The Great Delicious Exodus of 2010

Slide from Yahoo! webcast posted on Twitter by Eric Marcoullier

Everything old is new again. This is an essay about communities migrating between online services. I wrote it as a scholarly paper but never managed to get it published in a journal. But now that it seems like everyone is talking about defecting from Twitter and where to go next, I think it’s timely.

A little background: I wrote this essay in 2011–12. And it’s been rejected from every scholarly journal that I’ve submitted it to. Which is I think 5 by now, I honestly don’t even remember any more. (I have Thoughts about why that is, mostly having to do with journal editors saying that they want to publish interdisciplinary work, but reviewers not knowing what to make of interdisciplinary work when they have it. But yes, ok, it’s probably also that this essay is just not very good.) After a while I just got tired of looking at this thing and set it aside. I’ve been just sitting on it for 4 or 5 years at this point.

And yet — all available evidence to the contrary — I still believe that this essay is interesting and valuable. Especially now.

Forgive the academic-ese of the writing style. If I’d known from the start that I was writing this for Medium, I would have done a lot of things differently.

Introduction

In December 2010, the company Yahoo! held an internal webcast. A screenshot of a single slide in that webcast was captured by Eric Marcoullier, and posted on Twitter [1]. That screenshot showed Delicious — one of the first social bookmarking web services — included in a list of products under the heading “Sunset.” Within minutes, that image had been spread far and wide, the Web was ablaze with discussion and speculation about the future of Delicious, and options for alternative bookmarking services were being sought by panicked Delicious users.

Yahoo was relatively quick to announce that they would not be eliminating Delicious, but rather looking for a buyer… quick in business terms, perhaps, but not in Internet time. Within 24 hours of Marcoullier’s original tweet, the damage was done. Many Delicious users were, en masse, seeking alternative services to fill the gap that, many believed, Delicious was going to leave.

Enter Pinboard (pinboard.in). Pinboard is a bookmarking service that bills itself as “social bookmarking for introverts.” While Delicious had always been as much a social networking service as a bookmarking service, Pinboard’s focus “is less on socializing, and more on speed and utility” [2]. Prior to Yahoo’s accidental announcement of the sunsetting of Delicious, Pinboard’s relationship to Delicious was, in the words of Pinboard’s creator Maciej Cegłowski, “that of a tick to an elephant” [3]. On December 16th however, that relationship changed dramatically, as traffic to the Pinboard site increased by an order of magnitude within minutes — given the timing, presumably by panicked Delicious users. It’s unclear what percentage of former Delicious users migrated to Pinboard, or if that percentage rose to a level that would even raise the eyebrows of anyone at Yahoo. That migration was a significant boon for Pinboard, however, as it dramatically increased the number of Pinboard’s paying users.

The Great Delicious Exodus, as Cegłowski dubbed it [3] is an example of a phenomenon that has, to date, been almost completely undocumented: the migration of online communities. Online communities have existed for decades, and some of these communities have no doubt migrated from one service provider to another, but to date this has been largely hidden from view. It has been hidden from view only to the world at large, of course; these migrations are no doubt of great import to the communities involved in them. By and large, however, these communities have not documented these migrations, and few external reports exist.

Online communities have existed for decades, but humans have been migrating for probably as long as there have been humans. While human migration has not been systematically studied for quite that long, there is still an extensive body of literature on the subject. Scholars have identified a wide range of forces that motivate and govern migration, and many theoretical models have been proposed to account for migration. This paper will review some of these forces and models, and will discuss those that are relevant to investigating the migration of online communities. This paper will by no means be a complete treatment of this topic. Rather, this paper will bring together existing and well-established areas of study, to open the door to a new interdiscipline: the study of online human migration.

Online Communities

Any discussion of online communities is almost required by law to begin with Howard Rheingold’s 1993 book The Virtual Community: Homesteading on the Electronic Frontier. In this book, Rheingold both documents his personal experiences connecting to the WELL (Whole Earth ‘Lectronic Link) and participating in the online communities there, and discusses a wide range of issues around computer-mediated communication (CMC).

Most significantly here, Rheingold discusses the formation and social dynamics of online communities. The term he chooses to describe these communities — “Grassroots Groupminds” — says a great deal about the way these early online communities operated and members perceived these communities. While networked teleconferencing has existed since the late 1960s in government and research communities, over time these technologies became increasingly available to the public, and communities of interest formed. These communities of interest were able to develop their own social conventions, unique to their context — where context is the scope, problem, interest, or what have you, agreed upon by the community itself.

Rheingold emphasizes the volunteer nature of the WELL and other early online communities. He argues that the combination of free (much of which would now be called open source) software and contributed time and effort by users led to an environment amenable to free expression and experimentation — with all the pros and cons of free expression and experimentation familiar from the physical world. Social contracts, social norms, and forms of social capital emerged in these online communities, both by design and organically.

Rheingold, and much of the work that followed him, focuses on the online communities themselves, but largely ignores the online environment in which those communities exist. Rheingold discusses the WELL extensively, as well as other online communities in the United States and Europe — but he does not address the question of why those particular online communities formed in those particular online environments. What was it about the WELL and these other communities that attracted and fostered these communities? The earliest online communities may have sprung up on the WELL because it was the only game in town, but what about later communities, when there were other available options? And did any communities migrate from the WELL to other online environments, or vice versa? History does not record.

Online Communities for Migrants

An extensive, and highly interdisciplinary, body of literature exists on migration. While little attention been paid to the migration of online communities, a great deal of work has been conducted on the inverse: the use of the internet and other communication technologies among migrant populations. These populations are generally referred to in this body of literature as being “in diaspora.” Defining the term diaspora is challenging, and one that many authors undertake, with differing foci and agendas. But briefly, and over-simplified: a population in diaspora is one that is living in a region of the world away from its “native” land (e.g., Turks in the Netherlands [4], Palestinians in camps throughout the Middle East [5]), and self-identifies as distinct from other populations in this region [6]. Some authors also suggest that the concept of diaspora implies that the migrant population is currently or has been on the receiving end of colonization by an imperial power [7].

A great deal of this literature focuses on communication technologies, and the uses to which they are put by individuals in diasporic populations, with the goal of maintaining both connections to the community left behind, and a sense of identity as part of that community [8], as well as to integrate into the culture of their new home and to develop a new identity [9]. Much of this literature focuses on consumption and production of mass media: primarily radio, television, and print [10,11]. Some of this work focuses on what Dayan [12] refers to as “smaller” media (p.22), as opposed to mass media: newsletters, audio and video cassettes, and personal media such as home videos and letters. Discussions of “small” media inevitably gave way to discussions of email, mobile phones, and SMS for maintaining personal communication and contact [13,14]. Other communication technologies that have been used by populations in diaspora, that are particularly relevant to the current paper, are newsgroups [15], and private websites [6]. In short, perhaps every communication technology that exists, has been used at some point by some diasporic population, for maintaining interpersonal contact or for community-building and -maintenance.

We have here only scratched the surface of this extensive body of literature on media and technology use by diasporic communities. This literature demonstrates that extensive research has been conducted on the development and dynamics of mediated communities — and the media in question may be anything from newspapers to text messaging — that serve migrant and diasporic populations in the physical world. But, as stated above, the inverse is not the case: little research has been conducted on the migration of online communities.

