amazon highlights: Location is (still) Everything / David R. Bell

Location is (Still) Everything: The Surprising Influence of the Real World on How We Search, Shop, and Sell in the Virtual One by David R. Bell ----- Last annotated on September 25, 2014

if you and I live under different physical circumstances and in different physical environments, we will use the virtual world very differently—even if we are very similar people in terms of our ages, incomes, education levels, and so on. We’ll shop differently, search differently, and won’t be equally attractive to sellers.Formun Altı

So if we want to make sense of how we use the virtual world, we need to have a framework for understanding the different influences of the real world on how we shop, search, and sell online.

Young, wealthy, and well-educated citizens are the most active in the virtual world. Within a country, individuals who live in rural areas, or those with lower educational attainment and less income, tend to use the Internet less.

relationships that exist between our physical-world locations and our virtual-world behaviors will be robust.  

The virtual world is flat in terms of the opportunity that it delivers to all of us, but it is not flat in the way that we use it. Because the way we use it to search, shop, and sell depends on where we live in the real world, which is anything but flat.

 

Reilly’s law of retail gravitation, which describes in simple terms how you and I as consumers decide where to buy the goods and services that we consume.

In GEOGRAPHY I look at the key principles that explain how we decide where to live and what influences the kinds of goods and services that are available in our locations. Once we know how the real world is organized and varies from one location to another, we have a platform for understanding how the virtual world is used.

RESISTANCE shows how the real world and the virtual world interact with each other—in the consumption of both products and services as well as information and content. The real world often places obstacles in the way of our getting what we want, and the virtual world often helps us remove them.

ADJACENCY explains why it matters who is next to whom. Specifically, this chapter examines how demand patterns for virtual-world sellers spread systematically from one real-world location to the next.

VICINITY demonstrates that there is more to how things cluster in the virtual world than is explained by adjacency alone. When groups are formed in the virtual world, whether around the consumption of content or products, they are often made up of people who live far apart. The twist, however, is that these people who live far apart do share common real-world circumstances and preferences. This, in turn, drives similarities in their behavior.

ISOLATION explains the importance of the relationship we have with the people who live near us in the real world and how that affects what we do in the virtual world. On average (like it or not!), we are quite similar to our neighbors in many regards, but in situations in which we have different tastes for products and information, the virtual world liberates us in special ways.

TOPOGRAPHY concludes the book and looks to the future by explaining how products, information, and people move through and traverse the landscape of both the real and virtual worlds. I consider what happens, for example, when the virtual world goes mobile, and how the variation of real-world landscapes predicts the behavior of buyers and sellers in the virtual world.

Spatial Long Tail (SLT). In the SLT, initial sales arise from proximity among close neighborhoods in large markets or “hot spots,” and later sales arise from similarity among distant but comparable smaller neighborhoods. This concept shows that demand often comes first from customers who are adjacent to each other, and then through customers who are similar to each other, but live increasingly farther apart.

GE • OG • RA • PHY n. The study of the physical features of the Earth and its atmosphere, and of human activity as it affects and is affected by these.

Diapers.com offers exactly the same prices, assortment, and convenience (one-day free shipping for orders over $50) to both customers in our example. In other words, the virtual world is completely flat, offering the same “package” to both real-world locations. Since the real-world locations for Taylor and Paige are different, the relative attractiveness of shopping at Diapers.com from their given location (Houston or Philadelphia) depends on how it compares to their offline alternatives. Paige in Pennsylvania does not pay taxes on diapers bought offline, whereas Taylor from Texas does. Right now, neither shopper would have to pay taxes on her online purchases.

There’s also a lot of evidence that our preferences for individual brands are shaped by where we live. The word “undifferentiated” is important. By this I mean products that are more or less identical and indistinguishable from each other in blind taste tests or similar comparisons. Assume that both brands spend about the same amount on marketing. Then try to imagine why consumers in Los Angeles prefer brand A by a factor of 2:1, while consumers in Philadelphia prefer brand B by a factor of 2:1. The answer to this question about undifferentiated products illustrates the important way geography shapes our tastes. Just being first in a particular market allows you to dominate that market for decades and decades, even as new entrants come along. Formun Üstü

Academics and consultants refer to this persistence of brand loyalty as a first-mover advantage. The choices made by these customers are not shaped by inherent preferences; rather, their preferences have simply been shaped by where they live. Well, our location shapes what we want, and because of that we adjust very quickly to our new surroundings, but only partially. The “location of the past” stays with us, exerting influence over us, for quite a long time!

So we have two strong, robust, and perhaps surprising impacts of circumstances on preferences.

