INSIGHTSobering Up the Data
A bottle of red, a bottle of white, and some iffy insights.
The NY Times recently published a piece explaining that Millennials are not buying enough wine, threatening the industry. At first blush it reads like the sort of Boomers v. Millennials conversation I’d walk (run) away from at a party. In search of more wine. But the data used to justify this beverage crisis, and the environment from which it was harvested, both deserve a serious look. Let’s uncork this. (I apologize in perpetuity.)
Here’s a summation of the argument presented: Baby Boomers buy a lot of wine but won’t be around forever, Gen X buys wine like Boomers but there are too few of them to matter, and Millennials aren’t buying enough wine. Civilization is doomed.
First thought: Trying to decode generational secrets to determine how to spend a significant portion of $296BN (total advertising spend in the US in 2021, estimated) has weak legs. Next thoughts:
There is no such thing as a generation, really.
Object permanence is for toddlers.
It’s dangerous to ascribe permanence to trends, trend lines, and other representations of data (i.e., believing a line on a graph that identifies a trend, and what that trend line represents, is immutably real). Data insights are not empirical. Insights are interpreted from numbers, by humans, from specific points of view. Data can certainly support a position, but insights are subject to the bias(es) of the interpreter(s). Opinions exist. Backed by data.
That’s it. Data insights are meant to serve as trail markers, not GPS.
Let’s call this delusion “data object permanence.” What is really happening is that data insights are reinforcing bias and de-risking marketing outcomes for marketers. The data is relied on (Gen Xers are a tiny little generation!) and acted upon (Millennials are a bigger generation worth focusing on!), and then… the needle doesn’t move very much. Sometimes in the wrong direction a little.
But our campaign was so clever! The data was wrong. The algorithm was wrong. Our product is best-in-class! Programmatic waste! The wrong 50% of waste was wasted. Too bad the deck gets reshuffled before every hand.
Back to the wine. It turns out that Boomers love buying wine; Millennials not as much (preferring a mix of alcoholic beverages, I guess. But still some wine. Insight!) Marketers can’t figure out why… A dog-bites-man story.
Gen X—begrudgingly smooshed between wine lovers (Boomers) and H8RZ (Millennials)—buy similarly to Boomers. Too bad the article also notes that the Gen X population has been declared too small to really bother with at only 64.95 million Americans. By comparison, there are 70.68 million Boomers* and 72.26 million Millennials.
Express these numbers relative to one another, and the insight is that Gen Xers are about 91.9% the size of Boomers (5.7 million fewer), and 89.9% of the Millennial set (7.3 million fewer). While millions of people are significant and meaningful, we’re not exactly talking about orders of magnitude here.
Going further, the article informs, “…because of the size difference, Generation X has less buying power, although its wine-buying behavior does not seem that different from Boomers.” The same data that explains why not to market wine to the Most-Ironic Generation could also show us that they are a group that is more likely to live longer (and buy more wine as a result of still being alive) than the preceding generation. And… ask one more question and the data permanence is further undone: How much less buying power?
It is simply unsound reasoning to declare a generation of consumers to be flyover country. The difference in overall size of an already shaky concept (generations being distinct segments rather than a gradient) is just 8.1% in one direction and 11% in the other. This is not enough to forget nearly 65 million Americans and their buying power, especially when preference in this age group is somewhat understood. It stands to reason that adjusting budget for scale, to some degree, makes a whole lot more sense. However, this does not really even come close to explaining and factoring in “buying power” and then appropriately weighting/adjusting for scale. If there were a number that actually quantifies the dip in buying power, and if that number were significant, it seems worth at least a mention.
I have a slightly-buzzed guess hypothesis as to why data is being used to justify this bias. “Millennials” are self-identifying, believing in the data object permanence of “Millennials” in the first place. People who fit the Millennial birth range do represent a giant chunk of the professional workforce today, and quite a number of them are working in marketing. Outgroup homogeneity is negligent; mass narcissism is worse.
The point here isn’t that I feel bad for forgotten Gen Xers. The point is that data was manipulated in such a funny way and little justification was given for the twist; the miscalculation was tacitly acknowledged as reasonable (or entirely overlooked).
“Data insights” make the unreasonable seem reasonable, again.
Last Call
Explaining the downward trend in wine sales due to generational problems is goofy. Generations are neither fixed nor entirely reliable as dataset brackets, and the data interpretation is biased. This all adds up to an offbeat position for marketing a beverage that has existed for thousands of years. It may be (but likely isn’t) the end of the wine-selling business in its current form but not the end of wine.
There’s a sobering insight for marketers to swallow.
_______________
Matthew Thornton
29 March, 2022
*The Boomer epoch is measured as the 18 years between 1946 and 1964, three years longer than the 15-year standard established for all generations following. See Roger Maris.