Migration of Online Communities

Of the little scholarship that does exist that directly addresses the movement of populations between online services, danah boyd’s work on the evolution of online social networks stands out. In 2006, boyd documented the decline of Friendster and the rise of MySpace. Friendster’s decline, boyd argues, was a result of two factors: slow servers, and its adoption by “mainstream American 20/30-somethings.” The migration of users from Friendster to MySpace, in other words, was the result of both a failure of the network infrastructure, and a change in the social makeup of users.

boyd also discusses race and class divisions among American high school students playing out in Facebook and MySpace [16,17]. boyd’s writing on this topic is not primarily concerned with the movement of demographic groups of high school students from one online social network to another. Rather, she observes that the demographic makeup of users of these social network sites (SNSs) reflects pre-existing race and class divisions in American society. The movement of users from one SNS to another is therefore arguably a symptom of those divisions, and serves to reinforce those divisions — as (to use boyd’s example [17]), urban-to-suburban “white flight” was a symptom of and reinforced pre-existing race and class divisions in 20th century American society.

As online communities have become increasingly commonplace, a great deal of literature has been published analyzing the motivations of users who participate at different levels of involvement [18,19]. Of particular importance for the long-term health of online communities is the issue of how to convert “lurkers” into participating users [20]. The identification of a lifecycle of an individual’s participation in an online community naturally gave rise to the identification of a lifecycle of online communities themselves. Iriberri & Leroy [21] identify the stages of the lifecycle of online communities as follows: inception, creation, growth, maturity, and death. The “death” of an online community can occur when there is an insufficiently large group of participants to sustain the activity of the online community; Brandtzæg & Heim [22] identify a variety of reasons why users may choose to discontinue their participation in an online community.

This section has been an overview of the literature about online communities. We now turn to the literature on migration of populations in the physical world, to identify concepts and models from this area of study that may apply to the study of migration of populations online.

The Study of Migration

Human migration has been an object of study for well over a century. Scholars have identified a wide range of forces that motivate and govern migration, and many theoretical models have been proposed to account for migration. Consequently, it is not possible to present a comprehensive overview of the subject of human migration here. Such a task would require a book, or more likely, several. Instead, this section will engage in significant cherrypicking: only models of human migration in the physical world that are most relevant to migrations online will be discussed.

Economic Migration

Economic migration is probably the most familiar type: migration for the purpose of seeking improved employment prospects. Indeed, economic migration is so common and familiar that it is often referred to simply as “immigration” [23]. It could be argued that most of the history of the United States is a history of economic migration, and many regions of the world are currently experiencing major demographic changes due to increasing rates of immigration and emigration [24]. Of all forms of migration, economic migration is perhaps of the most concern in the arena of politics and policy-making [25].

Much early work on migration is economic in nature. Massey et al. [26] review decades of literature on migration, and articulate a macro and a micro theory of neoclassical economics as explanations of migration. According to the macroeconomic theory, migration (both international and within a single nation) is caused by differences in wages: stated overly simply, workers will “flow” to the location where they can receive the highest rate of pay. The microeconomic theory addresses migration from the perspective of the migrants themselves, rather than as simple pawns of larger economic forces. To again oversimplify: individuals make a cost-benefit calculation, and will move when the benefit of migration, in terms of increased expected earnings minus the costs of the migration itself, exceeds the benefit of staying put [27]. Migrations of populations are thus, according to theories based in neoclassical economics, the simple aggregate of many individual moves, made on the basis of individual decisions.

An important distinction must be made here, between the individual and the population of which that individual is a part. This is precisely the distinction between the microeconomic and macroeconomic theories of migration: the individual worker is the unit of analysis of the former, the population of workers of the latter. For the purposes of model-building and theory development, it of course is fruitful to use units of analysis larger than the individual. But it must be pointed out that (short of force, which will be discussed below) entire populations do not migrate; individuals and families do. For example, many years ago, the author of this paper moved from New York state to North Carolina. As it turns out, for many years New York has ranked among the top states in the nation in out-migration, and North Carolina among the top in in-migration [28,29]; thus the author was part of two of the largest migrations currently ongoing within the United States. The author did not make this move because of the larger migratory population, however, and indeed did not even know about these trends at the time; rather, the move was for entirely personal and family reasons.

Family is of course one of the most important reasons that individuals migrate. Massey et al [26] articulate a “new economics of migration” (p.436), which suggests that moves are not individual decisions, but are rather decisions made by families or households. While the microeconomic model emphasizes the “rational actor,” this model emphasizes the larger economic unit of the family: an individual may take a risk in moving to a new labor market, but a household is in a position to mitigate risk by spreading the labor of its members around.

The new economics of migration is based in part on social capital theory, and “family” is, in part, a stand-in for the larger construct of “social network.” Massey & Espinosa [30] argue that individuals are more likely to become migrants if they have a social network (family or friends) at the destination, “since these connections can be used to acquire information and assistance that reduce the costs and risks” of migrating (p.951). Massey [31] suggests that, further, migration “tends to feed back on itself through social channels” (p.68): as more individuals migrate, social networks form at the destination, reducing the costs and risk to later migrants. This feedback mechanism has the effect of increasing the likelihood of migration over time, to the extent that migration “becomes progressively independent of the economic conditions that originally caused it” (p.68). Further, after individuals migrate, they are more likely to remain at the destination if they build their own social networks: e.g., wives migrating after their husbands, children born at the destination, developing friendships, etc. (p.988).

The driving force in neoclassical economic theories of migration is wages, and wages are arguably absent as a consideration in the migration of online communities. To be sure, wage-earning is a significant force on the internet, and online communities exist in which wage-earning is allowed (e.g., World of Warcraft), or is even the primary purpose of the community (e.g., e-Bay). It seems unlikely that there are no cases of migrations between online communities caused by issues around wages. It is easy to imagine such a case: e-Bay, for example, might institute a policy that restricts the market for a particular good, and the community of buyers and sellers of that good migrate to, for example, craigslist. However, cases such as this have not been documented, and so cannot be discussed here. Far more relevant to the migration of online communities is the new economics of migration that takes social capital theory into consideration, acknowledging that the migration of individuals is explained as much by those individuals’ social networks as by any motivating push or pull factors. Ultimately, one uses the social networks that one’s friends use.

Environmental Migration

One form of human migration that is particularly relevant online is environmental migration. According to the International Organization for Migration (IOM):

Environmental migrants are persons or groups of persons who, for reasons of sudden or progressive changes in the environment that adversely affect their lives or living conditions, are obliged to have to leave their habitual homes, or choose to do so, either temporarily or permanently, and who move either within their territory or abroad.

Environmental migration is common in the physical world, but is hotly debated, even among scholars of migration. This is evident in the wide range of terms used to describe migration and migrants motivated by environmental change: “environmental migration, climate change-induced migration, ecological or environmental refugees, climate change migrants and environmentally-induced forced migrants” [32]. Piguet et al. [33] provide a brief history of the fall and rise of “environmental factors” in theories of migration, arguing that these were displaced by political and economic interpretations of migration during most of the twentieth century, to re-emerge in the late 1980s and 1990s as international awareness of climate change grew internationally (pp.2–3). Climate change itself is, of course, highly politically and economically charged, and so, as Dun & Gemenne point out, environmental theories of migration cannot be separated from political and economic theories.

