Harbingers of Failure

Harbingers of Failure

More Exciting Clerical Work

Going through more notes, I found a few snippets and links from a topic that caught my attention nearly five years ago. There’s been just a tiny bit more research in that area since 2015, so I figured it’d be fun to give a super high-level breakdown on it and share something mildly interesting.


What is a Harbinger of Failure?

Back towards the end of 2015, a group of marketing professors published a paper titled, “Harbingers of Failure“.

The study challenged the assumption that all positive feedback from customers equated to increased long-term success by identifying a subset of consumers they called harbingers of failure. Interestingly, early adoption of a product by these consumers was a strong indicator that the product would fail. The study claimed that a household’s tendency to buy successful or unsuccessful products was systematic, in that customers who bought a failed product were more likely than others to buy another failed product. Interestingly, the converse was true as well, in that customers who bought successful products were more likely to buy other successful products.

It is important to mention that in this context, a product is a failure if “its last transaction date (in the store-level transaction data) is less than three years after its introduction.” Basically, if a product can’t keep selling for three years, it is considered a failure. There are probably better metrics for humanity, but for people in the marketing world I’d say this is a decently fair and empirically evaluable definition.

In addition to measuring a household’s previous purchases of a failed product, the study examined customer repeat purchases. Purchasing Coca-Cola Blak one time to see what it’s like is partially informative, but repeat buyers begin to more obviously fall into the eerily named category of weirdos known as the Harbringers.

Forgot about me, didn’t you?

While the paper was able track aggregate data from a large sample of customers consuming ~9,000 products and see that there was a subset of the population that regularly bought stinkers, there wasn’t much beyond that and the fact that the amount of product that group bought tended to correlate with how poorly the product did.

In short, harbingers of failure are a group of consumers whose early adoption of a product (and the degree of that adoption) correlate with the failure of a product within three years.


Harbingers To: Failure Boogaloo

Fast-forward four years later, and we finally get some follow-up in The Surprising Breadth of Harbingers of Failure.

In this long-awaited sequel, the findings on individual harbingers of failure is expanded upon, and it is found that there are not only individuals which predict product failure, but harbinger zip-codes as well. This study has perhaps even more interesting implications than the first, as harbinger zip-codes are dependent on a variety of factors such as retail purchases, political decisions, and even housing decisions. (As far as sequels go, I’d say it’s like the Shrek 2 to the original paper’s Shrek. It’s got a few more dimensions, but it’s still built comfortably on a topic that the original brought us all to know and love.)

A number of the traits of these harbinger zip-codes aren’t terribly surprising (older, whiter, more suburban, more single family households), but one interesting takeaway was that households which moved away from one harbinger zip-code tended to move to another harbinger zip-code.

This seems to indicate that harbingers of a feather flock together, rather than there being some sort of weird invasion of the body-snatchers scenario where households move to a harbinger area and then “learn” harbinger behaviors.


My Thoughts

It is kind of weird to think that there are pocket areas where high-concentrations of rogue non-conformists with a strong affinity for failure group together and presumably lament the modern world’s distinct lack of accessibility to 8-tracks and support for Microsoft Vista. Entire zip-codes disproportionately filled with people who handle their money differently than the rest of the world, whether it comes to store-bought goods, or houses, or even political donations. Areas that are filled with people who consistently defy the norm, adopt early, and back the totally wrong horse.

While it is mildly entertaining to think about the type of person who loved Crystal Pepsi, Battlefield Earth, and voted for Ralph Nader all bajillionty times he ran for President, I think it is much more interesting to consider how marketing firms could try to use this kind of demographic information to their advantage.

If it is possible to deduce whether a household or a zip-code is a harbinger of failure, then it would certainly be prudent for any agency looking to release a big product to take particular note of how these harbingers react to their products at its outset. Similarly, as there seems to be a converse group that could be coined “heralds of success”, it would seem obvious that there would inevitably be an elevated interest in these households and zip-codes as well.

It is possible that in the future of marketing and product-development, consumer feedback will be further analyzed and reduced by presupposed notions of “winners” and “losers”, judging from a sort of historical track record of purchasing. Maybe instead of certain parts of Ohio deciding elections, we’ll have certain counties in Washington deciding whether the next gizmo or doo-dad makes it to production.


As a side note, I think “harbinger of failure” is a terribly harsh term. I would much prefer to use “flop friendly“, or “outcome agnostic“, as opposed to something that sounds like a very dire insult or a very underwhelming first act to the Four Horseman of the Apocalypse.


If you have any interest in further reading, I’d recommend these links (as well as the original papers):

  • https://www.nytimes.com/2020/03/05/opinion/harbinger-failure.html
  • http://news.mit.edu/2015/harbinger-failure-consumers-unpopular-products-1223

I hope this got your noggin’ joggin’ on the future of marketing products in a stat-driven world, and that you found this enjoyable. I know that I’m very much looking forward to these papers being made into a trilogy. Maybe we’ll learn that society was the harbinger of failure all along (eat your heart out, M. Night Shyamalan).

Thanks for reading,

-WellTree

Leave a Reply

Your email address will not be published. Required fields are marked *