Mobbo: Re-imagining and Monetizing Online Social Networking


Research Paper (undergraduate), 2012

32 Pages, Grade: None


Excerpt


Table of Contents

Paper Presented At The 1st Science & Technology Conference & Exhibition, 14th August 2012.

Department of Research, Science & Technology - MIST

Abstract

Acknowledgements

Dedication

Introduction

Social Relations

From Personal Relationships to Communities of Interest

Communities
Membership
Influence
Activities

Introducing Mobbo
Mobbo Canvas and Unified Communication
Mobbo Carousel

Quantifying Social Relationships
Social Actions & Weights
Personal Network
Personal Network Composition
Category Weights
MobboRank Derivation
Action Value
Interaction Index
Affinity Index
MobboRank
Applying MobboRank to compute a member's Personal Network
Graphical Represantation of The MobboRank
Ranking Example

Generalized MobboRank
Mobbo Affinity Index
Affirnity Between Mobbos

Perfecting the Activity Feed
Organizing Feed Content
Computing the ActivityRank Parameter
Relevance of an Activity
The ActivityRank

Mobbos for Brands
Brand Following
Computing a Brand's Following

Conclusion

Bibliography

Abstract

The advent of online social networking has created opportunities for brands to attract large numbers of followers, while individuals can 'add' thousands of other members as personal friends. However, how users and brands share with, manage, and control communication to and from these large 'personal' networks has become a challenge. The challenge is exacerbated further by having separate massaging interfaces for communication channels like chat, inbox messaging, status update, etc, as it currently obtains in leading social networking services.

In this paper, we present Mobbo1, a service that ranks and organizes people who share similar interests (eg fans of a brand, or a user's network of friends) into concise and manageable communities called mobbos, to enable targeted messaging and intuitive control of communication to and from other network members. Every activity in the network is thus associated with a community of interest, such that all messaging in the network takes place within these mobbos, and is targeted to only members of the mobbo. The mobbo approach unifies messaging interfaces into one, to improve interaction experience and centralize control and access.

An algorithm to compute and rank a brand's or user's 'personal' network is also derived, while a method to organize the activity feed is presented.

Keywords: Social Networking, Online Community, Affinity, Interaction, Interests, Followers 1 Mobbo is available at http://www.mobbo.me

Acknowledgements

A. I am grateful for the contribution of my co-founders Tumelo Serumola and Goemeone Seretse. Their feedback, suggestions, and encouragements as I worked on this paper made a lot of difference.

B. Many thanks to Chantelle Olefile Apadile for her part in conceptualizing Mobbo.

C. I also thank Professor Edward Lungu for cross-checking the Math.

D. A final bigup goes to my family and early users of the kMobic Symbian App which gave birth to Mobbo.

Dedication

This work is dedicated to beautiful Tony and Masa.

Introduction

Social networking is a natural phenomenon that forms fibre and fabric of any society. It describes the act of a person participating in the activities of a social network. (Wellman, 1996) defines a social network as a set of people (or organizations or other social entities) connected by a set of socially-meaningful relationships. Social networking forms an important part in the lives of individuals and their communities (Setuke, 2007). Any individual has a social network, or is a member of at least one community as a result of the different relationships that she has formed. (Wellman, 1988a) argues that, despite the traumatic changes of modernization, communities continue to flourish. He concludes that communities are necessary refuges from outside pressures, sources of interpersonal aid in dealing with large bureaucracies, and a useful means of keeping streets safe.

Social network theory uses relationships among people (as defined by "weak” and "strong” ties and relations) to determine a person's own social network (Dueber & Misanchuk, 2001). Hence, at the core of a social network are relationships or social connections between social entities, which define how members or member groups are associated and interact with one another. Members of a social network and their relationships are organized in a patterned manner (Wellman, 1996).

Social Relations

(O'Murchu et al, 2004) and (Wellman, 1996) define a social relationship as an association between two or more people based on shared interests. Such relationships usually involve some level of interdependence. People in a relationship tend to share thoughts and feelings, and engage in activities together.

Abbildung in dieser Leseprobe nicht enthalten

Figure A. A relationship between two persons A and B. The shared interest here is attending the same class.

In social network theory, social relationships are generalized by friendship degrees, as friends, friend of friend, friend of friend of friend (in essence, strangers), (O'Murchu et al, 2004). Social relationships are therefore friendships, whereas a friend is defined as someone whose company and attitudes one finds sympathetic, and to whom one is closely related (O'Murchu et al, 2004).

People's relationships with others strongly affect their social resource, mobility, happiness, work habits and many other important things about them (Wellman, 1996). Social relationships can be voluntary such as mutual friendship whereby one person chooses to associate with another (Schlenker, 1980). In such case, the relationship can be terminated. Relationships can also be involuntary such as in blood relations. In which case, the relationship cannot be terminated even if relationship intimacy changes.

From Personal Relationships to Communities of Interest

As noted earlier, in social network theory, social relationships are generalized by friendship degrees, as friends, friend of friend, friend of friend of friend (in essence, strangers). Therefore, the focus is on relationships between members, i.e. who is whose friend or friend of friend?