While this debate among migration scholars is fascinating in showing the intellectual evolution of a highly politically-charged field, it is unclear how much of the terminology employed in this literature is relevant here: the terms “climate change” and “refugee,” for example, seem not to make much sense in the online context. Nevertheless, we suggest that environmental migration is a useful model for discussing online migration, and that the constructs around environmental migration are appropriate to apply to online environments. This, however, requires that the characteristics of an online service that can be considered “environmental” be articulated. We will begin to articulate these characteristics below.

Trolldalen et al. [34] suggests that “a working definition of ‘environmental refugee’ should have two aspects”: coercion and external displacement. In other words, environmental refugees are “forced to leave their homes for environmental reasons,” and in doing so cross international boundaries. Environmental migration studies refer to the causes forcing migration as push factors: any factors that make it difficult or impossible to continue to live in a particular environment. A pull factor, by this conceptualization, is simply “a relatively safer environment” [35] (p.84).

Several categorizations of different aspects of environmental push factors have been articulated in the literature. Black [36], while rejecting the validity of the very construct of the “environmental migrant,” nevertheless provides a useful overview of these categories. Black (pp. 1–2) cites Jacobson’s [37] three categories of the permanence of environmental migration: temporary displacement due to natural or man-made disasters (e.g., earthquake, oil spill), migration over the long-term due to environmental degradation that interferes with livelihoods (e.g., drought), and permanent emigration due to environmental degradation that makes the environment permanently unlivable (e.g., desertification) (p.37–38). Laczko and Aghazarm [38], on the other hand, categorize environmental migration by the speed of the environmental change: sudden disasters (e.g., natural disasters such as tsunamis) vs. “slow-onset” disasters (e.g., climate change such as droughts and desertification). Yet a third dimension is proposed by The International Organization for Migration: natural vs. man-made causes of migration (e.g., earthquakes vs. industrial accidents). It is difficult, both environmentally and politically, to categorize certain types of environmental change according to these categories: for example, is desertification natural or man-made? Other authors therefore categorize environmental migration by more specific motivating environmental changes, as Suhrke [39] does in articulating “the five most common forms of environmental degradation” (p. 11).

Forms of environmental degradation in the physical world are of only limited relevance to online environments. An earthquake may cause a server to go offline, to be sure, but a drought is unlikely to have much effect on an online service. Other categorizations are more obviously relevant to migration in online environments: the speed of the factor leading to migration, whether the factor leading to migration is natural or man-made, and the temporariness or permanence of the migration. It is an open question which of these categorizations will be most fruitful in analyzing migrations online. Identifying what forms of “environmental degradation” even may be said to exist online will also be fruitful, from the natural disaster of an earthquake destroying a server farm, to the man-made denial-of-service attack. There may even be a third form, a hybrid of natural and man-made: for example, a hacking attack (man-made) that exploits a vulnerability in the server software (natural, in the sense that the server software is part of the environment in which the service operates).

Forced Migration

Another form of migration that is no doubt familiar to the reader is forced migration. Forced migration of course includes the all-too-familiar “displacement” caused by ethnic cleansing and other forms of persecution. Forced Migration Online [40] defines forced migration more broadly, however, to include conflict, development, natural disasters and environmental change, and human-made disasters as factors that may force migration.

To be sure, persecution exists online. Cyberbullying in particular is recognized as a significant issue for grade school-aged children, and as a result has become possibly the most familiar form of online persecution. Like more “traditional” forms of bullying, however, cyberbullying is generally targeted at individuals, rather than at groups or populations [41]. An extreme example of online persecution targeted at groups or populations is a program called Genetic Access Control, which requests data from the genetic testing tool 23andMe, and enables websites and apps to block a user’s access based on any genetic characteristic (gender, ancestry, etc.). Fortunately, Genetic Access Control was apparently never implemented, as 23andMe, to its credit, rapidly revoked permission for to access its data [42]. Despite these examples of online persecution at the individual and group level, however, we do not know of any examples of forced migration due to online persecution. It is certainly possible to imagine such a situation, however: an online community that, over time, sees a change in the makeup of its user community and becomes unfriendly or even hostile to a specific subcommunity, which then migrates to a new service provider or venue. It seems likely that such a situation has arisen, probably more than once, in the history of online communities — but if so, it has not been documented.

Instead, online communities tend to be self-restricting: witness the thousands, perhaps millions of groups on Facebook and Reddit and other social media, that cater to individuals with specific interests. These “communities of interest” tend to be joined, obviously, by those with an interest in the topic, and avoided by those without. In this sense, online communities tend to resemble physical neighborhoods, which are often fairly homogeneous demographically, due to homophily [43]. Of course there is a wide range in the demographic makeup of physical neighborhoods, but even diverse neighborhoods attract residents who value diverse neighborhoods, so there is a certain degree of homophily even there. Also like physical neighborhoods, the diversity of — and tolerance of diversity in — online communities can, at least to some extent, be manipulated by agreements and policies. Newell, et al. [44], for example, document a migration away from Reddit upon the adoption of “more restrictive content policies to protect users from harassment, hate speech, and offensive content.” User communities that engaged in hate speech — or, from their point of view, “who believe that freedom of speech should be paramount” — left Reddit for other online communities.

All that said, however, forced migration due to persecution is not of particular relevance to the current discussion. Of more relevance to the current discussion are other factors that may force migration: conflict, development, and disasters. Forced migration due to environmental and man-made disasters has been discussed above. Some suggest that areas particularly vulnerable to, and not resilient to climate change become “hotspots” that then breed conflict [45]. It is thus difficult to separate forced migration due to environmental factors and forced migration due to conflict. It is clear, however, that conflict is a factor that may force migration. In the physical world, that conflict is often between populations competing for the same physical locations or other scarce resources; online, it is not clear what resources could be scarce enough to prompt conflict. Conflicts online are therefore more likely to arise from factors other than scarcity. For example, Brandtzæg & Heim [22] identify conflict with other users and conflict with the administration of an online service as reasons why users may defect from a service.

More relevant to the discussion of online communities is development that forces migration. In a sense, development-induced displacement is the inverse of economic migration: economic migrants move, more or less voluntarily, to take advantage of opportunities elsewhere, but “development-displacees” [46] are forcibly moved by others so that those others may take advantage of opportunities. Those opportunities are generally for economic development: while there is a range of types of development projects, Stanley [47] identifies dams, urban infrastructure and transportation, and natural resource extraction as among the most important. Most often it is the state forcing the displacement, as in the case of China’s Three Gorges Dam project, or the construction of the Quabbin Reservoir in the author’s native Massachusetts — though large corporations, such as the oil industry, may also have the power to force displacement.

Development-induced displacement may occur at a variety of points in a project’s lifecycle. But displacement is due primarily to two factors: relocation to enable a project (as in the case of slum removal for urban renewal [48]), and consequences of the project (for example, environmental degradation, reduction or extinction of animal populations, human rights abuses [47,49]). Both of these factors may be relevant to online communities, though the definition of “development” is likely to be quite different online as in the physical world. A reduction in the number of servers allocated to a service, for example, may be the result of the service provider engaging in development on another project, but it may have the effect of increasing latency for users of the service, as in the case of Friendster, discussed above.