The first that is where you live impacts how you use the Internet and its related technologies to buy goods and services and to connect with others. This effect works through the force of the offline options that you have.
The second effect is more subtle, but nevertheless very powerful. The preferences you develop for the individual brands that you end up liking, buying, and consuming are also shaped by where you live.

There are at least three sets of factors that shape our choice of where to call home.

First, there is the force of inertia and family ties.
Second, there are economic barriers and incentives. Individuals who share the same neighborhood are likely to have similar levels of income or wealth, and to share characteristics that drive income and wealth.
Third, there is the old saying that “birds of a feather flock together.” Homophily is not just determined by cultural, ethnic, and demographic factors either. Preferences—of all types—can be an equally strong basis for it. There is a statistical relationship between the propensity of a location’s citizens to shop online and for e-commerce sellers to base their businesses there. Knowing what goods and services are offered in a location helps us to infer something about the preferences of its residents.

So for now, let’s keep three things separate: (1) economic factors, (2) cultural factors, and (3) preferences for local goods and services. Zipf’s law is based on the fact that the natural log of a number can be used to tell us something about the amount of time needed to achieve growth in a quantity of interest. The quantity of interest could be money, people (or population size), the number of bacteria in a dish, and so on.   

central place theory (CPT). CPT argues that, in the real or offline world, consumers visit the nearest “central places” that provide the goods and services that they need, and they make these decisions so as to minimize travel distance. Two concepts underlie CPT.

First and foremost: there is a minimum (or threshold) market size that must exist before a particular good or service will be supplied to a market.

Second, there is a maximum distance (or range) that consumers are prepared to travel to procure a good or service.   

You can’t succeed in the virtual world unless you understand where and how your potential customers are situated in the real one.

RE • SIS • TANCE n. A force that tends to oppose or retard motion.

Markets don’t always provide us with exactly what we’d like to know to make an informed choice, or exactly the right product for our needs and desires—let alone at the right price. We’ll use some helpful jargon and refer to them as “frictions.” This term really gets to the nature of these obstacles—they don’t absolutely prevent you from doing things, but they do make your life more difficult.

The first is “search friction,” or the inability to get the information that you and I would like to have before making a decision. The real world has an important and pervasive friction. You have to expend effort—incur search costs—in order to make better decisions. This is how the Internet solves what I’ll call geographic friction number one—you can live in Sioux City (a small market) but still have access to New York City (big market) variety. The Internet liberates everyone by aggregating people, independent of their location, and creating a large, addressable market. So the Internet creates markets made up of “similar” people, independent of where they actually live. So the Internet removes, or in many settings at least substantially mitigates, the two most intransigent and enduring market frictions. With it, you can find out the price of almost anything without leaving your house, connect with almost anyone, and buy pretty much whatever you like no matter where you live—so long as someone is willing to deliver it to you.   those of us who live in large cities tend, on average, to be heavier consumers of local information. farther you have to travel to reach particular offline stores, the more likely you are to shop online. Research shows that the larger the city, the more local sites there are that serve up content devoted to the people, places, and activities that define it.

BOPS means “Buy Online, Pick up in Store.” When a store offers BOPS, you can go online and check out, say, a new espresso maker or a duvet, buy it, and then wander by the store and pick it up from there. This phenomenon was coined the “delay premium,” and means that you and I need to be paid more to have to wait longer than expected, compared with how much we’d be willing to pay to shorten the wait time by the same amount. So in some sense, BOPS gives you the best of both worlds (full information before purchase and no waiting for delivery, although you do of course have to stop by the store to pick the thing up).

Most of us like to try clothes on before buying them. Clothes have tactile, or “touch and feel,” attributes that are hard to communicate online. Traffic at the website was up because shoppers could go online to get price and in-stock information before buying (the Internet had reduced their search friction), but they still wanted to turn on the coffee machine, sit on the sofa, or touch the duvet before buying. So they didn’t actually buy what they wanted online. Having confirmed the price and that the item was in stock, they went to the store to inspect the product. Hence, sales at the store went up. Nevertheless, BOPS was a great success because it led to—wait for it—ROPO (Research Online, Purchase Offline).

This is why sales of popular items at Amazon.com will go down in zip codes contained in the trading of the new store after it opens, but sales of niche items will not. They found that goods with a high value-to-weight ratio were the most likely candidates to be sold via e-commerce. For non-taste-dependent categories, distance doesn’t matter at all; i.e., there is no statistically significant effect of distance on consumers’ behavior. The theory underlying this finding is that the closer that creators and users of content are to each other in the real world, the likelier they are to have similar interests or tastes. Distance is a proxy for cultural similarity.

