck, write a series of blog posts analyzing the 2020 elections through a marketing lens (so meta!), complexity and nuance are required. But if you wanna sell? If you wanna stick in people’s minds like a puerile Reddit meme? Keep it simple.
Via BusinessInsider’s: Why Democrats and Republicans Use Different Words
Language is one way to do this, and it can be quite effective. As BI noted in their piece, “The substitutions often work — an April Ipsos/NPR poll found that support for abolishing the estate tax jumps to 76% from 65% when you call it the death tax.“
Not bad for such an easy change.
But when it comes to overall positioning and how brand associations and ideas are built in our minds, there’s a wrinkle…
Via The New Republic: What if Democrats’ Message Just Doesn’t Matter
When market researchers ask large audiences for their views about a particular brand, person, political party, or candidate, the “common wisdom” is driven by years and decades of formative experiences.
Hillary Clinton could no more re-brand herself in 2016 than McDonald’s could convince people it’s serving healthy food (this despite both spending veritable fortunes in the attempt). Donald Trump could claim Joe Biden was a radical leftist in the vein of Bernie Sanders, but, because of Biden’s lengthy career as a well-known moderate, the labeling didn’t stick.
In the corporate world, Hubspot has spent years trying to re-position their brand around salespeople, not just marketers, with relative success. But among the early audiences who got to know them, it’s a much tougher sell (no pun intended). Uber and WeWork’s rebranding attempts have struggled mightily. Microsoft’s transition from the software world’s most-hated behemoth to a gentler, more benevolent giant has, more or less, succeeded, but took over a decade (more than a quarter of the company’s entire existence), a CEO change, and a strategic overhaul.
Marketing Takeaway: If your brand is well known and cemented in the minds of your audience, it’s very, very hard to change views in the short term. I experienced this at my previous startup, where the brand I’d built became synonymous with one market segment (SEO), and attempting to shift to a broader one took years of effort that was largely unsuccessful despite a relatively massive investment.
Based on what we’ve learned from the worlds of politics and human psychology, it probably makes more sense to launch new brands, rather than attempt to re-brand a well-established one, especially if you’re trying to move away from one association to another. The Democratic party probably doesn’t have this option, and so must play the long game of re-positioning themselves in the eyes of younger voters (whose views are still malleable). But, in the world of brand marketing, changing names or launching sub-brands is a very compelling option.
If your business is struggling to effectively reach a new audience segment, or if you’re hamstrung by historical associations, a new brand might be the way to go.
Whatever you do, make your language accessible and your concepts simple. It might even be worth following in Frank Lutz’s footsteps, creating a language guide for your team to talk about features, benefits, product, market, and the problem you’re solving.
#4: Demographic Segmentation is Almost Always the Wrong Choice
Almost everything you read about US elections focuses on four demographic elements: race, gender, age, and (more recently), education. These defaults spur an important question: are these demographics the most useful segments we could possibly apply?
Clearly, in national politics, where targeting of hundreds of millions of Americans is required, these elements are deeply entrenched. Media has so ingrained race, age, and gender as our defining characteristics that there’s little chance of anyone in the mainstream using anything else. Americans are supposed to inherently understand that their race, gender, age, and education are central to their identities and voting patterns, and it’s only those iconoclastic few who dare buck the trend.
Via BBC US Election 2020: Results
But, as we’ve seen in this year’s polling misses, this broad approach creates a lot of inaccuracy, and its expected that in the years ahead, these predictors will get worse. The buckets are just too big, and with rare exceptions, demographic cohorts oversimplify the wide range of attitudes and behaviors.
Why do political analyses do this?
Inertia likely plays a role. The United States’ media and pollsters are accustomed to dividing us by race, gender, income, age, and education. Because they’ve done it for so long, we’ve come to expect it, and so it’s the shorthand for how people talk about and analyze politics (and far too many other aspects of society). But the past few elections have proven how misleading these purely demographic boxes can be.
Most of the media now focuses on the most obvious “miss” of the 2020 demographic: Latinx voters, whom pollsters and media treated as presumably voting against Trump since he and his followers so consistently and publicly demonize them. But as you can read in pieces like Time’s Why the Complexities of 2020’s ‘Latino Vote’ Were Overlooked or The Guardian’s How Joe Biden Lost Florida’s Latino Voters or MarketWatch’s Why Polls get the Latino Vote So Wrong, putting people into buckets based solely on race turned out to be just as problematic as it sounds.
Latinx voters may have been highlighted in this election because of the diversity of voting patterns when homogeneity was expected, but they’re not alone. Black men shifted more toward Trump than expected. Jewish voters shifted to Trump. White evangelicals shifted to Biden. White voters without college degrees shifted to Biden.
If this market research and these results were presented to any competent corporate executive, that person would (hopefully) reject the whole process. And rightfully so! It is:
- Inherently racist and sexist
- Prone to shifts in behavior that cannot be seen or predicted based on the limited criteria applied
- Consistently, provably delivering poor quality predictions
- Needlessly broad and oddly ignorant of other useful predictors
I don’t know for certain whether other approaches have been tried by professional, political pollsters (psychological profiles, purchase habit data, values-based cohorts, population density, etc), but I do know that in the world of marketing, we don’t have to tolerate this shoddy approach. With rare exceptions (multinational consumer brands) no self-respecting marketer would dare pitch a campaign based exclusively on race, gender, age, and education demographics. They’d be (rightfully) tossed out of the room.
What we want is useful, usable data about our audiences. Demographics alone are almost always too simplistic, but when layered with psychographics and behavioral data, it can get much better.
At scale, correlations between purely demographic groups and behaviors exist. But whenever we perform market research, our goal is to gain the best understanding we can of what works, how, why, and with whom. Segmentation is a useful tactic that most marketers will be familiar with through the concept of “personas.” These avatars are meant to represent our customer targets through things like job roles, values, hobbies, lifestyle, and yes, demographics, too.
Ideally, when we research an audience, we gain an understanding of their proclivities in a nuanced and comprehensive way. We want to dig deep enough to get not just “interesting,” but *actionable* data about them.
Via a SparkToro profile search for “liberal”
SparkToro’s software only exists because Casey and I were intensely curious about useful, usable behavioral data on nuanced audiences at scale. We believed that knowing what people read, watch, listen-to, follow, talk-about, and engage-with online could be immensely valuable for better marketing (and were lucky to find many other marketers did too).
Marketing Takeaway: Electoral polling (and the shoddy, exclusively demographic segmentation used there) is a great case study in why we should avoid such simplistic, poor-performing buckets. When we perform customer research, we should segment with caution and care, seeking out archetypes that…
- are not reduced to arbitrary stereotypes based on purely demographic identifiers
- share characteristics with enough consistency to be targetable as a group (with messaging, through ad targeting platforms,
- give us actionable differences in how we build, serve, and or market to them
If you’re attempting to market a product, determine who the buyers are, break them into groups that match up with behavioral criteria you can act on, and then discover what makes them tick, how to reach them, and what resonates. In most companies of any scale, you’ll find that there’s an existing process for this… and it’s probably wrong!
Don’t get biased by the past. You can dramatically change marketing results by breaking out of the shell of how things have been done, and starting fresh. Usually, just a small experiment or two showing the efficacy of such an approach (a few high-performing ads, for example) can light the way to a big overhaul, and even better results.
If you found this valuable, don’t miss yesterday’s analysis: How cultural identity drives behavior, and the power of common enemies. Tomorrow, we’ll cover survey methodologies and why it might pay to invest in new customers rather than changing minds. Stay tuned!