Analysis of Examples of Online Migrations

Having reviewed a few models of human migration in the physical world, we now turn to examples of actual migrations online. In this section, we will analyze these examples in light of the models addressed above.

Friendster to MySpace to Facebook

It’s probably unnecessary to explain that Friendster, MySpace, and Facebook are social network sites (SNSs); or that SNSs provide functionality to enable users to construct their own social network online [50] in which individuals are nodes and “friendship” are edges. The resulting network of connections between nodes displays regions of higher density that may be identified as communities [51]. SNSs are thus not communities, but rather collections of communities. With billions of users, it is unreasonable to call Facebook a community; even with their smaller user bases, the same can be said of Friendster and MySpace.

boyd describes migrations of one of these communities (or, more accurately, many related communities) from Friendster to MySpace [52], and from MySpace to Facebook [17]: communities of teen users. As discussed above, boyd attributes teens’ departure from Friendster to its “mainstream-ification and slowness”: the discovery of Friendster by “mainstream” users (what boyd refers to as “hegemonic society”) turned off many of Friendster’s early adopters, and a slow-down of Friendster’s servers “made it very difficult (if not impossible) for mainstream users to engage.” Upon leaving Friendster, however, users could have started using any of several SNSs that existed at the time, or stopped using SNSs altogether. boyd [52] goes on to discuss the reasons that many users migrated to MySpace in particular: features such as flexibility in creating a profile and a messaging system, and the effort by MySpace to promote events. In other words, MySpace enabled the creation of “subcultural capital” for its users, both through its functionality and the activity of its management.

The migration from MySpace to Facebook was due to an entirely different set of factors. boyd [17] attributes this migration to issues of race and class. In her 2007 essay on this topic [16], boyd expressed concern that addressing issues of race and class, especially for an American audience, “would piss many off,” but that the topic was “too critical to go unacknowledged” (¶3). In her 2011 chapter [17], however, she stated quite baldly that “white and more affluent individuals were more likely to choose and move to Facebook” (p.204). Like all issues wrapped up with the controversial issues of race and class, this MySpace-to-Facebook migration is not so simple: boyd argues that “taste and aesthetics,” and “the network structures of teen friendship” (p.204) — all of which are, of course, related to race and class — were also important factors. But fundamentally, the MySpace-to-Facebook migration was a symptom of pre-existing social divisions among teenaged users in the United States. boyd used the term “white flight” in the title of her 2011 essay very deliberately.

The factors contributing to the Friendster-to-MySpace and MySpace-to-Facebook migrations make for an interesting comparison. In the former case, environmental degradation played a major role. In the latter case, white flight was the dominant factor.

It is fairly straightforward to make an analogy between environmental degradation in the physical world and online, provided that one accepts the premise that an online service is an environment. As discussed above, factors that affect the servers on which a service runs may be said to be “natural disasters”: servers slowing down or dropping off the internet entirely has the effect of making the service “uninhabitable.” Desertification may be the best analogy for servers slowing down: users may be able to eke out an existence on the service, but fewer and fewer over time, and over time more and more users will move to a more hospitable service provider. The Friendster-to-MySpace migration thus displays push and pull factors, just as a migration in the physical world would: push factors, as in the physical world, are those that make it difficult or impossible to continue to live in a particular environment. Kniveton et al.’s [35] (p.84) definition of pull factors is “a relatively safer environment.” Safety may not be the primary consideration for online users; rather, “a relatively more hospitable environment” might be a better definition of pull factors — both online and in the physical world. A wide range of factors may then be said to make an environment hospitable.

The MySpace-to-Facebook migration, on the other hand, was the result of pre-existing social divisions playing out online. In the physical world, white flight is a form of environmental migration, though one in which the social environment is more important than the physical environment. Studies of social networks and “human ecology” in fact even refer to the “social environment,” as well as to narrower constructs, such as the “organizational environment” [43]. As boyd [16] observed, discussing issues of race and class (as any discussion of white flight must) is controversial. But squeamishness about these topics must be put aside in order to discuss the factors in physical-environmental migration that also apply to social-environmental migration.

White flight may be said to be a result of a perception of environmental degradation: “there goes the neighborhood.” Justified or not, residents move out of one environment because they “refuse[] to live near socially subordinated groups” [53] (p.25), and move into another environment that is deemed more desirable. In studies of white flight in the United States, the environment being fled is often urban city centers, while the environment being moved into is often suburban housing developments. An interesting feature of this form of environmental migration is that it is by choice, rather than by necessity. Natural or man-made disasters, whether sudden or slow-onset, force migration. White flight is, in a sense, the opposite of a forced migration: urban city centers do not become unlivable, they merely become perceived to be undesirable by a certain segment of the population (usually middle-class whites). Similarly, MySpace did not become unusable (as Friendster did, with its slow servers), it simply became perceived to be undesirable for certain communities of teenaged users — which more or less overlapped with middle-class whites.

The Friendster-to-MySpace migration also contained an element of white flight, though it was not necessarily the white users who fled. Prior to that, however, there was an element of gentrification in this migration. Gentrification is a form of forced migration due to environmental degradation, though again, it is not the physical environment that is degrading, but rather the social environment. Through their own efforts to make an (usually urban) environment more liveable, “first-stage” gentrifiers increase the perceived value of living in the environment, thus attracting “immigrants” (not in the usual sense of this word, but simply new populations of residents) who displace the first-stage gentrifiers [54]. In an urban area where housing is already built, housing is a zero-sum game: for someone to move in, someone else must move out. Thus, “immigrants” displace first-stage gentrifiers in the finite space of available housing. In most gentrification, however, the cost of housing also increases as new populations, often of greater economic means, move into the environment. Insofar as the cost of housing can be considered part of the environment, an increase in the cost of housing is a factor that will push some populations out of that environment.

To be fair, gentrification is an awkward metaphor for the Friendster-to-MySpace migration (and for online environments generally), since unlike an urban area, Friendster was not in some kind of state of urban decay prior to the arrival of its first populations. Indeed, a more accurate term for these users would be simply “early adopters” [55]. According to boyd [52], “Friendster’s early adopters were 20-somethings” (Social technologies succeed when they fit into the social lives and practices of those who engage with the technology section, ¶3), in particular, “freaks, geeks and queers” (People use the social technologies that all of their friends are using section, ¶1). It is worth doing a certain amount of violence to a metaphor, however, to point out the crucial fact that the early adopters of Friendster were demographically similar to “first-stage” gentrifiers of urban areas [56].

As in urban gentrification, these early adopters of Friendster were displaced by mainstream users. And here is where white flight becomes a factor. “Housing” online (or rather, being a user of an online service) is not a zero-sum game: with the addition of network infrastructure, Friendster, like any online service, could have accommodated both old and new user communities. But the early adopter communities believed that the influx of new users degraded the environment and made it undesirable to stay. The “mainstream-ification” of Friendster was, like the MySpace-to-Facebook migration discussed above, an un-forced migration due to a perceived degradation of the social environment. (Though to be fair, in the end, the servers slowing down may have forced migration by making Friendster uninhabitable.)