The real world puts obstacles in our way. Important information (e.g., about prices and product quality) can be hard to come by or costly to obtain, and not all products and services are offered in all locations. The virtual world helps us to overcome these search and geographic frictions, often eliminating them entirely. However, the Internet sometimes imposes frictions of its own—by making us wait for delivery of things bought online, or by making us uneasy about buying products with “touch and feel” components. So, the real and virtual worlds compete with and complement each other.

You might begin your journey searching in the virtual world, and end it by shopping in the real one.

Gravity still holds sway over the way we consume content and information goods in the virtual world!
Frictions frustrate in the real and virtual worlds—you’ll succeed in both when you identify and eliminate them.  

AD • JA • CEN • CY n. The state of being so near as to be touching. The state of being adjacent; contiguity.

Local stores with space constraints and relatively limited trading areas are not motivated to carry slow-moving items.

I need to point out an important connection between this chapter (ADJACENCY) and the next one (VICINITY). This one looks at what happens in the virtual world as a result of proximate individuals sharing locations. The next one looks at what happens when individuals share characteristics or preferences independent of the locations that they live in. When we look at social contagion derived from adjacency, we look at: who, where, how, and what.

Let’s say the “who” are the friends and acquaintances one comes across locally; the “where” is the local neighborhood; and the “how” is either via direct communication (a person-to-person conversation) or observation. The “what” is the content. The glue that holds everything together for our analysis is the real-world location, the “where.” On average, it looks as though there is more positive information contained in reviews on the Internet than negative information. Researchers at Yale found that a negative review not only slows the sales of whatever is being reviewed but also tends to reduce the flow of positive reviews. If we take these findings into account and consider that there is more positive information to begin with overall, then, in total, positive word of mouth will be at least as effective as negative word of mouth, if not more so. When reviewers reveal information about themselves, including their location, readers become more willing to trust what’s written in their reviews. we imagine that those living near us will have similar tastes and preferences as ourselves.

the New York Times reported in 2012 that about one in three online reviews is fake—a number determined by the data-mining expert Bing Liu. And one rather enterprising man, Todd Rutherford, was exposed as making nearly $30,000 a month selling fake reviews online! And that’s the important general point from the research that carries over to other types of sites: you should be more trusting of those that have some kind of verification mechanism built in. Sellers have an incentive to manufacture positive reviews for their own products and negative reviews for their rivals. In order for real-world conversations about a product to have an effect on virtual-world buying behavior, the product itself has to deliver.

The manufacturers of the energy drink Red Bull built upon this logic too. On college campuses in the United States, popular students were given cases of Red Bull to distribute at the parties they hosted. DJs at popular clubs were encouraged to drink Red Bull; bar owners often left the empty cans visible to customers.

Starbucks and Dunkin’ Donuts like to see customers take their coffee to go. The cups are visible to others.

In the study, over 300,000 leaflets (about one per resident) were dropped on the city from a plane. The leaflet contained a message asking anyone who came across it to fill it out and mail it back. About 7 percent of the leaflets (just under 20,000) were mailed back to the experimenters, which is a pretty decent response rate. Of the responses, 55 percent were from physical diffusion—i.e., someone picked up a leaflet, filled it in, and mailed it back. The remaining 45 percent of the responses were from so-called social diffusion. This means that the respondent got the leaflet from someone else first. So some people saw the leaflets and acted on them directly, while others passed them along to friends, who complied with the request. people who filled in the leaflets directly tended to live right underneath the places where the drops had occurred, whereas the people who learned about the leaflets socially tended to live farther from where the leaflets had been dropped. So in the real world, proximity and social transmission go hand in hand.

the total amount of influence to be exerted on any one location by the adjacent neighboring locations is 100 percent exactly, and no more. Let’s say that zip code number 2 has four times the population of zip code number 3. To reflect this, we could replace the ½, ½, entries that are currently in row 4 of figure 3.1 with ⅘, and ⅕. Physical neighborhoods can be defined not only in terms of their characteristics but also spatially and in relation to one another. Because people who live in immediate neighborhood clusters are also likely to be similar to each other in specific ways, it’s not that hard to imagine ideas, fads, and trends spreading among them. The structure of our physical neighborhood helps determine how information spreads through adjacent locations and how likely interaction among individuals is to begin with.

Three things are critical.

First, physical population density—the more people there are, the more opportunities there are for incidental contact.   
Second, when there are more people in your neighborhood, there are more offline establishments where interaction can occur.
Third, the people themselves are important to consider.

