Module: 55-7626-00N-A-20123 – Social Media Use in Organisations (A-2012/3)Social Network Analysisof M2M OrganisationsLee James CoxB0049872MA in Technical Communication1
First some definitionsSocial Network Analysis (SNA) is the study of structure1. It is the mapping andmeasuring of relationships and flows (ties) between members (nodes) within anetwork.Machine to Machine (M2M) refers to the technology that connects:a. a device (such as a sensor or meter) to capture an event (such as temperature,inventory level, etc.)b. which is relayed through a network (wireless, wired)c. to an application (software program), that translates the captured event intomeaningful information (for example, items need to be restocked).M2M Organisations include device, network and application providers; as well as:• Enterprise Customers: provide the services to end-users, e.g. Coca Cola, British Gas• Platform Providers: equipment and solution providers to operators and others• System Integrators: build solutions to join up incompatible systems2Node NodeTie
Five things to first consider about SNA31. The first endeavour should always be to define what the Nodes and Ties mean. Forour M2M analysis, the nodes are organisations and ties are contractual relationships.2. Nodes can have different weightings of importance, e.g.– Size (revenue/subscribers)– Geography– Track record– Associations– Other segmentation attributes3. Ties have characteristics that matter when it comes to identifying things like leadership,influence and strength e.g.– Direction (one-way, both)– Quantity– Contract date– Value– Frequency4. Matrix, graphs and other visualization tools are important for analysis and measurement.– Tools like NodeXL will be required for any network of reasonable size– Considerable time is usually required to capture and keep the data up to date.– The M2M example in later slides shows just a few of the tens of thousands of Enterprises, 200+ Operators &300+ Application Providers that Jasper, Vodafone & Ericsson have as M2M partners!5. The perspective of the analysis can centre on the complete network (socio) or anindividuals personal network (ego).
Social-centric or Ego-centric SNA?Social-centric (complete network)a. Allows analysis of nodes and ties in comparison to widernetwork, e.g.– Are Operator relationships tightly bonded, diversified orconstricted?– Is there density/clustering of contracts within a geographysuch as Europe, or are there more cross-continent?b. Identifies behaviours affected by positions andconnections, e.g.– Does the number of application provider ties influence thenumber of ties a platform provider has?– Does the distance between application and platform nodesaffect the number of operator relationships?Ego-centric (personal network)a. Only ties directly with the focal organisation (Jasper)plus those Jasper is aware of are included.b. Perception is reality and opinions count. Subjectiveattributes are likely to have weight in many businessmatters4Figure 1 –Socio-centric view of the sample M2M Eco-SystemFigure 2 – Ego-centric view of Jasper M2M Network
Visualization can also reveal positional relationships• Operators have the highest degreecentrality. AT&T is most central of theoperators and is the longest establishedwith largest customer base.• Application Providers have the highestcloseness centrality. They work withmultiple operators but rarely direct withEnterprise customers.• Device/SIM and Platform Providers arestructurally equivalent nodes.• System Integrators are the most peripheral,having the smallest number of connections.• Enterprise Customers has the highestbetweeness centrality providing the onlypath to System Integrations• No groups of nodes are connected to eachother (cliques). However if the definition ofties were extended beyond ‘contractual’then informal or personal relationshipswould show all nodes in this extracted viewas being connected.OperatorsEnterprise CustomersDevice/SIMApplication ProvidersPlatformSystem Integrators5
Benefits & Limitations of SNABenefits LimitationsProvides framework to describe any complexnetwork.Collecting and maintaining source data can bedifficultIdentifies important individuals and the influencethey haveDoes not describe meanings, motives or explainwhy actions happen, e.g. why a contract was won.A typological analysis is more suitable.Can identify previously unrecognised sub-groupsthrough visual clustersLittle examination of important attributes such asattitudes, opinions and behaviours that may behelping or hindering relationshipsHighlights areas for further inquiry and possibleimprovement, e.g. gapsMore sensitive to data omissions than othersurveys. >75% sampling is required.Even weak ties may be revealed as important forbridging disparate groupsPrivacy can be ignored when views of otherscontribute to the analysis.Useful for track changes over time to reveal paths. Visualization can lead to over simplification andmisreading of results. E.g. Network measures suchas density can be easily misrepresented whennetworks of different sizes are compared.6
Other take away learning’s from SNA1. Social Network Analysis is not necessarily restricted to connections fromtools such as Facebook or LinkedIn. SNA can be applied to a wide rangeof network subjects, such as how diseases spread, mapping films andinteraction of characters, influence of language throughout the world,etc.2. SNA focusses on relationships rather than attributes of the ‘nodes’ fortheir own sake, or the ideation behind the relationship.3. A Social Network does not in itself encourage co-operation or collectiveaction. Community tools such as Online Forums lend themselves betterto co-ordination of action.4. Connections within business networks often reveal companies tied tocompetitors. This has implications for trust and potential leaks within asocial network.7
References Wellman B, Berkowitz SD (1977). Social structures: a network approach. Greenwich: JAI Press. Pinheiro, C. (2011). Social Network Analysis in Telecommunications. John Wiley & Sons. Based on Davies R (2011). Network Visualisation and Analysis, Cambridge. Borgatti, Carley & Krackhardt (2006).8