The Friendster-to-MySpace and MySpace-to-Facebook migrations may have been un-forced, but “once mass departure began… it spiraled quickly” [52] (People use the social technologies that all of their friends are using section, ¶1). Stutzman [57] draws a distinction between “ego-centric” and “object-centric” networks: an ego-centric network is one in which the primary network node is the individual (e.g., SNSs) while an object-centric network is one in which some object other than the the individual is the primary network node (e.g., photos on Flickr). Stutzman suggests that an ego-centric network “has limited core-value — its value is largely in the network — making it highly susceptible to migration” (¶3). These migrations were rapid because there was little to keep users on one SNS, once members of their online community started migrating themselves. Among diasporic populations, as discussed above, communication technologies are used partly to maintain connection to the community left behind. In these cases, the community migrated together precisely because it was possible for technology to aid in maintaining that connection. Stutzman points out that “the main chore of migration [is] network reestablishment, a chore made ever-simpler as the migration cascade continues” (¶3). In other words, there is little cost for the individual user in moving service providers for an ego-centric network, except that one loses previously-established connections to one’s social network. This is a greater cost for the individuals in the vanguard of a migration, which is why, as boyd points out, the “mass departure began with a few pissed-off folks”: those in the vanguard had to be highly motivated to incur that cost. For those migrating later, it is easier and faster to re-establish those connections, as many members of their networks will already have re-established connections.

The Answerbag Diaspora

Answerbag (answerbag.com) is a social question answering (Q&A) site. Shah, Oh, & Oh [58] define social Q&A as an online “service that involves (1) a method for a person to present his/her information need in natural language…, (2) a place where other people can respond to somebody’s information need, and (3) a community built around such service based on participation” (p.206). Search engines exist at one end of the information seeking spectrum, with keyword searching and little or no community involvement; social Q&A exists at the other end of the spectrum, with information needs expressed as questions, and a dedicated community of users. Answerbag was one of the earliest social Q&A sites, though social Q&A sites have since proliferated. Some names in that space are likely to be familiar to the reader, including Yahoo! Answers, Quora, and Stack Exchange.

Gazan [59] describes Answerbag as “a social Q&A site designed around a one question-multiple answers architecture. Users submit and rate questions and answers, and the highest-rated answers are listed first, providing a collaborative filtering function that still allows people to view the full range of answers to any question. Users may also interact via comment threads beneath any answer, create personal profile pages, and ‘friend’ one another to be selectively notified about their friends’ activity on the site” (p.695). In other words, around the core functionality of question-answering, Answerbag has implemented functionality to enable a community to develop.

In December 2009, Answerbag underwent a significant site redesign, which led to the defection of a significant percentage of the site’s users. Gazan [60] describes this as “the Answerbag Diaspora” (p.2854). Gazan states that the redesign of Answerbag was launched to accomplish the following: add user-requested functionality, update the site’s interface (at that point unchanged for 3 years), and reduce spam. Upon the launch of the newly-redesigned site, however, several key pieces of functionality were broken, and the new interface “yielded the impression that the site had far less activity than before” (p.2850). In the wake of this redesign, 60% of Answerbag’s active users left the site, most for good. Many of these users established communities on other social Q&A sites as well as Facebook (p.2854).

On the face of it, the Answerbag Diaspora seems to be another case of environmental migration: the site redesign made the environment, if not uninhabitable, then undesirable. However because this environmental change was a side effect of a development project, this migration also bears characteristics of a forced migration.

As discussed above, forced migration includes development-induced displacement, and this displacement may occur to make way for the development or as an effect of it. The Answerbag user community was apparently quite active right up to the launch of the redesigned site; it was the effects of the redesign that led to migration. Gazan writes that these effects included “the inability to communicate with friends” and “lack of access to collective content” (p.2851): breakdown of established lines of communication, and an inability to access the community’s collective memory. In other words, many of the aspects of Answerbag that made it a community were damaged or eliminated. These were not necessarily intentional features of the Answerbag platform: comment threads enabled interpersonal communication, but threads became more than that as the community adopted some of them as part of the community’s collective memory. As with the WELL, the features of the Answerbag software environment enabled the development of aspects of the social environment. When that software environment was changed, it (perhaps inevitably) affected the social environment. Answerbag’s redesign changed — many users probably would have said degraded — the social environment to such an extent that it became undesirable to remain in that environment. In this sense, the Answerbag Diaspora was similar to the Friendster-to-MySpace and MySpace-to-Facebook migrations: migration due to degradation of the social environment. The critical difference, however, is in the cause of that environmental degradation: a development project changing the landscape, so to speak, in which that social environment existed, rather than a change in the demographics of the population.

The redesign of the Answerbag site also included making static content more visible, where previously there had been dynamic content. This led to a perception among users that there was a lack of activity on the site. This in turn may have given the impression that others were defecting from the site. (Gazan [60] includes a screenshot of one rather pathetic Answerbag post: “I feel so alone on this site now. Do you?” (p.2852).) boyd [52] wrote that “once mass departure began… it spiraled quickly”: she was discussing the Friendster-to-MySpace migration, but she might just as well have been discussing the Answerbag Diaspora, which took place over the span of a mere 1–2 weeks. Also like the SNS migrations, “many of those who had chosen to leave Answerbag actively convinced their friends to do likewise” ([60], p.2853). There is one significant difference between Answerbag and SNS services, however, and that is the nature of the network. As discussed above, an SNS service is an “ego-centric” network. Answerbag, on the other hand, is an “object-centric” network, in which the social network is layered on top, so to speak, of objects — those objects being questions and answers. Stutzman [57] writes that object-centric networks are less susceptible to migration than ego-centric networks, as the object-centric network has “core value” in the objects themselves, and the service can potentially exist without the social network. The Answerbag Diaspora, however, demonstrates that even an object-centric network is susceptible to migration when there is a strong social network layer, and perhaps also when the objects in question are ephemeral. This raises the question of where the core value of a network really lies, and how to weight the factors that affect users’ choices to discontinue their participation in an online community.

The Great Delicious Exodus of 2010

The title of this paper was taken from a post to the Pinboard Blog titled “Anatomy of a Crushing” [3] in which Maciej Cegłowski, the developer of Pinboard, documented the history and technical details of the rush on Pinboard that followed the announcement about Delicious, narrated at the start of this paper.

The Great Delicious Exodus began with the announcement of the “sunsetting” of the Delicious service. This is an important point, one where this story both adheres to and diverges from models of migration in the physical world. If Delicious had been shut down, or an announcement had been made of an imminent shutdown, this would be closely analogous to a natural disaster: an unexpected event that suddenly makes the current environment unlivable, by destroying property or even the environment itself. The property, in this case, being the bookmarks that users had stored with Delicious, and the environment being the Delicious service itself. An immediate shutdown of Delicious would have perhaps been analogous to a sinkhole opening up and swallowing an entire village; an imminent shutdown would have perhaps been analogous to a hurricane, where the effects can be devastating but those affected have some advance warning. (It’s worth noting that sinkholes swallow entire villages on the internet all the time… though usually those villages come back: sites go down for any number of reasons, from maintenance to maliciousness. As Cegłowski writes, “sites like Daring Fireball or Slashdot… are notorious for crashing the objects of their attention” ([3], ¶7).)