Consider the diagram below:

Abbildung in dieser Leseprobe nicht enthalten

Figure B. A network of friends

Social network theory describes Person Aas a "friend of’ Person B, and Person Cas a "friend of friend of’ Person A.

However, the phrases "friend of’ and "friend of friend of’ are general descriptions of social relationships. In Figure B above, the friendship between persons A and B could be arising from the fact that they go to the same church. It could be that they spend time praying together or going to a local church branch. In that case, a more accurate description of their relationship would be "church mates”. Or maybe their friendship could be arising from the fact that they regularly hangout together at a local cafe, drinking and chatting. In which case, a more accurate description of their friendship would be "drinking buddies” or similar description. As such, we could represent graph A as shown below:

Abbildung in dieser Leseprobe nicht enthalten

Figure C. A network of friends, with friendships explicitly defined.

Representing a social network graph with explicit relationships as shown in figure C above, allows us to make more accurate deductions from the graph, about the social network. Such representation could aid analysis better than generalizations of"friend of’ and "friend of friend of’, which are popular in social network theory scholarly work (Finin et al, 2005), (Von Erb et al, 2008), (O'Murchu et al, 2004).

Taking a closer look at the graph in Figure C above; it can be noted that Person A and Person B are church mates. Persons C and D are also church mates. It could be that all Persons A, B, C and D go to the same church. Or it may be that Persons A and B go to one church while Persons C and D go to a different church. Either way, an accurate conclusion can be reached to the effect that Persons A, B, C and D belong to the same community of people who are interested in church. In Figure C above, members with similar friendships (and hence interests) can also be grouped together such that, instead of having members connected by one-on-one friendships, we have them grouped according to interests. This act is consistent with the definition of a relationship which describes it as an association of people with a shared interest (O'Murchu et al, 2004). When representing graph in Figure C, we can link or associate Persons A, B, C and D with the community of people interested in church. Other persons in the graph, with similar friendships can also be linked to a community which is common to their friendships. Hence, graph in Figure C can be represented as follows:

Abbildung in dieser Leseprobe nicht enthalten

Figure D. A social network represented in terms of connections between various communities that form it.

Figures B and D represent the same social network. While graph in Figure B represents the social network in terms of relationships or connections between members, graph in Figure D represents the social network in terms of relationships between the various communities of interest that form it. Connections between communities are defined by their respective membership compositions. Each arrow between two communities represents one member who is in both communities. If there are more arrows between any two communities, it implies that the communities are closely connected. We can therefore conclude that a social network represents connected communities of interest.

And thus, the focus of this paper is to represent the social network in terms of various communities that form it, and to map relationships between these communities as determined by their membership composition. Such representation of the network in terms of communities enables targeted and customized services to be provided.

Communities

(McMillan & Chavis, 1986) define community as a feeling that members have of belonging, a feeling that members matter to one another and to the group, and a shared faith that members' needs will be met through their commitment to be together. (Bellah et al, 1985) define a community as a group of people who are socially interdependent, who participate together in discussion and decision making, and who share certain practices that both define the community and are nurtured by it.

From these definitions, it is evident that communities are made up of people who are connected first and foremost by the common interest or purpose of the community, and secondly by their interaction with one another. Members of a community have a sense of belonging together, and this sense of community is a result of interaction and deliberation by people brought together by similar interests and common goals (Westheimer & Kahne, 1993).

McMillan and Chavis in their seminal 1986 article, describes community through the concept of "Psychological Sense of Community”. Among others, they mention membership and member's influence as being core components and characteristics of a community.

Membership

In any community, membership is clearly defined. (Dueber & Misanchuk, 2001) notes that, the ability to identify another member of a community allows people to better determine how to spend resources and with whom to feel comfortable. Boundaries define community membership, and can be created and enforced in more subtle ways such as group's use of language, styles of dress, myths and ritual (Dueber & Misanchuk, 2001). The concept of boundaries is also important in protecting group intimacy.

Influence

(Dueber & Misanchuk, 2001) further posits that in order to be attached to the group, an individual must have the potential of influencing the group. The ability of the group to influence its members is also crucial to maintaining cohesiveness. Hence influence with respect to community is bi-directional. The need for people to influence others is manifest in a community setup. (McMillan & Chavis, 1986) observes that, people possess an inherent need to know that the things they see, feel and understand are experienced in the same way by others. As such, people will go to great lengths to re-assure themselves that their world-view makes sense to others, hence develop the desire to influence others.

Activities

Community members engage in activities that benefit the cause of the community and are fulfilling to their individual needs. (Dueber & Misanchuk, 2001) notes that the individuals association with the group must be rewarding for the members, and that successes associated with group activities bring members closer together. Members always find the need to keep track of group activities and to interact with one another. Interaction between members can be one-on-one, or members can interact as a group.

When people interact and pursue shared interests through computer mediated communication, the community is no longer defined by the physical proximity of the members. (Wellman, 1999) argues that, when community is viewed as what people do together, rather than where or through what means they do them, community becomes separated from geography, physical neighborhoods, and campuses. As in offline communities, members of online communities engage in activities that benefit the cause of the community and are fulfilling to their individual needs. They interact and perform social actions that satisfy their social networking needs.