Entire industries, starting with the government census, are built up around collecting data about zip codes. And just as important, everyone who shops or sells on the Internet either gives out or receives zip code information in the form of billing and shipping addresses. New customers popped up in areas adjacent to areas that already had customers.

number 1 and 2, and people in both of these zip codes have already started using Netgrocer.com, chances are it’s only a matter of time before someone in zip code number 3 does as well. This second effect—contagion resulting from adjacency—proved very robust in the case of Netgrocer.com. First-time orders from new customers were more likely to show up in a zip code when there had been previous orders by other customers in contiguous zip codes. This effect persisted even after accounting for the “birds of a feather” effect, or the tendency of co-located people, such as those in the same zip code, to behave similarly. So holding all else constant, as our economist friends might say (and it is important to do that), there are very strong real-world contagion dynamics that create these demand-spreading patterns for virtual-world sellers. In short, the trading area for Netgrocer.com expanded systematically. Once a few customers sprouted up somewhere, their neighbors soon joined them. This kind of contagion resulting from real-world adjacency has shown up in the data of every virtual-world seller that I’ve looked at.28 In chapter 4, VICINITY, I’ll look more closely at where customers pop up to begin with, expand the idea of “proximity,” and also examine some of the best ways to “seed” locations with new customers. Virtual-world businesses benefit tremendously from real-world adjacencies among customers and neighborhoods. Intelligent sellers understand this and do their best to cultivate activity that leverages adjacencies. with the two great benefits of adjacency—homophily and influence.

Homophily implies that people who live next to each other share characteristics and circumstances, and so make similar trade-offs, and can be drawn to use similar virtual-world products and services. When they communicate and emulate, influence happens. Homophily can be leveraged, but influence can be fanned. That’s why it’s critical for virtual-world sellers to consider how to make their products and services socially visible.

Formun ÜstüIt was no accident that my Soap.com boxes came in vibrant mixed colors like teal, orange, chocolate, and magenta, with “Soap.com” written on them in white letters. If your workplace “family” is bigger than your family at home, then your virtual-world seller should encourage you to get your items shipped to your office as well. That means more exposure—my Wharton colleagues have seen many of my orders over the years! 

 

In the virtual world, most sellers have unlimited reach—through a whole country, or perhaps even the entire world—yet their sales don’t occur just anywhere.

 

The patterns of demand that emerge when a seller conquers an “unlimited market” have special properties. They are rooted in adjacency, as new buyers spring up in areas contiguous with ones that have buyers already.

 

Adjacency rules. You’ll succeed when you seed it and let proximity power demand.

VI • CIN • I • TY n. The state of being near in space or in relationship; proximity.

Homophily and influence result from, and are made possible by, adjacency.

Why did Fischer argue the Chicagoan and the Angeleno were more likely to meet? Even though the physical distance from Chicago to Springfield is about one tenth of the distance to Los Angeles, the social distance between Chicago and Springfield is much greater than it is between Los Angeles and Chicago. just because locations are separated whopping physical distances, it doesn’t mean that they won’t behave the same way when it comes to adopting new products and services.

Also, thanks to the Internet, a large physical distance between locations no longer precludes influence or interaction flowing between them. They found that households living in relatively close proximity to each other were more likely to own the same model of car (and even the same color). The important principle uncovered in the study is that demographic similarity (as well as geographic proximity) has a big effect on car choice as well.

People pay attention not only to what kind of cars are on the block, but also to what kind of cars people like themselves are driving, regardless of where they saw the car. So there is quite a bit of evidence indicating that similarity, more than distance alone, matters a lot for choices and behaviors. basic concept of adjacency extends beyond just individuals and neighborhoods to corporate entities as well.   At the same time, two other forces pointed to the fact that certifying companies were in the same vicinity as each other as well.

First, individual companies were more likely to obtain management standard certification (ISO 9000) when other companies with similar trading patterns had already done so.

Second, companies sought environmental certification (ISO 14000) when other early adopter companies in “culturally similar” countries had already done so.

One of the most fundamental things that the virtual world does is to allow the formation of groups and markets that don’t (or can’t) exist in the physical world. This raises interesting questions about how the virtual world is used and affects human behavior. On the one hand, the Internet could lead to more diversity of interests. On the other hand, for almost any kind of interest or activity, there is now also access to a sufficient density of others who are just a clique away.The authors show, mathematically, that as access to others improves, “agents” (or individual consumers) tend to seek out similar types of individuals in the virtual world and give up some of their local connections in the real world. assume that you could somehow “seed,” or start one hundred conversations about your show in order to try to get a bit of buzz humming along.

Would you rather have ten different groups of people engaged in ten different chats about your show in ten different locations, or one single group of people discussing it intensively in one hundred different conversations? It turns out that if you want to generate awareness about what you’re doing, it’s generally better to have “width” rather than “depth,” holding the total volume of word of mouth constant. In the virtual world, as in the real one, “spread” or dispersion of information about your product, helps a lot.

