Should You Bet on What The Wealthy Do Today?

In 2011, Hal Varian, chief economist at Google, wrote an essay published in Foreignpolicy which generated lots of discussions.

He asserted that the march of technological progress was leading to an unprecedented access to power for corporations, even for the smallest ones. Communication and computation innovations were so affordable that the smallest company could have access to much more than what a large corporations had access to 15 years before. “If the late 20th century was the age of the multinational company, the early 21st will be the age of the micromultinational: small companies that operate globally.

Yet, this access to power wouldn’t stop at the borders of corporations. Technology is progressing so quickly that the products and services offered to end-users by these corporations would become more and more affordable overtime. Thus, Hal Varian concluded “a simple way to forecast the future is to look at what rich people have today; middle-income people will have something equivalent in 10 years, and poor people will have it in an additional decade. Think of VCRs, flat-screen TVs, mobile phones, and the like. Today, rich people have chauffeurs. In 10 years or less, middle-income drivers will be able to afford robotic cars that drive themselves, at least in some circumstances”.

Andrew McAfee, a principal research scientist at MIT, later called this the Varian Rule in an article published in the Financial Times, adding “scale economies and online efficiencies will combine to keep driving prices down within this business model, and entrepreneurs will soon realise that the middle class is a huge market, and so tailor offerings toward it.”

Having the Varian Rule in mind, the VC job seems quite easy: just spend time with the wealthy, look at the products and services that they use, which others can’t afford (yet), and bet on them.

Optimism Or Credulity? The Varian Rule And Its Limits

The belief in the Varian Rule has certainly driven many VCs to bet heavily on the “on-demand economy”, which tried to replicate the convenience of services dedicated to the ultra wealthy for mainstream consumers. Yet, many failures in that field can easily be explained: even if the transaction costs (to find, book and pay, for instance) are lowered, most of the costs of these services are not lowered by technology. Unlike TV Screen, you cannot divide the price of your cook or home-cleaner by 100. Damn humans.

This of course doesn’t mean that betting on these startups is always a mistake. Many of them manage to build huge companies and bring high returns to their investors, founders, and employees (not that often to people providing the service, but that’s another story). However, as long as cars don’t drive themselves, and as long as cleaning is not done by little drones, the on-demand economy is mostly NOT following what the Varian Rule states. The decreased transaction costs, increased convenience & decreased barriers to entry, might increase the market and change the habits of the upper middle class in the megalopolis, but it would be very surprising -all things being equal- if these habits would become international mainstream. It is just too expensive, since a large percentage of the costs are not compressible.

Let’s just focus on Uber, one, if not the biggest success story of the on-demand economy. If you think they are already mainstream, here are some figures to have in mind (worldwide stats for 2017, if not stated otherwise):

  1. Uber vs other digital successes:
    Uber is soon reaching 40 million monthly active users (MAUs). Netflix has 110m MAU, including 50+ million in the US alone. LINE has 170m MAU. Amazon has 310m active customers, including 65m prime members. Alibaba has 507m MAU. Instagram has 800m MAU. Facebook has 2B MAU.
  2. Uber vs traditional alternatives
    You could say that there are “only” 80m cars sold each year worldwide. Yet there are 1.2B cars owned in the world. So we should not forget that Uber users represent only 3% of people riding in a car. You can say that a more correct comparison would be taxis. It might be right, but taxis are also not mainstream. The point is to show that TV became mainstream (1.7b of households own a tv) by following the Varian Rule, but Uber never will.
  3. Framework about diffusion of innovation (more details below)

Everett M. Rogers, one of the most influential scholars in the field of diffusion of innovation, stated that diffusion follows a normal law of distribution, and to become mainstream it must cross the chasm between the early adopters and the early majority, which occurs at about 16% of the population. That would mean, only focusing on adults, 7.6b*65%*16% = 790m of users in the world. In France, that would mean (broadening the target audience to the 15–69): 67m*68,4%*16%= 7.3m. Uber just announced having reached 2m users in France. The chasm is still far ahead. Yet, in major cities, with higher levels of income, it indeed became mainstream reaching 26% of users.

Ok, the Varian Rule is not universal. It does not work that well for the on-demand economy. And it does not work outside the boundaries of the economic realm where technology is removing most of the costs.

But does that mean that VCs should not spend all their time observing the wealthy, playing tennis, golf and drinking cocktails on their boats?

Shouldn’t we just be a little more picky about the bets, choosing only the activity where technology can transform a high-end service into a mainstream service?

Are The Usages From The Rich A Good Glimpse Into the Future

I would love to commit fully to the Varian Rule and spend all my time on boats. Yet, if Apple and Uber could have been predicted by the Varian Rule (personal computers and personal drivers were not mainstream), I’m afraid, it does not do a good job at predicting every kind of innovations.

The key to determine how much time we should spend observing the behaviours of the wealthy is about understanding the diffusion of innovations.