Through "activity feeds”, they keep track of what is happening within their networks. Forms of relations between community members has remained the same, what has changed is the way these relationships are maintained (Setuke, 2007).

Introducing Mobbo

Mobbo is a service that attempts to connect like-minded people, enabling them to share and achieve together. Emphasis is placed on simplifying messaging within the network, such that interfaces for different communication channels like chat, inbox messaging, status update, email, etc are unified into one. A brand's or user's personal network of followers or friends is ranked and organized into concise and manageable groupings called mobbos, while brands are also organized and classified into categories. In this way, users and brands are able to intuitively, and with minimal effort, share with, manage and control communication to and from their 'personal' networks, and other network members.

At the heart of the service is the mapping of relationships between these mobbos, as determined by their membership composition. All interaction in the network takes place within mobbos, enabling context to be inferred from every activity in the network. Just as an algebraic set may contain sub sets, a mobbo may contain one or more sub-mobbos. For instance, a user's personal network of friends is a mobbo: all the people in such network share a common interest of 'being friends with the user in question. These friends can be further divided into sub-mobbos by affinity to the user. Some may be very close to the user, while others may be distant acquaintances.

It is such grouping of users into interest-based sets that provides opportunity to simplify messaging, enable targeted communication, infer context from interactions in the network, and enable users to share and achieve in groups. Recommendation and prediction of user behavior can also be easily done when analyzing membership trends between these groupings.

Mobbo Canvas and Unified Communication

All interaction takes place within a single interface called the Mobbo Canvas. A member may share a text message or an image, collectively referred to as ‘mobbs', and it will be displayed on the Mobbo Canvas regardless of whether the mobb is an inbox item, chat item or status update. The same input area used to send an inbox message is also used to share a status update or for live chat messaging. And the same area used to display inbox messages is also used to display communication from other channels. The format of presenting messages is also the same. In this way, different communication interfaces are unified through the Mobbo Canvas. There is no need to open a small chat box commonly placed on the bottom-right corner to start chatting with an online friend. And there is no need to navigate to the inbox interface to interact with one-on-one inbox messages. The inbox is represented by a mobbo of two, and inbox messages also appear on the 'activity feed', unlike the popular approach of treating inbox messages differently.

Mobbo Carousel

For scalability purposes, Mobbo has a key feature that allows users to navigate between many mobbos in the shortest time possible, getting a preview of how much activity has or is taking place in any mobbo of interest. The carousel allows users to preview activity in many mobbos continuously, by scrolling backward or forward. Mobbos in the carousel are ranked according to the user's affinity to each.

Quantifying Social Relationships

Social Actions & Weights

Social actions are movements with a meaning and purpose, intended to address another directly or indirectly, and which may solicit a response from another person. The degree of relevance of a social action to a relationship differs from action to another. Some social actions have more impact in a social relation than others. For instance, buying one's friend a vehicle on their birthday does not have the same impact on the two's relationship as smiling at the friend on their birthday. Though both actions may have a positive impact on the relationship, one will certainly have more impact than the other. As a result, there is need to assign weights to social actions when quantifying social relationships.

We note that social interactions can be direct or indirect. When an interaction is targeted or addressed directly at another person, then it is direct. For instance, when person A comments on a statement made by person B, then the interaction between persons A and B is direct. But when two people interact with the same person in relation to a common topic, then their interaction is indirect. For instance when persons A and B both respond to person C regarding a statement she made, then the interaction between persons A and B is indirect.

In the model we use to represent a social network, we weigh social actions according to the amount of effort made to complete the action, i.e.

-How many clicks it took to complete the action?
-Was there text input involved?

Other criteria such as the time taken to complete the action are left for future work.

Accordingly,

-Viewing another member's profile will have weight 1 unit, because there is only one mouse click involved.
-replying another member's post will have weight of 3 units, because there is 1 mouse click to get focus of textbox, 1 unit for text input, and 1 unit for submitting the reply.

If the social action is indirect, we multiply its weight by half.

In the same way, we represent the weights of other relevant actions as:

Abbildung in dieser Leseprobe nicht enthalten

Table 1. Some actions and their assigned weights

Personal Network

Member interactions among others help us to determine who is in one's personal network, i.e. who is a mutual friend, who is in one's inner circle, who is following who or who is a stranger to whom?

The member's personal network describes the people that have had some direct or indirect contact with the member at some point. When a member views another member's profile, or comments on their status, she becomes part of that member's extended network. This is not different from real life scenario such as; when a person sees another, say at a conference. Even if they might not interact, the fact that one has seen the other creates a connection between the two. Next time when they meet at a different place, the other person can say, 'Hi, I saw you at XYZ Conference; your presentation was good!' The ensuing conversation will be different than if the first person had never seen the other. Hence, one-way actions such as seeing a person are significant in social networks.