A study conducted at MIT by Catherine Tucker found that Facebook advertising by a brand that is targeted at people who are “friends of people who are friends of that brand” has click-through rates that are twice those for ads sent to a random sample of people. friendship in the real world can, on average, be a reliable proxy for response to stimuli in the virtual one. This is an extremely powerful insight for sellers: physical-world relationships impact virtual-world responses to marketing and advertising. After studying tens of thousands of users and their interactions, the authors arrived at three pretty fascinating conclusions.

First, they found that there is significant variation in the extent of influence that happens in virtual-world networks, both in the capacity of one individual to influence another in that network and in the susceptibility of an individual to fall under the influence of others.

Second, we are each influenced, on average, by about 20 percent of our friends.
Third, and perhaps most interesting, the study found that patterns of influence in virtual-world social networks among users in the same vicinity, i.e., clusters of friends, are also driven and shaped by real-world factors.

Males tend to systematically overstate their income and women tend to systematically understate their weight.

So there are systematic discrepancies between what people say they are and will do in the virtual world, and what they actually are and are willing to do in the real world.

I focused on three key factors: (1) the number of existing customers within a zip code, (2) the number of existing customers across all the zip code’s geographic neighbors, and (3) the number of customers across the zip code’s demographic neighbors.

First, there is a very strong immediate neighbor effect.

Second, we found that sales took off initially in what we came to term “hot spots” and spread out from there. 

Spatial Long Tail (SLT). when we plotted the number of new customers in a histogram against the real-world locations that generated them, we uncovered an intriguing phenomenon. In the SLT the x-axis shows all the counties in the state of Pennsylvania, ranked from those with the most customers to those with the least (going from left to right, so that the resulting graph slopes downward). The y-axis just shows the number of customers in each county. It starts with an observation about product supply and availability, namely, that the Internet has dramatically reduced the costs of offering variety—Amazon can carry millions of books whereas even a very large physical store would struggle to carry more than, say, 100,000. If you were running Amazon and selling millions of books, you could attempt the following: At the end of the year, you could total the sales for each book and rank the books from best- to worst-selling. If you did, you’d notice a pattern. You’d have a few books that sold really well—books like Harry Potter and the Bible. You’d also have hundreds of thousands of books that sold hardly at all, perhaps just one or two copies.

If you drew a histogram of product sales, starting from the most popular product all the way down to the least popular product, you’d see that the total sales from all those rather slow sellers would actually add up to a pretty significant number. Specifically, if you drew a vertical line in the histogram with all the books with sales ranks of, say, 1 to 100,000 to the left of the line and those ranked 100,001 and higher to the right, you’d observe that books to the right of the line, i.e., those in the “tail,” might account for up to a quarter of all your profits!

The “head” of the diagram is the popular stuff (the books that would actually make it onto a physical bookstore shelf), and the “tail” is the niche or relatively unpopular stuff.

Customers in the head locations were acquired largely via proximity effects. Adjacency among physical locations facilitated the spreading of word of mouth and emulation as well. Customers in the tail locations, on the other hand, were generated via similarity. This is because there is an 80/20 weighting at the beginning of the virtual-world sales process, in which 80 percent of the new customers initially came from proximity effects, which eventually gave way to a 50/50 balance once the business stabilized. But the best approach is to focus initially on relatively well-populated areas and thereby take advantage of the proximity effect. Then, as the rate of new-customer growth stabilizes, online sellers should then alter their strategies to pick up new customers via the similarity effect.

Given the opportunity, you might end up “linking” only to others who are just like you!

Virtual-world influence comes in a variety of real-world flavors too. If you want to have sway over others, it helps to be younger, female, and an early member of the virtual-world community.

The end result is the Spatial Long Tail, in which the head is demand from customers connected through “proximity,” and the tail is demand from customers connected by “similarity.” A large pool of customers still comes from the “main” locations, but all the smaller locations matter too, and collectively add up to a lot!

Value comes from vicinity—you’ll succeed when you collect head and tail locations together.

I • SO • LA • TION n. The act of isolating, the quality or condition of being isolated.

Now, there are two kinds of real-world isolation that are important drivers of how we use the virtual world.

We discussed the first one, geographic isolation, RESISTANCE. general, the amount of “pull” a commercial establishment like a store or restaurant has over you is a function of how close it is and how “good” it is—in terms of offering you what you want.

