Musings on Entrepreneurship and Innovation
Saturday, May 26, 2007Uncertainty, Consumer Demand, and the 80/20 Rule
Ask the chairman of P&G - the $70 billion consumer goods giant - who is more important to his company's success: Wal-Mart or housewives. Sure, it's a tough choice, but I suspect that he'd answer "housewives." In other words, although the Long Tail distribution of market outcomes in the highly competitive world of consumer products is undoubtedly the result of the interaction of supply and demand, it seems likely to me that the dynamics of demand provide the foundation for the emergence of the winner-takes-most characteristic of consumer markets. Furthermore, the fundamental uncertainty of such dynamics are critically important for an individual inventor (and others in the innovation chain) to understand and appreciate when making investment decisions.
Skewed Demand
Robert Cialdini wrote Influence: The Psychology of Persuasion, one of the best selling books of its kind over the last decade or so. In it, he notes the following:
In general, when we are unsure of ourselves, when the situation is unclear or ambiguous, when uncertainty reigns, we are most likely to look to and accept the actions of others as correct.
In the context of buying decisions, we tend to look to others to the extent that we're uncertain about what movie to watch, book to buy, or music to listen to. It's a reasonably rational strategy in an uncertain world. More generally, our buying behaviors are more likely to be socially contingent to the extent that the promised value is qualitative, experiential, and emotional. That's important, because the dynamics of "social contagion" or "information cascades" are characterized by threshold rules or tipping points that are the signatures of "power law", "scale free", or "Long Tail" distributions.
As Chris Anderson has explored in a book, his weblog, and a downloadable manifesto, many consumer markets can be described in terms of a "short head" composed of a handful of hits and a "long tail" composed of everybody else. This is popularly known as the 80/20 rule, in which 80% of the revenue is generated by 20% of the products. Although directionally correct, the relative proportions aren't fixed. For example, data from the book business shows that something like 87% of book sales are attributable to just under 8% of the 1.2 million books published in a recent year. In other words, the book publishing industry reflects something like a 87/8 rule, which might look something like this:
(Distribution generated by demonstration tool developed by Fiona Maclachlan.)
Movies, similarly, seem to be distributed according to something like an 80/6 "rule". Clearly, books and movies fit the criteria of qualitative, experiential, and emotional. But, what about more tangible and objectively functional consumer products? Well, even sporting goods, for instance, are well described by something along the lines of a 60/10 rule:
In any case, the improbable (i.e. hits) dominate the distribution, a point that Nassim Nicholas Taleb makes more generally in Fooled by Randomness and The Black Swan. The implication for inventors of consumer products is sobering: Given a market composed of 100 distinct offerings, a 10% market share implies a #2 market ranking. If your product's sales rank falls outside of the top 10%, your market share is likely to fall below 1%. That is why, in my previous post, that I suggest that even in a success case, the range of revenue for a new, compelling product in a $400 million annual market might range between $4 million and $40 million and his most likely to be at the low end of the range.
Many times when we use words such as likely or expected, we are referring to the average or the mean of a distribution. That makes a lot of sense if the distribution is the familiar bell-shaped curve, or normal distribution. But, the average has no meaning (i.e. there is no representative scale) in the case of highly skewed Long Tail distributions. In such cases (which, ironically, may be the more normal when it comes to the demand for consumer products), the median or mode of the distribution are better indicators of "likely" outcomes.
Democratizing Supply Effects
Although I think it is appropriate to focus on demand dynamics, Chris Anderson has some interesting hypotheses regarding changes on the supply side. More specifically, he sees three key trends:
- The tools of production are becoming less expensive and more widely available.
- The cost of inventory and distribution is falling.
- The cost of search - the matching of supply and demand - is falling.
While most pronounced in the digital world of entertainment products such as movies and music, these trends extend to the physical world, as well. P&G has explicitly acknowledged the democratization of invention - the fuzzy front end of production - in its embrace of Open Innovation. As I wrote about here, the emergence of readily accessible and relatively inexpensive CAD software and expertise, rapid prototyping, rapid manufacturing, just-in-time inventory management, and various direct marketing techniques may mean that the distinction between digital products and physical products is more blurry than we commonly assume.