Personal Network Composition

Any member interacts differently with people in her personal network. She is closely related to some, such as close friends or close relatives. She might also be closely following some, while others may be following her. Other members in her network might be general mutual friends, and the rest strangers (who are in essence outside her network). We classify the personal network into six broad categories determined by affinity to the subject user. This classification enables us to better represent any individual in relation to another in the network.

The categories are, Inner Circle, Mutual Friends, Followers, Those the Member is Following, Extended Network, and Strangers.

Inner Circle

This category defines those members who are very closely connected to the member in question than any other group of members. These people may be close friends, close family members or any persons that are strongly connected to the subject person.

Followers

This category defines those members who have great interest in the subject member. They might be interested in knowing about her latest activities, thoughts, or about general aspects of her life.

Following

This category defines those people that the subject member has great interest in. The member might be interested in knowing about their latest activities, thoughts, or about general aspects of their life.

Mutual friends

This category defines those people that the member is generally connected to. They are not in the member's inner circle but the member has directly interacted with them in the past and maintains this interaction at some level.

Extended Network

This category defines those people that the member has had indirect interaction with. For instance, if the member and person B comment on person C's status update, and the member and person B have never interacted directly, then they are in each other's Extended Network.

Strangers

This category defines those people that the member has never had any direct or indirect interaction with. This people in essence are not in the member's personal network. However, we define them here because they are part of the social network, and we have to represent the member with respect to them as well.

Category Weights

We argue that, of all the six categories described above, a member is more likely to want to know what is happening in her inner circle before wanting to know what is happening in the rest. The events happening in one's inner circle directly affect the member, and the member holds them important. For instance, the death of a close friend is more important than the death of one's favorite celebrity icon. Hence, when assigning weights to the categories, the Inner Circle carries the most weight.

Of the remaining five categories, a member is more likely to want to know the latest from people she is passionately following, as opposed to say, mutual friends, stranger or followers. An argument can be advanced that there is more passion associated with following someone than there is when one is just mutual friends with them. Hence, the category of'Those the Member is Following', carries the second most weight after 'Inner Circle'.

With regards to the remaining four categories (Followers, Mutual Friends, Extended Network & Strangers), a member is more likely to want to know the latest from mutual friends because she interacts with them more than she does with the other three groups. It then follows that the members would be more likely to want to know the latest from followers as opposed to extended network and strangers. This is so because there has at least been some direct interaction between the member and her followers (by definition of follower), unlike with extended network or strangers.

We can therefore, assign weights to the categories and re-arrange them as:

Abbildung in dieser Leseprobe nicht enthalten

Mobbo Rank Derivation

In the following section, we derive an important parameter called MobboRank. This parameter will enable us to perform tasks such as computing a user's personal network, ranking and determining the relevance of various mobbos to a user, making recommendations to a user, and predicting user behavior in the network. We use interactions among members to define relations between them, focusing on the level and depth of interactions between a particular member of interest and other members of the social network.

Action Value

We define a variable called Action Value, which describes the significance of an action made by a member. The Action Value, represented as Vi, addresses the question; what is the worth of an action t, made by a member toward another? The Action Value is a function of time elapsed since the action was made and the assigned action weight. We first note that recent social actions are generally more significant than older ones, i.e. the older the action, the less significant it is. If t represents the action age, we have;

(1) t= Today's Date - Action Date

Action age is inversely proportional to Action Value: i.e. if action age becomes bigger, Action Value decreases.

Hence, Action Value represented by Vi , for action i of assigned weight Ei is given by;

Abbildung in dieser Leseprobe nicht enthalten

Therefore,

[Abbildung in dieser Leseprobe nicht enthalten] where and Vi is the Action Value for action i and Ei is the assigned weight of action i.

Interaction Index

We define another useful parameter called, the interaction index, which we will use to sum up Action Values of all actions made by one member toward another. We can then use this

parameter to determine a member's affinity toward another. In this way, we are able to compare the affinities of two members and determine the associated MobboRank value.

By definition, the Interaction Index of member X with respect to member Y, represented as Ixy, is the sum of Action Values arising from actions made by member X toward member Y.

Hence, we have;

[Abbildung in dieser Leseprobe nicht enthalten] where b is the total number of actions and Vi is the Action Value for action i made by member X toward member Y.

Affinity Index

We define a final parameter, the Affinity Index of a member X toward member Y as the extent of attraction of member X toward Y. It is a function of Interaction Indices of both members X and Y toward one another.

The affinity of member X toward Y is determined by the number of interactions made by X toward Y, as taken against the overall interactions between the two. Hence, we are looking at X's contribution to the relationship.

Therefore, X's contribution to the relationship, represented as Cx, will be;

Abbildung in dieser Leseprobe nicht enthalten

Hence,

[Abbildung in dieser Leseprobe nicht enthalten] where Ixy is the Interaction Index of X toward Y, and Iyx is the Interaction Index of Y toward X.

Abbildung in dieser Leseprobe nicht enthalten

However, from member X's perspective, interactions she makes toward member Y are significant if they compare well with the interactions she makes toward other members in the network. For instance, say X on average makes 2000 interactions to other network members in a month, but to member Y she has made only three interactions this month. The three interactions made by X to Y as compared to X's overall involvement in the social network are less significant in her world.