The second condition—preference isolation—is more subtle but just as powerful. What this means in practical terms is that sellers in a particular location will tend to cater to the tastes of the majority. adding certain “types” to a market creates more variety for those types, but not necessarily for others. it’s probably going to be hard to find any males or females that are isolated in their gender-specific preferences. Second, preference minority status could be attributable to having a characteristic that is temporary, such as being the parent of a newborn child. we have two different and very interesting forces of real-world friendship on virtual-world behavior. (Recall that the study was a comparison of the effect of actions taken by one’s friends versus actions taken by random others in the network.)

When portraying ourselves in the virtual world, we seem to like to diverge from what others do and retain a sense of individual expression—except when our choices will be publically announced and our friends are all doing the same thing. It’s nice to express your individuality unless you really have to swim against the tide of your friends in a very public way! This suggests an interesting “balancing out” between the real and virtual worlds, in the sense that one makes up for the other. In locations where individuals were underserved by their local markets—online sales were higher. In locations where they were less isolated, in the sense of having greater access to offline apparel, online sales were significantly lower. Niche products are bought by those with minority tastes, so virtual-world sellers can cater to these individuals without fear of competing head on with offline stores.

Researchers at MIT wanted to understand what would happen when retailers in the real and virtual worlds offered identical product assortments and prices. Would they have identical patterns of sales? In particular, they were interested to see if there would be differences in sales of products from the Long Tail—those desired by customers with more niche or isolated preferences. So, it’s real-world (catalog) versus virtual-world (website). The researchers found that even though the seller’s offerings were identical in the two channels, the virtual-world channel of the Internet saw higher levels of sales for products located in the “tail” of the long-tail plot.

Unit sales of niche products were 14.8 percent in the Internet channel, and dollar sales were 15.0 percent, compared with only 12.7 percent for both metrics in the catalog channel. The effects were highly statistically significant. Since prices and assortments were identical, what could explain this discrepancy?

The big difference between the two channels is that shoppers with niche preferences were more able to find products in the “tail” on the Internet. This is because the virtual world of the Internet comes replete with recommendations from customers and with the ability to search. Regardless of their station in the real world, shoppers in the virtual world are more able to get into the tail and discover niche items. If you’re a seller and you can find customers who are isolated and who want niche or long tail products as well, demand from them will be very high. Customers will be happy, and also relatively price-insensitive—which is a winning combination for sellers.

In RESISTANCE, we saw that the benefit of buying online depends a lot on where a person lives.   In this chapter we added another very important condition: It’s not only where we live but who we live next to as well that matters for our choices. If the majority of those who live next to you are too different in characteristics or stage of life or preferences, then you’ll be a preference minority and turn to the virtual world for liberation. On average, we share characteristics with others in our locations, but that doesn’t mean that our preferences will always coincide with those of our neighbors.

In the real world, isolation deflates and dictates that you’re not going to get what you want. Furthermore, in the real world, product variety is targeted. There have to be sufficient numbers of people needing a certain good or service, or it just doesn’t get offered locally at all. So, for virtual-world sellers our status with respect to local others is of critical importance. Virtual-world sellers have a more difficult time competing with real-world sellers offering mainstream products. Instead, they should focus on target locations in which potential customers form a preference minority.

These customers have high offline searching and shopping costs, and are therefore not price sensitive.

As a result, a disproportionate amount of demand for a virtual-world seller comes from “preference minority” locations

Isolation offline becomes liberation online. You’ll succeed by identifying, serving, and investing in the isolated.

TO • POG • RA • PHY n. The arrangement of the natural and artificial physical features of an area.

The real-world population keeps expanding, and the virtual-world one does too. Internet penetration is about 35 percent on average, with the most-penetrated countries (these include Germany, Japan, Korea, the United Kingdom, and the United States) at around 80 percent or above. is also interesting to note that Internet penetration contributes to GDP, in the sense that countries with more connection to the virtual world enjoy all kinds of efficiencies, cutting-edge business models, and so on.

One of the most important things shaping the topography of commerce in both worlds is the steady move by shoppers and sellers from fixed Internet-connected devices (desktops and laptops) to mobile ones (smartphones, feature phones, tablets, and so on). I’m using “fixed” to mean desktops and laptops because while laptops are of course movable, consumers typically don’t use them (literally) while moving about, and, more important, a mobile device is more likely to be connected to just one single user.

I want to share a little bit of background research on what is known about browsing behavior in general.