But a shutdown is not what happened. Delicious was not shut down immediately, nor was any timeframe given for when it would be shut down, nor was it even clear that it would be shut down at all. Rather, the mere threat of a natural disaster was enough to cause a migration. This is an important difference between migrations in the physical world and online. In the physical world, moving is a major undertaking. There are many risks and costs in the process of moving: the dangers of physical travel, the cost of cutting oneself off from one’s existing social network, the risk of moving to a possibly unknown and uncertain destination. Online, moving service providers is not without risks and costs, but is considerably lower-risk and lower-cost. There is no physical travel, and as one website is as accessible as any other, the costs are only those associated with bandwidth consumption (once the initial hurdle of access to the internet is cleared, of course). Because of the user’s ability to access the new site (even if not as a subscribed or paying user) prior to becoming a full-fledged user, as well as the existence of online communities in which discussion of the service has taken place, the individual moving may know as much as it is possible to know about the service she is moving to.

As discussed above, however, one significant risk in moving online service providers is in cutting oneself off from one’s existing network. As Stutzman [57] points out in his discussion of “ego-centric” and “object-centric” networks, “the main chore of migration [is] network reestablishment, a chore made ever-simpler as the migration cascade continues” (¶3). In other words, the cost to moving service providers for an ego-centric network is that one loses previously-established connections to one’s social network. This is more of a risk and a cost for early adopters: as more of one’s social network moves to a new service provider, it is easier and faster to re-establish those connections. Delicious and Pinboard, however, are object-centric networks: web bookmarks, rather than individuals, are the primary network node. Stutzman argues that object-centric networks are less susceptible to migration than ego-centric networks, since “object-centric social networks offer core value, which is multiplied by network value.”

If object-centric networks are less susceptible to migration than ego-centric networks, it must have been a powerful natural disaster to cause such an exodus. Except that, as has already been established, there was no natural disaster. One therefore has to ask, what caused this exodus that saw, as Cegłowski somewhat hyperbolically writes, “every Delicious user in North America race[] for the exit” ([3], ¶2)?

Another way to look at the Great Delicious Exodus is as a development-induced displacement: the equivalent of the state (Yahoo!) announced that it would be removing a large number of properties (as Brazil has in removing favelas [48]). The difference, of course, is that this removal was not immediate or even imminent; it was the mere threat of removal that caused the exodus. It was as if the Brazilian government had announced that it would be removing a favela eventually, and this announcement was enough to cause a migration. And indeed, this may be the case in the physical world as well: some percentage of a population, perhaps the least risk-averse or the most dissatisfied with current conditions, may require no more than the threat of change to prompt migration. While the literature on migration and disapora discusses motivating factors, not even the literature that treats migration as an individual decision addresses what ultimately prompts an individual’s decision to migrate. This question of what pushes an individual “over the edge,” so to speak, would be an interesting area for future “microlevel studies” of migration [61].

The reestablishment of one’s social network is a cost in migrations both online and in the physical world. It is easy to imagine, however, that this might be easier in the context of an online service than in the physical world: reestablishment can be aided by functionality of the online service, in ways that are difficult or impossible in the physical world. Take, for example, Facebook, which actively attempts to aid in expanding users’ networks by suggesting “People You May Know,” based on users’ existing connections. Thus the bar for making migration decisions online may be lower — perhaps quite a bit lower — than for making migration decisions in the real world.

Upon Yahoo’s announcement, many users began actively seeking alternative bookmarking services, posting questions and seeking recommendations on social media and other discussion venues. Many sites published instructions for how to export saved bookmarks from Delicious [62,63]. It would be difficult to claim that these disaffected Delicious users formed a community — any more than a flashmob is a community — but it is not a stretch to call this a community of interest. And as members of this community of interest recommended Pinboard, it may have made the migration decision easier for other members. Thus, even though Pinboard is an object-centric network, even though the social network is actively de-emphasized — “social bookmarking for introverts,” after all — it was the social network that lowered the costs associated with immigration to Pinboard.

Indeed, Cegłowski himself made a significant contribution to lowering the social costs associated with immigration to Pinboard, by maintaining active communication with new users: as the Pinboard site slowed under the increased load of a massive influx of new users, Cegłowski writes, “my wrists burned from typing reassuring emails to nervous new customers explaining that their bookmarks, would, in the fullness of time, actually show up on the big blank spot that was their homepage” ([3], ¶30). In other words, even while maintaining the functionality of the site — maintaining the object-centric network — Cegłowski was simultaneously maintaining the ego-centric network. Thus it seems that migration online is motivated heavily, perhaps even primarily, by social factors. Object-centric networks may be less susceptible to migration than ego-centric networks, but when there is a migration decision to be made, it is the social network that proves to be the decisive factor.

Discussion

The migration of online communities may be understood in many ways, as migration in the physical world is. Only a few models of human migration were reviewed in this paper, those that are most relevant to migrations online. This paper demonstrates, however, that some of the forces and models that have arisen from analyses of physical migrations are also fruitful for analyzing the migration of online communities.

Naturally, other methods of analyzing online migrations are possible. These methods may come from studies of physical migrations of human populations, as in this paper, or from other disciplines. For example, the study of customer defection [64] seems a natural fit for understanding why users emigrate from a service, as is relationship marketing [65] for attempting to prevent users from doing so. The study of the diffusion of innovations, specifically the characteristics of adopters and of innovations, and the factors that influence adoption decisions [55,66] would likely be fruitful for understanding why users immigrate to a service.

There is also a great deal of literature on human migration that has not been addressed in this paper. In particular, models of cultural assimilation (the process by which an immigrant population changes in a new environment), and cultural appropriation (how a population is changed by external influences) could be fruitful.

We do not have to look far for an example of a population changing the online environment into which they immigrated. In the wake of the Great Delicious Exodus, the fandom community arrived on Pinboard. Cegłowski [3] documents this migration, and some of the short-term effects it had on the service. One of the almost immediate actions of the fandom community was to develop what Cegłowski calls an “epic feature spec” (¶11), articulating the community’s feature requests. This was not cultural appropriation by the pre-existing community, so much as it was an immigrant population demanding a voice, but the effect was similar: a change to the online environment. The complexity of the fandom community’s feature spec also seems to fly in the face of the “cultural simplification that often attends concerted human migration” ([67], p. 14). The arrival of the fandom community on Pinboard would make an interesting case study, to explore the extent to which models of cultural change among migrant and native populations in the physical world apply to online migrations.

As mentioned above, very little literature exists that documents migrations between online environments. This stands in stark contrast to the extensive literature on migrations in the physical world, which spans multiple disciplines. In order for the study of online migration to mature, it is necessary that more online migrations be documented. Without such a rich set of case studies, it will be difficult or impossible to develop the detailed models of online migration that exist for physical migration. Furthermore, as the online environment increasingly becomes as much a part of the “real world” as any physical location [68], adding documentation of online migrations to the extensive literature on migrations generally will enable models of migration to be tested and refined with regard to this “new” environment.