Hence, the Affinity Index of member X toward Y will be a product of the interactions she made toward Y (her contribution to the relationship) and their significance. This significance is in fact a control variable.

If Sxy is the significance ofX's interaction toward Y as taken against her average interactions toward other members, we have;

Sxy = Interactions of X toward Y / Average of Total Interactions of X in the network Hence,

Ixy

[Abbildung in dieser Leseprobe nicht enthalten]where Ixy is the Interaction Index of X toward Y and Ax is the

Average Interaction Index of X toward other members.

Let Pxy represent the Affinity Index of X toward Y; from equations (5) and (6) we have,

Pxy = Interactions of X toward Y * Significance of the Interactions

Abbildung in dieser Leseprobe nicht enthalten

Hence,

Abbildung in dieser Leseprobe nicht enthalten

Similarly, Affinity Index of Y toward X will be;

Abbildung in dieser Leseprobe nicht enthalten

Note: Pyx has an upper bound of and [Abbildung in dieser Leseprobe nicht enthalten] has an upper bound of[Abbildung in dieser Leseprobe nicht enthalten]

MobboRank

The MobboRank represents the difference between the Affinity Indices of two members being evaluated, toward one another. In essence, it addresses the question; how is the affinity of a member X toward a member Y different from the affinity of Y toward X? It is this value that can tell us if a member is following the other, is a mutual friend, or is in another's inner circle. If there is little difference between their affinities to each other, and their Interaction Indices toward each other are high, then the members are strongly connected. Whereas, a big gap between their affinities to one another suggests that one is a follower of the other.

Therefore, from equations (7) and (8) we have,

The MobboRank value between X and Y, as determined from X's perspective, represented by Mxy, is therefore given by;

Abbildung in dieser Leseprobe nicht enthalten

Hence,

[Abbildung in dieser Leseprobe nicht enthalten], Ixy and Iyx cannot be both zero, and both Ax and Ay cannot be zero.

Note that the value of Mxy has an upper bound Max [Abbildung in dieser Leseprobe nicht enthalten] where Ixi is the maximum Interaction Index of member X in the network. To determine Iyi and Ay, we consider the member with the highest Interaction Index toward subject member X.

Hence, the MobboRank values for member X and all her friends (or members who have had direct interaction with her) can be interpreted as falling within a circle of radius Max [Abbildung in dieser Leseprobe nicht enthalten]

This situation can be represented by the following diagram:

Abbildung in dieser Leseprobe nicht enthalten

Figure E. MobboRank values for a member with respect to other members in her personal network are bounded. They are depicted by the bold blue line, with end points marked.

Applying MobboRank to compute a member's Personal Network

To determine membership composition of categories for the subject member X's personal network using the MobboRank, we first note that; when the absolute value of the MobboRank is greater than 3, the relationship is completely one-sided. Hence if MobboRank value is greater than 3, the subject member X is following the other member, and if the value is less than -3, the other member is following the subject member X. If absolute value of MobboRank is less than or equal to 3, then that situation represents Mutual Friends. And, within Mutual Friends is the Inner Circle.

We assume that a member's Inner Circle is about 3% of her Mutual Friends. Hence, if Mutual Friends are represnted by an absolute value of the MobboRank which is less than or equal to 3, then the member's Innner Circle will be represented by an absolute value of the MobboRank which is less than or equal to 0.09. And finally when the absolute value of the MobboRank is greater than 0.09 and less than or equal to 3, member X and the other member are Mutual Friends.

Hence, we can summirize as follows:

Abbildung in dieser Leseprobe nicht enthalten

Table 3. Ranking a user's personal network

There critical point ([Abbildung in dieser Leseprobe nicht enthalten]is the center of the hyperbola.

To determine the remaining critical points, we express equation (11) in the standard form of a unit hyperbola by completing the square.

Equation (11) becomes:

Abbildung in dieser Leseprobe nicht enthalten

Equation (14) above is of the form: ^ = 1 where,

Abbildung in dieser Leseprobe nicht enthalten

(15)[Abbildung in dieser Leseprobe nicht enthalten]

(16)[Abbildung in dieser Leseprobe nicht enthalten]

But [Abbildung in dieser Leseprobe nicht enthalten]where C will determine the foci of the graph.

(17)[Abbildung in dieser Leseprobe nicht enthalten]

We can therefore make the following summary:

Center of the Hyperbola is at: [Abbildung in dieser Leseprobe nicht enthalten]

Foci are:[Abbildung in dieser Leseprobe nicht enthalten]

Vertices are:[Abbildung in dieser Leseprobe nicht enthalten]

Asymptotes are given by: [Abbildung in dieser Leseprobe nicht enthalten]

Hence, for b > a and m < 0 we have the following graph:

Abbildung in dieser Leseprobe nicht enthalten

We only consider points of the graph which are within the shaded area, since both x and

y are always positive. As b increases, the slope of the asymptote[Abbildung in dieser Leseprobe nicht enthalten] increases, hence the asymptote (and hence the graph) rotates counter-clockwise about the center until it is almost parallel to the y-axis. When b = a, the asymptotes have slope 1 and -1 respectively; hence they are respectively parallel to y = x and y = —x. To obtain the graph for the case b < a , we note that the slope of the asymptote [Abbildung in dieser Leseprobe nicht enthalten]decreases as a becomes bigger and bigger than b. Hence we rotate the graph clockwise about the center until the asymptote [Abbildung in dieser Leseprobe nicht enthalten] is almost parallel to the x-axis.

