« Web 2.0 , user generated content and social networking course - Hong Kong | Main | Which telegraphs do we still hang on to? »
July 11, 2008
Cost per related audience
OVERVIEW
Peter Drucker, widely regarded as the father of modern management said that - What gets measured gets managed . Between the eras of Web 1.0 to Web 2.0 – the advertising model on the Web has now reached maturity. However, the advertising model on the social web(social networks) – is far from accepted. One of the reasons for this is the lack of relevant metrics since existing metrics like CPM do not apply to social networks.
The mathematical theories underlying social networks have existed for some time now. Also, online social networks have become mainstream. However, the applicability of the advertising model to social networks remains the missing link – the solution of which is commercially very significant especially if it is computable as a metric.
In a nutshell, we need a set of quantifiable advertising metrics (like CPM) - applicable to social networks. In previous blogs, I discussed the calculation of Personal CPM. In this blog, we discuss the idea behind Cost per related audience(CPRA). CPRA involves the concept of identifying a set of related profiles within a social network by navigating the social graph. Consequently, the advertiser would ‘buy’ advertising against a set of profiles through a metric called Cost per related audience. Further, reputation within the social network can be used to further refine the results. All the above factors are quantifiable within a social network - (including reputation) – and thus we can create a truly useful metric that can benefit advertisers
BACKGROUND
In previous blogs, I had been exploring the idea of how to measure personal CPM. (Calculating personal CPM and extending the idea of personal CPM beyond the Web to Telecoms) and the belief that we need new metrics if we are to get the support of advertisers for Web 2.0 applications and user generated content.
In this blog, we extend that discussion to other metrics beyond Personal CPM.
For the purposes of this discussion:
a) The unit of advertisement (i.e. against which the advertiser places an advertisement) is a social networking profile. Note that we can place advertisements against other objects in a social network (and hence lead to different metrics – maybe a topic for a follow-on blog ) but for now – let us consider the profile as a target unit of advertisement.
b) We take the principles of personalised advertising as a given. Traditional channels like TV are a mass market, interruptive medium for advertising. Advertising in such media is not personalised to the audience. Today, with the rise of the Web, most of us accept the idea that advertising should be personalised and targeted to the user.
c) The profile on social networks is one dimensional – it can be enhanced in many ways for instance by including social(relationship/network/social graph), behavioural(incorporating roles) and demographic profiling(traditional segmentation)
Extending the above ideas ..
a) If we extrapolate the idea of personal CPM, the next logical unit for advertisements is to consider is a ‘group of profiles’ – i.e. advertisements against a ‘related audience’ defined by a collection of related profiles.
b) If we are to target a relevant audience within a social network, then a ‘shotgun’ approach would be to place advertisements against a ‘group’ (such as a facebook group). However, that does not solve the problems mentioned in my previous blogs (for instance a group about ‘Nike’ may not be flattering to Nike – and hence not a place where Nike might want to advertise against)
A more interesting and sophisticated approach is to identify the relevant audience by navigating the social graph. Consequently, the advertiser would ‘buy’ advertising against a set of profiles through a metric called ‘Cost per related audience’
THE MISSING LINK
The term ‘social graph’ is currently in vogue – even more so after it was popularised by Brad Fitzpatrick(now at Google) . The social graph and its applications are now becoming significant because of the emergence of Web based social networks. However, the principles underlying social networks
and Graph theory have been with us since as early as 1736 following a paper written Leonhard Euler on the Seven Bridges of Königsberg in that year
However, while Graph theory exists in it’s mathematical form and the usage of social graphs have become popular due to the rise of social networking, the application of the advertising model to social networks via the social graph remains a missing link – the solution of which is commercially very significant especially if it is quantifiable through metrics
THE SOCIAL NETWORK
The application of advertising models to social networks can be an extensive and a complex subject since there are many possibilities.
Adapted from Wikipedia, we can describe social network analysis as follows
A social network is a social structure made of nodes (which are generally individuals or organizations) that are tied by one or more specific types of interdependency, such as values, visions, ideas, financial exchange, friendship, kinship, dislike, conflict or trade. The resulting structures are often very complex.
Social network analysis views social relationships in terms of nodes and ties. Nodes are the individual actors within the networks, and ties are the relationships between the actors.
In its simplest form, a social network is a map of all of the relevant ties between the nodes being studied. The network can also be used to determine the social capital of individual actors. These concepts are often displayed in a social network diagram, where nodes are the points and ties are the lines.
An overview of social networks at orgnet and a simple diagram below offers some initial ideas based on social network analysis
Remember that our goal is to find a set of related profiles which are suitable for advertising based on a topic.
We are concerned here with the location of actors in the network – measurement of which gives the centrality of a node. The computation of centrality in turn gives us the roles and groupings within the network. A connection exists between two nodes if they regularly talk/interact with each other. The above diagram illustrates three types of centrality : Degree Centrality, Betweenness Centrality, and Closeness Centrality.
Degree centrality represents the number of direct connections a node has – for instance Dianne has the highest degree in the above diagram (hub). In contrast, Heather has few direct connections - but exhibits a high level of betweenness since she lies between two important constituencies and hence has a higher power. Finally, while Fernando has fewer connections than Dianne, he has the shortest path to all others in the network – and hence Fernando has the highest visibility into the information flow within the network.
Thus, even the most rudimentary analysis exhibits relations between profiles which can be used to find related profiles to advertise against.
REPUTATION AS A CURRENCY
How else could we extend this?
The underlying ‘currency’ of a social network is the reputation of it’s members – and by navigating the social graph and complementing it with reputation metrics – we can create truly useful metrics that can benefit advertisers. Note that all the above factors are quantifiable within a social network - (including reputation) are quantifiable.
CONCLUSIONS
• The mathematical theories underlying social networks have existed for some time now.
• Online social networks have become mainstream.
• However, the applicability of the advertising model to social networks remains the missing link – the solution of which is commercially very significant especially if it is computable as a metric.
• We need a set of quantifiable advertising metrics (like CPM) - applicable to social networks.
• ‘Cost per related audience’(CPRA). CPRA involves the concept of identifying the relevant audience within a social network by navigating the social graph.
• The advertiser would ‘buy’ advertising against a set of profiles through a metric called ‘Cost per related audience’.
• Further, reputation within the social network can be used to further refine the results.
• All the above factors are quantifiable within a social network - (including reputation) – and thus we can create a truly useful metric that can benefit advertisers
Posted by ajit at July 11, 2008 10:56 AM
Trackback Pings
TrackBack URL for this entry:
http://www.opengardensblog.futuretext.com/mt-tb.cgi/937