As the online environment increasingly becomes — and is recognized to be — part of the real world, an understanding of the factors that motivate online migrations, and theoretical models of such migrations, will become increasingly important. Scholars of social media have already begun to discuss this phenomenon, but this work is still in its infancy. The online world is highly dynamic; in order to fully understand it, we must come to grips with that dynamism. Part of that dynamism is the migration of communities across online environments and services. The author hopes that this paper has opened the door to a greater interdisciplinary understanding of these migrations.

References

1. Marcoullier E. Aaaaaand that’s all she wrote for Delicious and MyBlogLog. #EndOfAnEra /cc @joshu http://yfrog.com/h3z89p. Twitter; 2010. Available: https://twitter.com/bpm140/status/15473591558602752
2. Nine Fives Software. Pinboard: About The Site. Available: https://pinboard.in/about/
3. Cegłowski M. Anatomy of a Crushing. In: Pinboard Blog. 8 Mar 2011. Available: https://blog.pinboard.in/2011/03/anatomy_of_a_crushing/
4. Ogan C, d’Haenens L. Do Turkish women in the diaspora build social capital? Evidence from the Low countries. Ethn Racial Stud. 2012;35: 924–940. doi:10.1080/01419870.2011.628034
5. Aouragh M. Palestine Online: Transnationalism, the Internet and the Construction of Identity. Reprint edition. London: I.B.Tauris; 2012.
6. Androutsopoulos J. Multilingualism, diaspora, and the Internet: Codes and identities on German-based diaspora websites. J Socioling. 2006;10: 520–547. doi:10.1111/j.1467–9841.2006.00291.x
7. Dasgupta S. Language, Identity and Culture: the Issue of Multi-Culturalism, E-Diaspora & Global Culture Networks with Special Reference to Call Centres in India. IPEDR. 2012;33: 21–25. Available: http://www.ipedr.com/vol33/005-ICLMC2012-L00012.pdf
8. Brinkerhoff JM. Digital Diasporas: Identity and Transnational Engagement. 1 edition. Cambridge ; New York: Cambridge University Press; 2009.
9. Nguyen T, Cunningham S. The Popular Media of the Vietnamese Diaspora. Media Int Aust Inc Cult Policy. 1999;91: 125–147. doi:10.1177/1329878X9909100113
10. Karim KH, editor. The Media of Diaspora. 1 edition. New York: Routledge; 2003.
11. Bailey O, Georgiou M, Harindranath R, editors. Transnational Lives and the Media: Re-Imagining Diasporas. 1st ed. 2007 edition. S.l.: Palgrave Macmillan; 2007.
12. Dayan D. Media and Diasporas. Television and Common Knowledge. Psychology Press; 1999. pp. 18–33.
13. Hiller HH, Franz TM. New ties, old ties and lost ties: the use of the internet in diaspora. New Media Soc. 2004;6: 731–752. doi:10.1177/146144804044327
14. Barber PG. Cell phones, complicity, and class politics in the Philippine labor diaspora. Focaal. 2008;2008. doi:10.3167/fcl.2008.510104
15. Helland C. Diaspora on the Electronic Frontier: Developing Virtual Connections with Sacred Homelands. J Comput-Mediat Commun. 2007;12: 956–976. doi:10.1111/j.1083–6101.2007.00358.x
16. boyd danah. Viewing American class divisions through Facebook and MySpace. In: Apophenia. 24 Jun 2007. Available: http://www.danah.org/papers/essays/ClassDivisions.html
17. boyd danah. White Flight in Networked Publics? How Race and Class Shaped American Teen Engagement with MySpace and Facebook. In: Nakamura L, Chow-White P, editors. Race After the Internet. New York: Routledge; 2011. pp. 203–222.
18. Nonnecke B, Preece J. Why Lurkers Lurk. Americas’ Conference on Information Systems (AMCIS) 2001 Proceedings. 2001. Available: https://aisel.aisnet.org/amcis2001/294
19. Ridings C, Gefen D, Arinze B. Psychological Barriers: Lurker and Poster Motivation and Behavior in Online Communities. Commun Assoc Inf Syst. 2006;18. Available: https://aisel.aisnet.org/cais/vol18/iss1/16
20. Beamish A. Contributors and Lurkers: Obstacles to Content Creation in a Professional Online Community. In: Foster J, editor. Collaborative Information Behavior: User Engagement and Communication Sharing. First edition. Hershey PA: Information Science Reference; 2010. pp. 36–54.
21. Iriberri A, Leroy G. A life-cycle perspective on online community success. ACM Comput Surv. 2009;41: 11:1–11:29. doi:10.1145/1459352.1459356
22. Brandtzæg PB, Heim J. User Loyalty and Online Communities: Why Members of Online Communities Are Not Faithful. Proceedings of the 2Nd International Conference on INtelligent TEchnologies for Interactive enterTAINment. ICST, Brussels, Belgium, Belgium: ICST (Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering); 2007. pp. 11:1–11:10. Available: http://dl.acm.org/citation.cfm?id=1363200.1363215
23. Geddes A, Scholten P. The Politics of Migration and Immigration in Europe. Second edition. Thousand Oaks, CA: SAGE Publications Ltd; 2016. Available: https://us.sagepub.com/en-us/nam/the-politics-of-migration-and-immigration-in-europe/book234419
24. Zagheni E, Weber I. You Are Where You e-Mail: Using e-Mail Data to Estimate International Migration Rates. Proceedings of the 4th Annual ACM Web Science Conference. New York, NY, USA: ACM; 2012. pp. 348–351. doi:10.1145/2380718.2380764
25. Geddes A. Immigration and European integration: Towards fortress Europe. 2nd edition. Manchester: Manchester University Press; 2008. Available: http://www.manchesteruniversitypress.co.uk/9780719074660/
26. Massey DS, Arango J, Hugo G, Kouaouci A, Pellegrino A, Taylor JE. Theories of International Migration: A Review and Appraisal. Popul Dev Rev. 1993;19: 431–466. doi:10.2307/2938462
27. Todaro MP, Maruszko L. Illegal Migration and US Immigration Reform: A Conceptual Framework. Popul Dev Rev. 1987;13: 101–114. doi:10.2307/1972122
28. McMahon EJ. Empire State Exodus : Empire Center for Public Policy. Albany, NY: Empire Center; 2009 Oct. Available: https://www.empirecenter.org/publications/empire-state-exodus/
29. United Van Lines. 39th Annual National Movers Study. 2016 Jan. Available: https://www.unitedvanlines.com/contact-united/news/movers-study-2015
30. Massey DS, Espinosa KE. What’s Driving Mexico-U.S. Migration? A Theoretical, Empirical, and Policy Analysis. Am J Sociol. 1997;102: 939–999. Available: https://www.jstor.org/stable/2782024
31. Massey DS. The Social and Economic Origins of Immigration. Ann Am Acad Pol Soc Sci. 1990;510: 60–72. Available: https://www.jstor.org/stable/1046794
32. Dun O, Gemenne F. Defining “environmental migration.” Forced Migr Rev. 2008;31: 10–11. Available: https://www.fmreview.org/climatechange/dun-gemenne