Similarly, when b > a and m> 0 , the corresponding graph is represented by;

Abbildung in dieser Leseprobe nicht enthalten

Figure G. MobboRank graph for b > a and m> 0

We can therefore rank and represent a member's personal network as follows:

Abbildung in dieser Leseprobe nicht enthalten

Figure H. A member's personal network ranked using the MobboRank.

Ranking Example

Consider the table below which depicts 8 relationships (RI... R8) that member X has with other network members. The Interaction Index of member X toward each member is represented by Ixy and the Interaction Index of each member toward X is represented by Iyx . Ax is the average Interaction Index in the network, while Ay represents the average Interaction Index in the network, for each member who has a relationship with X.

Abbildung in dieser Leseprobe nicht enthalten

Table 4. Sample values for ranking a member's personal network

Generalized MobboRank

The MobboRank formula (Equation (9)) can be generalized and applied to compare affinities of any two entities in a network of many, which are not necessarily persons in a social network. In such case, it is necessary to treat the entities as sets, and define a relation from one entity to the other, before substituting into the MobboRank formula. In this way, we're able to use the MobboRank to find out which entity is 'following' the other.

Generalizing the MobboRank can be summarized with the following steps:

1.1 Define the domain/universal set within which the entities exist. For instance, one domain could be 'interactions between people in a social network'; another could be 'movement of people between mobbos'.

1.2 Regard each entity as a set, and define its membership composition according to the defined domain. For instance, consider entities X and Y. For the domain of 'interactions between people in a social network'; set X will consist of interactions associated with person X and set Y will consist of interactions associated with person Y. While for the domain of 'movement of people between mobbos’; set X will consist of movements associated (inbound and outbound) with mobbo X, and set Y movements associated with mobbo Y.

1.3 Determine the intersection of the entity sets. For instance, for the domain 'interactions between people in a social network'; the intersection between the entity sets X and Y will be all interactions that took place between the two persons X and Y only. While for the domain 'movement of people between mobbos’; the intersection will be all the movements of people between the two mobbos X and Y only.

1.4 Divide the intersection of entity sets into two subsets; such that one subset consists of elements which represent a relation from entity X to entity Y, while the other subset consists of elements which represent a relation from Y to X, according to the defined domain. For instance, for the domain 'interactions between people in a social network'; the first subset of the intersection will be defined by the relation from person X to Y, which is the set of interactions from person X toward person Y. And the other subset of the intersection will be defined by the relation from Y to X, which is the set of interactions from person Y toward person X. While for the domain 'movement of people between mobbos'; the first subset of the intersection will be defined by the relation from mobbo X to Y, which is the set of people who were in mobbo X and later joined mobbo Y. And the other subset of the intersection will be defined by the relation from mobbo Y to X which is the set of people who were in mobbo Y and later joined mobbo X.

1.5 Substitute into MobboRank equation. The number of elements in each subset of the intersection as defined above represents the Interaction Indices. For the averages, we consider the intersections of the entity sets with other entities in the network.

Mobbo Affinity Index

We can determine the Affinity Index of a member toward a mobbo (called Mobbo Affinity Index) by substituting into equation (7).

Abbildung in dieser Leseprobe nicht enthalten

Hence,

[Abbildung in dieser Leseprobe nicht enthalten] where Ixa is Interaction Index of user X toward mobbo A, and Ax is the average number of interactions member X makes to other mobbos.

Affirnity Between Mobbos

Using the MobboRank formula, we can determine how close two mobbos X and Y are, depending on their membership composition. This can help us to make suggestions and recommendations to network members depending on their memberships of different mobbos. To determine the affinity between two mobbos, we substitute for the Interaction Indices in equation (9), using the number of members who were in one mobbo and later joined the other mobbo. For two mobbos, X and Y, if more members of X joined Y than members of Y who joined X, then mobbo X is likely a 'follower' of mobbo Y. And therefore, we can recommend mobbo Y to new members of mobbo X.

When applying equation (9) to two mobbos, we define Ixy as the number of members who are in mobbo X and later joined mobbo Y, and Iyx as the number of members in mobbo Y who later joined mobbo X. Ax is the average number of members of X who later joined other mobbos, and Ay is the average number of Y who joined other mobbos.

Example

Consider the example below for two mobbos X and Y.

Abbildung in dieser Leseprobe nicht enthalten

Figure I. Affinity between two mobbos. The arrows represent members who were in one mobbo and later joined the other mobbo.