Below the ads are the top “organic links,” that is, the links to the sites that Google thinks will be the most helpful to you. In contrast, for sponsored links, the basic idea is that the link that shows up first is the link that Google thinks is the best fit for you, after taking into account what a seller has bid for the terms so that its link gets shown to you. how do you (and I) respond to these sponsored ad links? the first link is typically the one most likely to be clicked on, the second the second most likely, and so on. the seller with the highest rank in the paid search process gets the best response or, in industry language for the most common metric for measuring response, the best click-through rate (CTR). on a fixed device like a laptop, one move up in the ranking delivers a 25 percent lift in the CTR. If the chance that you click on the link ranked second is 10 percent then the chance that you click on the top link is 12.5 percent, or 25 percent more. Bear in mind of course that the cost to the seller goes up as well—you have to bid more on the search terms if you want to come out on top. we know there’s a “rank premium” of 25 percent on fixed devices. What about on mobile devices? Is it the same, more, or less? The researchers found that moving up one position in the ranking increased the click-through rate by 37 percent, a lift of about 50 percent over what happens on the fixed devices, for exactly the same kinds of ads.

Going mobile increases your virtual-world search friction, because you’re less likely to click on the lower-ranked links and there’s a higher lift for the higher-ranked links. when you’re searching for local stores and brands on your fixed Internet device, a one-mile decrease between where you are and where the store is increases the CTR by about 12 percent.

The researchers found that searching on a mobile device doubles the “one mile premium.” using a mobile device increases geographic friction, because it makes closer things even more attractive to us. The increased rank premium or increase in search friction is due to the size of the device. The screen is smaller, so lower-ranked results are more difficult to access. the mobile Internet, relative to fixed Internet, has five critical features that are going to dramatically alter searching, shopping, and selling.

what are the five rules?

First, mobile has what Rob calls “instant-on.” Wherever we are in the real world, we can be using our “dead time” to participate in the virtual one:

Second, sensors on the device allow us to take photos, share videos, and of course communicate location information.

Third—and perhaps more controversially—the mobile device is the domain of apps.

Fourth, at no time in human history, have more people been connected to a device that is essentially a distribution channel (for product and content) connected to a payment system. mobile devices greatly amplify the power of word of mouth.

fifth and final point—show and tell, as he did by turning up the heat on his family via iPhone and Nest, is a compelling way to spread the word.

many of us are building our own personal brands and being ever more public in various social forums. if you’re an Internet seller and you want to enlist a celebrity to promote your brand or business, it’s best to find someone with a highly engaged fan base. Such fans have a strong affinity with the celebrity and are highly connected to him or her.

The researchers used some clever mathematics to tease out two different types of motivation and to measure them from the data.

The first kind of motivation is extrinsic. Extrinsic motivations are related to ego. We feel good about ourselves because people want to follow us!

The second kind of motivation is intrinsic. This kind of motivation is far more altruistic—our goal is primarily to help others through the sharing of our insights and information.

the researchers discovered two very interesting things.

First, certain kinds of execution improved the CTRs. Specifically, ads that were more “obtrusive” got more clicks. That is, ads that moved, popped up, popped down, utilized video, and so on just got more attention. Next the researchers looked at ads that were “contextually targeted” and found that they got more CTRs too. Contextually targeted means that the content of the ad matched the content of the site. So, obtrusiveness works (more CTRs). Contextual targeting does too (more CTRs as well). That’s the first finding.

The second finding tweaks this first one in a somewhat disturbing way (if you’re a seller, anyway). While obtrusiveness and contextual targeting work individually, doing them at the same time is not a good idea. Zappos.com, the famous shoe seller, pioneered “two-way free shipping,” and this is now de rigueur in e-commerce.

The fashion eyewear seller WarbyParker.com offers a “home try-on” program in which customers get five frames for five days to try at home for free. It’s an ideal company for the purposes of our study, because all the items it sells have “touch and feel” attributes. Our analysis showed that up to half of all the first-time customers of Bonobos had gotten some helpful information about the company from existing customers of Bonobos who shared their offline neighborhood. Bonobos.com sales were always higher in neighborhoods with more real-world social capital, i.e., in neighborhoods with more trust and interaction among neighbors. In neighborhoods with more social capital, information transmission is just better. What gets communicated has more veracity. If what is getting communicated is “good,” as in, “These Bonobos pants are great!” then a seller benefits.

“Social capital does not operate on (sales) directly; rather, it improves the learning process and therefore indirectly drives sales when what is communicated is favorable.”

Big media like the New York Times and the Washington Post reside in big towns like New York City and Washington, DC. They have a large local base and a broad national following. Let’s adopt the researchers’ nomenclature and call these kinds of news services Traditional Media Outlets (TMOs). In addition, there are more organic and typically smaller sources of information about products and services as well. These include review sites and blogs and other forms of user-generated content (UGC). These are known as Social Media Outlets (SMOs). So what’s better for sellers? TMO coverage or SMO coverage?

TMOs provide the greater overall lift in sales for an online seller. They have a broader reach, and when an event like a media mention happens, sales spike. Places like the New York Times just have a lot of readers.