Anderson believes that these trends will shift the relative proportion of demand away from the short head toward the long tail. But, Anderson acknowledges the continuing importance of hits, while offering hope for the masses:
I've described the Long Tail as the death of the 80/20 Rule, even though it's actually nothing of the sort. The real 80/20 Rule is just the acknowledgment that a Pareto distribution is at work, and some things will sell a lot better than others...What the Long Tail offers, however, is the encouragement to not be dominated by the Rule. Even if 20 percent of the products account for 80 percent of the revenue, that's no reason not to carry the other 80 percent of the products. In Long Tail markets, where the carrying costs of inventory are low, the incentive is there to carry everything, regardless of the volume of its sales. Who knows - with good search and recommendations, a bottom 80 percent product could turn into a top 20 percent product.
Much of Anderson's writing describes the emergence of Long Tail aggregators such as Rhapsody, iTunes, and NetFlix. More broadly, Google, eBay, and Amazon.com are aggregators in that they provide a platform for reducing the cost of distribution and search across a broad range of products - digital and tangible, alike. Guthy-Renker, HSN, and EIP are variants on what might be called Long Tail prospectors - firms that search the "Long Tail of things" in order to connect the next hit with the capacities that can take advantage of economies of scale and scope, which remain important in the tangible world. (In a very real sense, EIP represents a search innovation.)
Implications for Inventors
- Having a great product helps, but there is no linear, causal relationship between the quality of your invention and market success. (See the suggested reading list below for more.)
- Look at the potential market from the top-down and the bottom-up. Try to be realistic. Express your conclusions in terms of a range of revenue. Divide the result by 10 to give you an idea of the "home run" potential of your product. Divide that number by 10 to give you a "likely success case" revenue estimate. Think in terms of probabilities; avoid point forecasts.
- Look for ways to run an in-market experiment as quickly and inexpensively as possible. In highly information-based markets, Anderson notes, "It's more expensive to evaluate than to release. Just do it!." (As I noted here, some venture capitalists are beginning to favor market experimentation over analysis, as well.)
- Invest assuming the likelihood of modest returns, but recognize that there is at least a small chance that your product will trigger a global cascade of demand. That is, it could really be a hit. Consequently, give early thought about how you might sell or license your product to a company that has the operational capacity to take advantage of the hit when it happens, because the window of opportunity is likely to slam shut very quickly.
- Respect the unavoidable uncertainty that surrounds the innovation process. Acknowledge the risk that others - including licensees and customers - are taking in adopting your product as theirs. On the other hand, don't fear uncertainty - it is the source of disproportionate opportunity.
More Reading
Feedback-driven, complex, non-linear systems are hard to understand and, quite often, counter-intuitive. In addition to the resources mentioned above, you may find the following useful in your exploration of the Long Tail:
- Critical Mass by Philip Ball - This book goes deeper than The Tipping Point and brings together some of the ideas presented in books such as Emergence, Ubiquity, Nexus, and Linked.
- Six Degrees by Duncan Watts - Includes a nice description of the dynamics of information cascades as well as to the limits of our ability to engineer social epidemics.
- Hollywood Economics by Arthur De Vany - This is for more advanced readers, but is a terrific resource.
The Long Tail: A Practical Guide for Inventors
The likely sales of a compelling product in a large consumer market is smaller than you think. On the other hand, there is small - but meaningful - chance that your product could be a runaway hit. Many, maybe most, individual inventors tend to fixate on the latter and ignore the former. The results can be disasterous.
The trap is easy to fall into. Just the other day, an inventor sent me some research that suggests that the total U.S. retail market for products like his may be as large as $400 million per year. To be honest, my initial reaction was, "That's pretty big. If this product achieved a mere 10% market share, it could generate $40 million of retail revenue per year."
Maybe, but highly unlikely.
A better range for planning purposes would be to assume that in a success case the revenue for this product may be between $4 million and $40 million. Furthermore, the most likely outcome will be $4 million or less. In contrast, a typical inventor might assume that $40 million represents a success case, while, say, half of that - $20 million in revenue - might represent a conservative case. The difference in expectations is very likely to drive quite different investing behavior, for starters. In other words, an inventor who is convinced that his product can generate retail revenues of from $20 million to $40 million is likely to behave differently than an inventor that believes that his product could generate revenues as high as $40 million, but is most likely to generate revenues of $4 million or less. The distinction could be the difference between a strained marriage and divorce; a smaller savings account and personal bankruptcy.
Consequently, I strongly recommend that inventors devote some time toward understanding the implications of the Long Tail distribution. To that end, I'm going to summarize my understanding as succinctly as possible in my next post. I hope it helps.