5 members of mobbo X went on to join mobbo Y and 2 members of mobbo Y went on to join mobbo X. Assume on average, 5 members of mobbo Y join other mobbos in the network, while 1 member of mobbo X on average joins other mobbos. Then from equation (9) we have;

Abbildung in dieser Leseprobe nicht enthalten

The absolute value of Mxy is greater than our threshold of 3; hence mobbo X is 'following' mobbo Y. As such, when a member joins mobbo X, we can recommend mobbo Y to her. Or we can predict that she is likely to join mobbo Y. Comparing affinity between two mobbos is the same as comparing affinity between any two interests.

Perfecting the Activity Feed

The composition of the activity feed as presented to a member should be relevant to her. As such, the feed needs to be ordered by relevance; such that the most relevant activities appear at the top, while the least relevant appear at the bottom of the feed.

Organizing Feed Content

Besides ranking activities in the feed, it is necessary to present feed content in a way that is easy to be consumed. Leading social networking services mix multimedia and text into a single stream.

However, on a particular browsing session, a user may be interested in just browsing photos. While on another, she may be interested in engaging in discussion, and replying to topics. It is for this reason that we present the activity feed in panes. One pane is for media, and another for text. There is possibility to switch panes by expanding one and minimizing the other.

Computing the ActivityRank Parameter

Relevance of an Activity

The overall relevance of an activity to a member is determined by five factors, namely;

1. Action Weight

2. Mobbo Affinity Index

3. Activity Author

4. Activity Age

5. Activity Feedback

Except for the Activity Age, all the above factors are directly proportional to the relevance of an activity.

Action Weight

The Action Weight, represented by E, is the assigned weight to the activity according to Table 1.

Mobbo Affinity Index

Mobbo Affinity Index, represented byZ , is the affinity of the subject member to the mobbo within which the activity is authored. This parameter factors the context of an activity into the ranking. It is given by equation (18).

Activity Author Weight

Activity Author Weight, represented by n, is the assigned weight of the author according to Table 2.

Activity Age

Activity Age, represented by t, is the difference between today and the date the activity was authored.

Activity Feedback

Activity Feedback represented by G, is the amount of feedback the activity has received. This feedback is in the form of actions, and is a function of the assigned weights of those actions, their authors and their ages. The Activity Feedback is given by;

[Abbildung in dieser Leseprobe nicht enthalten] where n is the weight of author of activity i, E the weight of activity í, t the age of the activity, and b is the total number of actions that are given as feedback to some activity.

The ActivityRank

If Ri represents the ActivityRank parameter, we can therefore sum up the above as follows;

Abbildung in dieser Leseprobe nicht enthalten

Mobbos for Brands

Just like users, brands can have mobbos. These mobbos organize followers of a brand into one community. Brands are also able to offer services and interact with followers directly from the mobbos. For instance, a church that usually collects Sunday Offerings can do so directly on its mobbo. A musician can also do ticket sales for her upcoming concert on her mobbo. In this way, services associated with a brand are accessed instantly through the brand's mobbo, offering an improved approach for brands to serve and share with their followers. Enabling brands to offer various services directly on their mobbos allows Mobbo to retain a small percentage of any monetary transaction in the respective mobbo; thereby utilizing a proven monetization model in eCommerce. Mobbos for brands are organized into categories such as Sport, Religion, Politics, Workplace, etc. Such categorization enables users to interact with mobbos by subject of interest.

Brand Following

Followers of a brand are ranked and organized into groups by affinity to the brand, to make possible targeted interaction between followers and brands. The four standard categories are Active Followers, New Followers, Average Followers, and Dormant Followers.

Active Followers

As the name suggests, these are followers who interact with the brand often, and actively participate in the brand's mobbo.

New Followers

These are followers who recently interacted with the brand's mobbo for the first time.

Average Followers

These interact with the brand's mobbo from time to time, but not as actively and often as Active Followers.

Dormant Followers

These followers have not interacted with the brand in a long time

Computing a Brand's Following

When ranking and organizing followers of a brand, we compute the Mobbo Affinity Index of each follower, as represented by equation (18). The higher the Mobbo Affinity Index, the more passionate about the brand the follower is. We note that, possible Mobbo Affinity Indices of Ixi brand followers are always greater than zero and less or equal to[Abbildung in dieser Leseprobe nicht enthalten] Where Ixi and Ax are taken from the follower with the highest Mobbo Affinity Index toward the brand.

When the Mobbo Affinity Index is greater than 1, the follower's Interaction Index toward the particular brand is above her average Interaction Index toward other brands. The follower is therefore an 'above-average' follower, and hence an Active Follower. When the Mobbo Affinity Index is less than 1, the situation represents Average Followers. But within Average Followers are Dormant Followers. We defined Dormant Followers as those who have not interacted with the brand in a long time. Hence a value of Mobbo Affinity Index less than 0.1 is suitable for this category. New Followers are not ranked by Mobbo Affinity Index but by recency of their first interaction with the brand.