Now, on the other hand, individuals participating in SMOs are more committed to the issues and causes that define the SMO, so their engagement is stronger. That leads to a nice twist to the result: In totality SMOs were more effective for two reasons. First, even though the per-event (or mention) effect of being named and discussed in an SMO is smaller than the per-event effect of a mention in a TMO, the mentions are more frequent. Net-net, more frequent but smaller impacts add up to a greater total effect than large per-event, but infrequent mentions in major media. The four main ways that most Internet retailers acquire customers are: (a) offline word of mouth (WOM), (b) online WOM, (c) traditional advertising, and (d) online search.

Since the real and virtual worlds compete with and complement each other, the next logical step is to weave them together. If you’re a virtual-world seller of “touch and feel” products, you need real-world touch points with customers. Virtual world advertising is more effective when obtrusive or contextually targeted, but not both. the fixed Internet dramatically reduces search and geographic frictions, but the mobile Internet brings them back. On mobile we want more seamless information (i.e., we are less tolerant of search friction) and we’re more attracted to options that are locally close (i.e., we are less willing to travel). So, we’re becoming simultaneously more spontaneous and more local in our behaviors. And this will create huge opportunities and enhancements for all local commerce. Your real-world life, made better by the virtual one.

No longer “either/or,” just a single transforming topography—you’ll succeed when you use one world to enhance what you do in the other.

Making GRAVITY Work for You

So if you are trying to get a virtual-world business going, it makes sense to choose a real-world location carefully—one where, on average, customers and employees are co-located.

Many forward-thinking e-commerce companies, such as Warby Parker and Bonobos, create open environments at corporate headquarters where customers can visit.

Step one in applying the GRAVITY framework is to understand the implications of GEOGRAPHY, particularly in terms of access to early customers and employees. they implemented a Home Try-On program, enabling potential customers to receive five pairs of glasses (without lenses) for free, and have five days to try them on.

Thus, step two is about countering RESISTANCE to information search and purchase by giving customers the means to reduce their uncertainty. Specifically, it’s about giving customers access to the information required to enable them to make better choices and feel comfortable doing so. Market research shows that customers who request a Home Try-On packet share the contents with another three to four people on average.

Warby Parker also leverages adjacency effects in real-world locations—by putting showrooms in places where potential customers are going to be not only the most receptive to the brand, but also most likely to tell others about the product. You want to make sure that you give potential customers the ability to share your product or service and therefore grant access to a proximate group that will then deliver more customers for you.

So as a general principle, you need to make sure you have a strategy that allows you to link and access clusters of customers who are distant from each other yet belong in the same marketplace—yours. So make sure that you identify the real-world locations that will give you access to the most receptive customers: those most starved of local choices that are relevant to them. So, this real-world event was reported, shared, amplified, and commented on through a variety of virtual-world sources. In early 2012 the team decided to retrofit an old yellow school bus (a familiar, iconic vehicle of sorts), and drive it all over the country. And so began the Warby Parker Class Trip. This was essentially a movable real-world showroom journeying all over the country, including places that did not have Warby Parker stores or third-party retailers that stocked Warby Parker glasses. In addition to experiencing the Trip when it rolled into town, customers could share the journey virtually and chat about it via Twitter, Instagram, and a variety of other virtual-world platforms. Customers also had the opportunity to enter a contest involving the virtual-world sharing of these real-world activities. My colleagues Toni and Santiago and I collected and analyzed data on the locations where the bus stopped. In all of these locations, we found significant increases in online sales, in addition to the sales that took place on the bus itself.

The upshot of this model for your own entrepreneurial efforts? Make sure that you allow your customers to access your business in both real-world and virtual-world contexts in one seamless and natural experience.

First, you should ideally be addressing a very common problem—one that affects a large market.

Second, the founders of your team—and most research shows that two (or more) is better than one—should exhibit magnetic personalities, with a passion for the company that is obvious and infectious. These qualities will inspire others to come on board and see the vision through to fruition.

Third, and this is nonnegotiable, the team must be able to efficiently and effectively get things done. Ideas are a dime a dozen. Successful execution of them is something else.

Always keep in mind that the benefits of shopping, selling, and searching online cannot be separated from the locations where customers live in the real world. This means that you must take into account the cities and towns that potential customers reside in, whom they choose to live among, and what they have access to.

Your virtual-world business, whether selling goods, services, or information, must quickly create sufficient density—often by connecting and aggregating individuals from disparate locations. Density breeds critical mass and engagement, and is necessary for more density to follow.

Remember that the tactics and strategies required for gaining followers, disseminating content, or selling products online need to be customized for different locations. Moreover, the virtual world is a great amplifier for whatever gets done locally in the real one.