Diffusion of innovations has been studied since the 1940s, within different research traditions, each focusing on one specific kind of innovation. For instance, rural sociologists investigated the diffusion of agricultural innovations to farmers, while education researchers studied the spread of new teaching ideas among school personnel. Interestingly, despite the divergences on the methodologies, most of the research schools at that time found similar patterns, such as the S-shaped curve of adoption and the overall higher socioeconomic status of early adopters.

The S-shaped diffusion curve, Rogers

The different diffusion traditions — anthropology, early sociology, rural sociology, education, public health & medical sociology, communication, marketing, geography and general sociology — have studied and highlighted lots of models of diffusion.

Crossing the chasm, your favorite book (and its wonderful framework*), does NOT cover the full picture. Yes, I know, this is bad news.

* Adopter Categorization, developed by Rogers in 1985, made famous by Moore

There is more to understanding diffusion of innovations than just knowing who the 5 ideal-types (innovators, early adopters, early majority, late majority and laggards) and the two inflection points are.

Crossing the chasm from early adopters to early majority enable to reach the tipping point, Rogers

Yet, since the S-curve has been discovered in almost all research traditions, it might be a good idea to take a closer look at the characteristics of these ideal-types and understand where innovation comes from, to know whether the wealthy are indeed the people to observe. We need to know who launches innovations and, even more, who helps trigger the critical mass and enables an innovation to become mainstream.

The Socioeconomic Characteristics of the 5 Ideal Types

Let’s start by discussing your refusal to categorize — we are all unique, I know. Why should we create so-called “groups of people?” And how are they constructed?

Well, researchers use ideal types to make the world modelisable and comparison possible. Ideal types are not average in all observations. Ideal types are indeed based on abstractions from empirical investigations. Exceptions to the ideal types can be found, because if no deviations exist, there would be no use of ideal types.

Everett Rogers in Diffusion of Innovation, summarized the main generalizations of the 5 ideal types found in the voluminous research literature about diffusion.

Here are some of the key discoveries:

  1. Earlier adopters are no different from later adopters in age. (!)
  2. Earlier adopters have more years of formal education than do later adopters.
  3. Earlier adopters are more likely to be literate than are later adopters.
  4. Earlier adopters have higher social status (income, standard of living, occupational prestige, possession of wealth…) than do later adopters.
  5. Earlier adopters have larger-sized units (farms, schools, companies, and so on) than do later adopters.
  6. Earlier adopters have a greater degree of upward social mobility than do later adopters.
  7. Earlier adopters are more cosmopolite (the degree to which an individual is oriented outside a social system) than are later adopters.
  8. Earlier adopters are more highly interconnected through interpersonal networks in their social system than are later adopters.

I think we should make at least three observations about these discoveries:

  • The Varian Rule seems eventually quite compelling in the light of these generalizations (earlier adopters have a higher social status, and overall, more wealth).
  • These relationship are mere correlations. No causality can be concluded on the basis of available cross-sectional data. In other other words: we cannot know whether earlier adopters innovate because they are richer or if they are richer because they innovate (mostly thanks to an increase of productivity).
  • If you assume that earlier adopters become richer because they innovate (and point 6 could be an indicator of that), the Varian Rule could be (at least partially) invalidated: innovations could not go from the wealthy to the non-wealthy but from the soon-to-be wealthy to the already wealthy (and then to the non-wealthy).

And in fact, if the most expensive products might be the privilege of the wealthy at first, lots of other innovation, be it services or products, are not a question of sole affordability.

Innovation also comes from non-wealthy people. Innovations can come from the street. Think hip-hop culture. Innovations can come from outside the boats. Think AirBnb.

But if we say that the Varian Rule covers only a subsegment of the segment of early adopters, who are the other people to observe?

Opinion Leadership, Diffusion Networks, Weak Ties & Change Agents

If you manage to discover patterns of diffusion, the best way to forecast future usage would be to just follow the flow of information within and between networks, and observe the nodes where innovations first appear.

The best way should be to look for opinion leadership. Let’s use Rogers (yes, once again) definition: “opinion leadership is the degree to which an individual is able to influence informally other individuals’ attitudes or overt behavior in a desired way with relative frequency.” (Diffusion of Innovation, Rogers)

Networks provide a certain degree of structure and stability in the predictability of human behavior. When you look for innovations, you look for change within a given system, which comes from another system. So you look for the exchange potential of certain links.

And there is a ton of research in that domain too! One of the very interesting discoveries is that the exchange potential of network links (and thus ability to bring change) is negatively related to the degree of (1) communication proximity (overlap of the individuals’ networks) and (2) homophily (similarity of attributes within the network). Consequently, heterophilous (difference of attributes) links of low proximity play a central role in the diffusion of innovation.

This is a quite fancy conclusion to share during a social event. But in fact, you might have known it already. Remember the strength-of-the-weak ties theory? The theory comes from the work of Granovetter (1973), who discovered that most people find new jobs not through close contacts but by “[individuals] marginally included in current network of contacts, such as an old college friend or a former workmate or employer, with whom sporadic contact had been maintained.”. Low proximity. And often, heterophilous. Bingo!