We can summarize as follows:

Abbildung in dieser Leseprobe nicht enthalten

Table 5. Ranking a brand's followers

Ranking followers of a brand allows it to target services and messaging. For instance, a brand can send an introductory message to only New Followers, or offer a discounted service to them as a customer retention strategy. Useful analytics can also be deduced as followers interact with the brand is such groupings.

Conclusion

In this paper, we suggested how an online social network can be represented in terms of the various communities that form it. We have introduced Mobbo, a social networking service that is based on grouping interactions by communities of interest. Emphasizing Mobbo's focus on targeted communication, we derived an algorithm to rank and organize a member's network of friends and a brand's followers, to allow for targeted messaging. We have also suggested how the concept of mobbos can unify messaging interfaces for different communication channels such as chat, inbox and status update; noting that in a mobbo, there is no distinction between an inbox message and a status update, for instance. We finally presented an algorithm to rank activities in the 'activity feed'.

The Mobbo approach as outlined in this paper provides a new method of sharing and communication within an online social network. The possibility to retain a small percentage per monetary transaction on services offered within brand or user mobbos provides a monetization opportunity.

We conclude that Mobbo offers a viable monetization model, improves messaging and provides an improved and worthwhile sharing experience for users of online social networking services.

Bibliography

1. Wellman, B. 1996. An Electronic Group is Virtually a Social Network. University of Toronto, Canada .MS5 1A1.

2. Hofstede, G. 1984. Culture's Consequences: International Differences in Work-Related Values. Sage Publications.

3. Setuke M. 2007. MSN Platform: A Mobile Social Networking Platform Targeting Mobile Phone Users in The Developing World. MSc Thesis.

4. O’Murchu, I., Breslin, J. & Decker, S. 2004. Online Social and Business Networking Communities. DERI - Digital Enterprise Research Institute. National University of Ireland, Galway, University Road, Galway, Ireland.

5. Dueber & Misanchuk, 2001. Sense of Community in a Distance Education Course. Mid­South Instructional Technology Conference Murfreesboro, TN.

6. Pacuit, E. 2005. Topics in Social Software Information in Strategic Situations. PhD thesis, The City University of New York.

7. Kersten G, Kersten M & Rakowski, 2002. Software and Culture: Beyond the Internationalization of the Interface. 86 Journal of Global Information Management, 10(4), 86-101

8. Al-Ghorairi, M. 2005. The Rise of Social Software: What Makes Software “Social”? Dissertation submitted in partial fulfillment of the requirements for the degree of MA in Electronic Media School of Humanities, Oxford Brookes University.

9. Wellman, B., Carrington, P., & Hall, A. (1988). Networks as personal communities. In B. Wellman & S. D. Berkowitz (Eds.), Social Structures: A Network Approach (pp. 130-184). Cambridge: Cambridge University Press.

10. McMillan, D. W. & Chavis, D. M. (1986). Sense of Community: A Definition and Theory. Journal of Community Psychology, 14, 6-23.

11. Tim Finin, Li Ding, Lina Zhou and Anupam Joshi, 2005. Social networking on the semantic web. The Learning Organization, Vol. 12 No. 5, pp. 418-435 Emerald Group Publishing Limited, 0969-6474.

12. Bellah, R. N., Madson, N., Sullivan, W. M., Swidler, A., and Tipton, S. M. 1985. Habits of the heart: Individualism and commitment in American life. Berkeley: University of California Press.

13. Westheimer, J. & Kahne, J. , 1993. Building school communities: An experience-based model. Phi Delta Kappan, 75(4), p. 324-28.

14. Sasha A. Barab, 2003. An Introduction to the Special Issue: Designing for Virtual Communities in the Service of Learning. The Information Society, 19: 197-201, Taylor & Francis Inc. ISSN: 0197-2243 print / 1087-6537.

15. Schlenker, Barry R. 1980. Impression management: The self-concept, social identity, and interpersonal relations. Brooks/Cole Pub. Co. ISBN 081850398X.

16. Bert N. Adams, 1967. Interaction Theory and the Social Network. Sociometry, Vol. 30, No. 1 pp. 64-78. Published by: American Sociological Association.

17. Von Arb M. , Bader M. , Kuhn M. , Wattenhofer R. 2008. VENETA: Serverless Friend-of- Friend Detection in Mobile Social Networking. Networking and Communications, 2008. WIMOB '08. IEEE International Conference on Wireless and Mobile Computing,

Excerpt out of 32 pages

Details

Title
Mobbo: Re-imagining and Monetizing Online Social Networking
College
International School of New Media at the University of Lübeck
Grade
None
Author
Year
2012
Pages
32
Catalog Number
V201835
ISBN (eBook)
9783656283133
ISBN (Book)
9783656287834
File size
784 KB
Language
English
Keywords
mobbo, re-imagining, monitizing, online, social, networking
Quote paper
Mpho Setuke (Author), 2012, Mobbo: Re-imagining and Monetizing Online Social Networking, Munich, GRIN Verlag, https://www.grin.com/document/201835

Comments

  • No comments yet.
Look inside the ebook
Title: Mobbo: Re-imagining and Monetizing Online Social Networking



Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free