The Anti-influencer curse: Can a certain set of individuals buying your product indicate that your product is bound to fail?

Marketers have always tried to find ‘the right influencer(s)’ who could boost the brand’s appeal, be it companies like Boost who have been advertising with cricketers like Virat Kohli or Ford Fiesta running mass influencer campaigns with 100 influencers in 2009. Studies have long tried to find the customers who are the right set of influencers who can boost the chances of a product's success. But it is very interesting to note that a certain set of customers (the anti-influencers) have a knack for buying the products that are going to fail. The paper: “Harbingers of Failure” authored by MIT Professors Duncan Simester and Catherine Tucke explains this eccentric phenomenon.

Who are the “Harbingers of Failure”?

A study was done on two large data sets from a large chain of convenience stores that have their reach across the U.S. The first data set consists of the aggregate weekly transactions from 111 store branches, from January 2003 to October 2009. The second set of data consists of individual-level transaction data from November 2003 to November 2005. The researchers examined 77,744 customers who purchased 8,809 new products between 2003 and 2005 and then tracked the aggregate data longer to see these products fared in the longer run.

As a key focus area of the study, the researchers studied consumers ‘whose purchases flop’ at least 50 percent of the time and saw pronounced effects when these harbingers of failure buy products. When the percentage of total sales of a product accounted for by these consumers increases from 25 to 50 percent, the probability of success for that product decreases by 31 percent. And when the harbingers buy a product at least three times, it’s really bad news: The probability of success for that product drops 56 percent.

When the mainstream market didn't find takers for products such as crystal Pepsi, Coca Cola black, and Oreo Watermelon cookies, a rare and curious breed of consumers— “harbingers of failure”—would probably like many such items. Image Credit: MetroMBA

This interesting phenomenon doesn’t stop here! In their paper, The Surprising Breadth of Harbingers of Failure, authored by Duncan Simester, Catherine Tucker and Clair Yang build on their earlier research. First, the findings document the existence of “harbinger zip codes.”, that is, if households in these zip codes adopt a new product, this is a signal that the new product will fail. Second, comparisons reveal that households in harbinger zip codes make other decisions that differ from other households. The first comparison ‘identifies harbinger zip codes’ using purchases from one retailer and then evaluates purchases at a different retailer. The comparison revealed that ‘households in harbinger zip codes purchase products from the second retailer that other households are less likely to purchase’. The analysis next compared the donations by households to congressional election candidates; the study found that households in harbinger zip codes donate to different candidates than households in neighboring zip codes, and they donate to candidates who are less likely to win. Finally, house prices in harbinger zip codes also increase at slower rates than in neighboring zip codes.

Why do they behave so?

When purchases happen in such a manner where the basic human instinct may suggest that these set of people might be irrational to pick the losing product with high frequency. But Professor Tucker, one of the authors says that there is no conclusive evidence for this to be the case, while also jovially mentioning that she herself was one of these Harbingers who selectively choose the products that are going to fail. One of the possible explanations could be that these people have a higher preference for risk as per Professor Simester. Another possible explanation could be that the tastes of these people are generally leaning towards unconventional products.

An interesting opinion article in Times of India by a Columnist further throws some light on the decisions made by her (an anti-influencer) such as her enthusiastic adoption of Google Plus, the liking for films that fare poorly at the box office such as the 1987 Hollywood movie Ishtar, her goto moisturizer going out of the market, her liking for the losing politicians, etc. and her self-diagnosis on why she might have taken such decisions.

When further studies were performed, attributing the ‘zip code effect of Harbingers’ to behavioral spread also proved to be false. In this study the Professors compared the buying habits of households before and after they moved out of the Zip Code, to see if transitioning to a new area affected them. It didn’t and even when they moved to a neighborhood of non-harbingers, their liking for failure-prone products persisted. This shows that they didn’t rub off on their neighbors, either. Most likely these harbingers had clustered together in choosing the same product, but a non-harbinger may not start making these peculiar decisions just because they moved to a harbinger neighborhood.

This incredible finding suggests that the clustering is not a result of social learning but of water seeking its own level. Evidently, the same irregular effect that had happened while browsing the aisles of supermarkets and clothing stores also guides their decisions about where to live, leading them to the same neighborhoods.  

Why this might be of interest to managers and tech enthusiasts?

Even though studies to date couldn’t find the exact reason for such behavior, this phenomenon might have an impact on how managers and data scientists make their decisions in the near future.

Firstly, if your product or service is reaching a very small niche with needs or preferences far outside the mainstream, it may be difficult to scale enough for success. Thus, the conventional wisdom of building a product for the niche and scaling up should be approached with a caveat as it becomes very important to find the 'right set of niche' who make up a decent size of the population to avoid falling into the trap of satisfying only the anti-influencers.

Secondly, a direct application of this could be that retailers (both online and offline) could make more qualified decisions on what products should be stocked less/ shown less on display so that they can cut down overstocking costs in the offline mode and restrict the digital marketing budgets of products that are bound to fail based on the buying patterns of these anti-influencers.

Finally, the role of data and AI in the field of marketing would start to become more important than ever, as analyzing such voluminous purchase data would and prescribing the right course of action has the potential to deliver promising results. This would also mean that the boundaries within which the technologists and marketers operate start to become thinner and investments in technology such as AI would become almost inevitable over the years.  

References

1. Dizikes, P. (2015, December 23). Are you a "harbinger of failure"? Retrieved November 24, 2020, from https://news.mit.edu/2015/harbinger-failure-consumers-unpopular-products-1223
2. Kottke, J. (2019, December 10). The "Harbinger Customers" Who Buy Unpopular Products & Back Losing Politicians. Retrieved November 24, 2020, from https://kottke.org/19/12/the-harbinger-customers-who-buy-unpopular-products-back-losing-politicians
3. Where are the harbingers of failure? (2019, December 24). Retrieved November 24, 2020, from https://www.warc.com/newsandopinion/news/where-are-the-harbingers-of-failure/43072
4. Simester, D. I., Tucker, C. E., & Yang, C. (2019). The Surprising Breadth of Harbingers of Failure. Journal of Marketing Research. doi:10.1177/0022243719867935
5. Gopalakrishnan, A. (2020, March 08). The anti-influencers who can't sell a thing! Retrieved November 24, 2020, from https://timesofindia.indiatimes.com/blogs/to-name-and-address/the-anti-influencers-who-cant-sell-a-thing/
6. Stone, A. (2020, March 05). Are You an Anti-Influencer? Retrieved November 24, 2020, from https://www.nytimes.com/2020/03/05/opinion/harbinger-failure.html
7. Evans. (2020, March 05). Do YOU Recognize Anti-influencers -- You Really Should. Retrieved November 24, 2020, from https://evansonmarketing.com/2020/03/10/do-you-recognize-anti-influencers/
8. Anderson, E., Lin, S., Simester, D., & Tucker, C. (n.d.). Are you a 'harbinger of failure? Journal of Marketing, 52(5), 580-592. doi:10.1509/jmr.13.0415
9. Cathleen O'Grady - Dec 26, 2. (2015, December 26). Harbingers of failure: Meet the customers you don't want to love your product. Retrieved November 24, 2020, from https://arstechnica.com/science/2015/12/certain-customers-spell-doom-for-new-products/
10. Insight of the Week: When Niche is A Non-Starter. (2015, November 27). Retrieved November 24, 2020, from https://www.cardwellbeach.com/insight-of-the-week-when-niche-is-a-non-starter/