And when researchers looked closely at these weak links, to try to identify bridge links (the individuals linking different networks and bringing change), they discovered that they shared a common attribute. The bridge links are most often opinion leaders, who play a critical role in diffusion networks. Compared to followers, opinion leaders have “greater mass media exposure, more cosmopoliteness [remember: the degree to which an individual is oriented outside a social system], greater contact with change agents [individuals who strategically influence innovation-decisions & thus behavior in a given direction], greater social participation, higher social status, and more innovativeness” (ibid).

If we try to summarize all of this and develop our understanding of diffusion of innovation, we shall say that: opinion leaders are super effective in bringing behavior change and diffusing innovations. They can be rich. But they not always are. They are the people to observe closely.

The big question is: how to make sure these opinions leaders are really diffusing innovations? And why exactly do people engage in an activity before the critical mass can be reached (a sufficient number of people engaged in an activity)?

Diffusion of innovation: critical mass & threshold

To understand the role of opinion leaders, and the dynamics of the diffusion of innovation, we need to talk about the concept of threshold. A threshold is basically the number of individuals who must be engaged in an activity before a given individual will join that activity (Granovetter, 1978). An early majority individual has a much lower threshold than a late majority individual. Let’s take the illustration provided by Granovetter to understand why some people engage in an activity before the critical mass is reached, and why opinions leaders are so important:

Imagine 100 people milling around in a square — a potential riot situation. Suppose their riot thresholds are distributed as follows: there is one individual with threshold 0, one with threshold 1, one with threshold 2, and so on up to the last individual with threshold 99. This is a distribution of thresholds. The outcome is clear and could be described as a “bandwagon” or “domino” effect: The person with threshold 0, the “instigator” engages in riot behavior — let’s say (s)he breaks a window. This activates the person with threshold 1. The activity of these two people then activates the person with threshold 2, and so on, until all 100 people have joined.

If we remove the individual with threshold 1, no one will follow the trouble-maker. There would not be a riot. Or just a one-person riot. No critical mass would be reached.

This is precisely the point of Derek Sivers in his quite famous “How to start a movement” TED talk. People with threshold 1 are super important (too).

And opinion leaders are super important because they are central nodes in different networks, which can help the reach of the critical mass.

A good way to think about it is to visualize someone’s personal network. Maybe the overall network is far from reaching critical mass (let’s say that less than 1% of the French population is using device X). Yet, if 40% of your colleagues are users of device X — because it was brought and promoted by one opinion leader in your company — you might reach your personal threshold and thus increase the overall adoption rate, which would eventually lead to a critical mass. Another way to frame it is to think of the locality of network effects: I don’t care if there are enough Uber drivers in France or in the world, I care about having enough drivers in the city I am currently in. And that’s why opinion leaders are so key to bringing innovation: they manage to convince people by talking to early adopters (with a low threshold) and via the snowball effect they can start to trigger the threshold of later adopters overtime, which are also in their network. And, by aggregation of opinion leaders, they can diffuse innovations and bring a niche usage into a future mainstream usage.

CONCLUSION

  • Research traditions indicate that the main challenge for a given innovation to become mainstream, is to cross the chasm from the early adopters to the early majority. When the chasm is crossed, you reach a critical mass (tipping point).
  • Research suggests that, overall, innovators and early adopters are wealthy, so it can be a good idea to just observe what the wealthy do to predict future usages.
  • Yet, innovations and early adopters might only become wealthy by their use of innovations (increased productivity & economic attraction), which has been the case for many innovations brought to the market “by the street”. So solely focusing on the wealthy would NOT allow to fully predict future usages (it would be only partial).
  • Observing the wealthy, as suggested by the Varian Rule, is especially a good idea when the products and services can become significantly cheaper through technology, at a fast pace (which would explain why VCs focus on scalability and on products and services where marginal costs can be very low).
  • To make sure not to miss an innovation, we should observe opinion leaders (and many rich people are, so it’s not a pure contradiction of the Varian Rule), the more cosmopolite and mobile they are, the better bridge links they will be. These people are often at the crossroad of many social, economic and cultural systems and often a good way to predict the future.
  • If you apply this topic for VCs, two big questions remain:

1. Should VCs only focus on innovations when they already reach the opinion leaders, or should they also fund products and services before they do (R&D stage or before launch)? Market practices suggest that institutional investors focus on the former (excepting seed & pre-series A investments), especially because the expected value of these investments are higher (higher probability of success).

2. If VCs focus on the innovations that already reached opinion leaders (and started to have some tractions), the other big question is: how to determine which innovations you should bet on? If there is a chasm between the early adopters and the early majority, it’s because many innovations talked about by opinions leaders will never work… But that’s for another post.

Eventually, what we should remember, maybe even more than the Varian Rule, is what Saint Exupery said in Wisdom of the Sands: the best way to foresee the future is to enable it.

register to our weekly newsletter: http://bit.ly/daphni-newsletter

Thanks to Isabelle Thonicke & Mathieu Daix for their proofreading & numerous inputs.