33. Piguet É, Pécoud A, de Guchteneire P. Migration and Climate Change. Paris : Cambridge, UK ; New York: Cambridge University Press; 2011.
34. Trolldalen JM, Birkeland N, Borgen J, Scott PT, Coles G, World Foundation for Environment and Development, et al. Environmental refugees: a discussion paper. Oslo: World Foundation for Environment and Development : Norwegian Refugee Council; 1992.
35. Kniveton D, Smith C, Black R, Schmidt-Verkerk K. Challenges and approaches to measuring the migration-environment nexus. In: Laczko F, Aghazarm C, editors. Migration, Environment and Climate Change: Assessing the Evidence. Geneva: International Organization for Migration; 2009. pp. 41–112. Available: https://publications.iom.int/books/migration-environment-and-climate-change-assessing-evidence
36. Black R. Environmental refugees: myth or reality? Brighton, United Kingdom: University of Sussex; 2001 Mar. Report No.: №34. Available: http://www.unhcr.org/research/RESEARCH/3ae6a0d00.pdf
37. Jacobson JL. Environmental Refugees: a Yardstick of Habitability. Bull Sci Technol Soc. 1988;8: 257–258. doi:10.1177/027046768800800304
38. Laczko F, Aghazarm C. Introduction and Overview: Enhancing the knowledge base. In: Laczko F, Aghazarm C, editors. Migration, Environment and Climate Change: Assessing the Evidence. Geneva: International Organization for Migration; 2009. pp. 7–40. Available: https://publications.iom.int/books/migration-environment-and-climate-change-assessing-evidence
39. Suhrke A. Pressure points: environmental degradation, migration and conflict. Cambridge, MA: American Academy of Arts and SciencesPeace and Conflict Studies Program, University College, University of Toronto; 1993.
40. Forced Migration Online. What is forced migration? In: Forced Migration Online. 27 Jan 2012. Available: http://www.forcedmigration.org/about/whatisfm/what-is-forced-migration
41. Shariff S. Cyber-Bullying: Issues and Solutions for the School, the Classroom and the Home. 1 edition. London ; New York: Routledge; 2008.
42. Clark L. How one coder used 23andMe to create a race wall around the web. WIRED UK. 23 Jul 2015. Available: http://www.wired.co.uk/article/23andme-api-blocks-based-on-race-gender. Accessed 17 Oct 2016.
43. McPherson M, Smith-Lovin L, Cook JM. Birds of a Feather: Homophily in Social Networks. Annu Rev Sociol. 2001;27: 415–444. doi:10.1146/annurev.soc.27.1.415
44. Newell E, Jurgens D, Saleem HM, Vala H, Sassine J, Armstrong C, et al. User Migration in Online Social Networks: A Case Study on Reddit During a Period of Community Unrest. Proceedings of the Tenth International AAAI Conference on Web and Social Media. Palo Alto, CA: The AAAI Press; 2016. pp. 279–288. Available: http://www.aaai.org/ocs/index.php/ICWSM/ICWSM16/paper/view/13137
45. Diamond JM. Guns, Germs, and Steel: The Fates of Human Societies. 1st edition. New York: W. W. Norton & Company; 1999.
46. Cernea MM. Development-induced and conflict-induced IDPs: bridging the research divide. Forced Migr Rev. 2006; 25–27. Available: https://www.fmreview.org/brookings/cernea
47. Stanley J. Development-Induced Displacement and Resettlement. Forced Migration Online; N.D. p. 33. Available: https://www.alnap.org/help-library/development-induced-displacement-and-resettlement
48. Perlman J. Myth of Marginality: Urban Poverty and Politics in Rio de Janeiro. Berkeley, Calif.: University of California Press; 1980.
49. de Wet C. Development-induced Displacement: Problems, Policies and People. New York: Berghahn Books; 2005. Available: http://www.berghahnbooks.com/title/deWetDevelopment
50. boyd danah m., Ellison NB. Social Network Sites: Definition, History, and Scholarship. J Comput-Mediat Commun. 2007;13: 210–230. doi:10.1111/j.1083–6101.2007.00393.x
51. Kleinberg J, Lawrence S. The Structure of the Web. Science. 2001;294: 1849–1850. doi:10.1126/science.1067014
52. boyd danah. Friendster lost steam. Is MySpace just a fad? In: Apophenia. 21 Mar 2006. Available: http://www.danah.org/papers/FriendsterMySpaceEssay.html
53. Pulido L. Rethinking Environmental Racism: White Privilege and Urban Development in Southern California. Ann Assoc Am Geogr. 2000;90: 12–40. doi:10.1111/0004–5608.00182
54. Ley D. Gentrification and the Politics of the New Middle Class. Environ Plan Soc Space. 1994;12: 53–74. doi:10.1068/d120053
55. Rogers EM. Diffusion of Innovations, 5th Edition. 5th edition. New York: Free Press; 2003.
56. Castells M. The City and the Grassroots: A Cross-Cultural Theory of Urban Social Movements. New edition edition. Berkeley: University of California Press; 1984.
57. Stutzman F. Social Network Transitions. In: Fred Stutzman. 5 Nov 2007. Available: http://fstutzman.com/2007/11/05/social-network-transitions/
58. Shah C, Oh S, Oh JS. Research agenda for social Q&A. Libr Inf Sci Res. 2009;31: 205–209. doi:10.1016/j.lisr.2009.07.006
59. Gazan R. Microcollaborations in a social Q&A community. Inf Process Manag. 2010;46: 693–702. doi:10.1016/j.ipm.2009.10.007
60. Gazan R. Redesign As an Act of Violence: Disrupted Interaction Patterns and the Fragmenting of a Social Q&A Community. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. New York, NY, USA: ACM; 2011. pp. 2847–2856. doi:10.1145/1978942.1979365
61. De Jong GF, Gardner RW, editors. Migration Decision Making: Multidisciplinary Approaches to Microlevel Studies in Developed and Developing Countries. Pergamon; 1981.
62. Beta B. How to: Export, Import and Migrate Your Delicious Bookmarks. In: The Next Web. 17 Dec 2010. Available: http://thenextweb.com/lifehacks/2010/12/17/how-to-export-import-and-migrate-delicious-bookmarks/
63. Purdy K. The Best Alternatives to Delicious. In: Lifehacker. 17 Dec 2010. Available: https://lifehacker.com/5714329/the-best-services-for-migrating-your-delicious-bookmarks
64. Reichheld FF, W. Earl Sasser J. Zero Defections: Quality Comes to Services. In: Harvard Business Review. 1 Sep 1990. Available: https://hbr.org/1990/09/zero-defections-quality-comes-to-services
65. DeSouza G. Designing a Customer Retention Plan. J Bus Strategy. 1992;13: 24–28. doi:10.1108/eb039477
66. Wejnert B. Integrating Models of Diffusion of Innovations: A Conceptual Framework. Annu Rev Sociol. 2002;28: 297–326. doi:10.1146/annurev.soc.28.110601.141051
67. Peregrine P. Ancient Human Migrations: A Multidisciplinary Approach. 1st edition. Salt Lake City: University of Utah Press; 2009.
68. Tufekci Z. Who Acquires Friends Through Social Media and Why? “Rich Get Richer” versus “Seek and Ye Shall Find. 2010.