Chapter 6
Understanding Social Capital
Platforms such as Klout, Kred and PeerIndex seek to provide a measure of social capital based on social media data and are a novel addition to effective influencer relations. However, these relatively new tools require human, real-world awareness and an understanding of the limitations of an algorithmic approach.
What is social capital?
Social capital is not new, nor is it something that relates specifically to social media; it has been the subject of academic study and debate for decades. There are two key perspectives on social capital; one is that it is attached to an individual and the other that it is attached to society as a whole. The UK Office for National Statistics, for example, considers social capital a measure of society's collective “health”.37 A noted collective definition is by Robert Putnam, Professor of Public Policy at Harvard University:
“Social capital refers to the collective value of all ‘social networks’ [who people know] and the inclinations that arise from these networks to do things for each other [‘norms of reciprocity’]38
An individual view, as described by Nan Lin, Family Professor of Sociology at Trinity College, Duke University, defines social capital as:
“…investment in social relations by individuals through which they gain access to embedded resources to enhance expected returns of instrumental or expressive actions.”39
What Lin calls “embedded resources” are the amount and variety of wealth, status and power of those with whom an individual has direct or indirect ties. “Instrumental actions” here mean actions taken in order to make economic, political or social gains, and “expressive actions” are actions taken to preserve an individual's resources.
In an influencer relations context, it is this individual perspective that is perhaps most useful.
Influence and social media capital
Influence is “the capacity or power of persons or things to be a compelling force on or produce effects on the actions, behavior, opinions, etc., of others”.40 The term is often overused or misused, but it is everywhere, and was even described by Michael Conroy of Temporo41 as the social media industry's “white whale”.42
The potential to exert influence is purported to be a key element of social capital. Lin wrote that it could also produce benefits such as access to information, greater “social credentials” or authority, and a feeling of wellbeing created by public recognition of an individual's claim to these resources.
Influence is an inherently difficult thing to observe, and even more difficult to measure. To do so, we need to isolate the element exerting influence and the resulting “action, behaviour, opinions etc.” caused by it.
Given its relationship to social networks, there has been a drive to measure social capital and its resulting influence within social media. There are many tools and platforms that claim to do this, and three of the best known at the time of writing are Klout, Kred and PeerIndex.
All three talk about measuring influence. Klout's strapline is “The Standard for Influence” but, as we have already noted, influence is not easily measured. What they attempt to measure is arguably closer to social capital, highlighted by Brian Solis in his report The Rise of Digital Influence,43 and as they mainly focus on signals and measurements from social media; “social media capital” might be a more accurate description.
Relevance is King
In his report, Solis highlights what he describes as “the Pillars of [Digital] Influence”, reach, resonance and relevance. Reach is a function of popularity and goodwill, but also physical proximity. Resonance is the frequency of a message or object, how long it appears and the engagement it generates. Relevance is the authority that a person has on a subject, the trust placed in them, and their underlying social relationships.
Solis suggests that it is the combination of reach and relevance that is measured by any of these scores. What this highlights is that relevance is a necessary condition: reach, or popularity, is worthless without relevance. Therefore, the relevance of those that can be reached is also important, and vital to understanding the value of scoring approaches to social media capital.
Social media capital scoring approaches
The scoring approaches of three popular platforms are summarized below.
Klout: Scoring (out of 100)
The Klout Score incorporates more than 400 signals from seven different networks. We process this data on a daily basis to generate updates to your Klout Score.44
These “signals” include comments, retweets, mentions, followers and subscribers. Some signals are only available when Klout has access to networks other than Twitter, e.g. LinkedIn and Facebook.
Klout: Relevance
Accounts are allocated topics about which Klout thinks they achieve influence. You can use these topics to see the accounts that Klout ranks most highly for each. At the time of writing, I am apparently most influential on recessions, Facebook and apps, about none of which I'm especially knowledgeable or vocal.
Kred: Scoring
Kred has two scores:
Influence is the ability to inspire action. It is scored on a 1,000 point scale. We measure Influence by assessing how frequently you are Retweeted, Replied, Mentioned and Followed on Twitter.
Outreach reflects generosity in engaging with others and helping them spread their message. We measure Outreach on Twitter by your Retweets, Replies and Mentions of others.45
Outreach is a cumulative score that has no maximum, although the highest score at the time of writing is 12.
Kred also takes into account similar data from Facebook if you have given it access. Kred is the most transparent of the three, letting you see an audit trail of where your Kred points have come from.
Kred: Relevance
Accounts have Global scores, but they also have Community scores based on Kred's assessment of their relevance to particular areas. I rank as most relevant on Advertising, which is at least a little more accurate than Klout.
PeerIndex: Scoring (out of 100)
PeerIndex is the most opaque in terms of its scoring methodology.
The PeerIndex algorithm recognizes the speed and quantity by which users spot, share (and thus endorse) content on any specific topic. Our content recommendation decisions can thus be used as a proxy to measure our knowledge and authority in a specific subject area. Your authority on a subject is affirmed when the content you share is approved, i.e. Retweeted, Facebook Shared, +1'ed or commented on, by someone else with authority on the subject.46
PeerIndex: Relevance
PeerIndex is in the process of adding support for topics. It currently gives an account a score for the topics it judges relevant. These topics are very broad, e.g. “News, Politics and Society”.
Comparing Klout, Kred and PeerIndex
An argument can be made that PeerIndex's opacity is an asset, as the more an individual knows about their score, the more they can take action to try to increase it. For instance, the knowledge that linking other social media platforms to your profile can boost your score may encourage people to do so. The more the score can be gamed, arguably the less valuable it is.
When we understand how each scoring system works, we then need to understand the scale that applies to the scores. Both Klout and PeerIndex use logarithmic scales, while Kred takes a normalized approach.47 A logarithmic scale means that as you ascend it becomes more and more difficult to increase your score, so there will be proportionally fewer accounts ranked between 20 and 29 than there were between 10 and 19, and so on. A rough estimate suggests Klout may have a logarithmic “base” of around 3–4 in relation to these bands. The more transparent Kred publishes its influence scoring distribution online, and their data shows that on 11 November 2011, accounts with scores of greater than 850 only made up approximately 0.02% of the total, while accounts with scores between 501 and 550 accounted for 12.5%.
This is important when considering outreach since, despite what the numbers might suggest, logarithmic scoring means someone with a Klout score of 90 is likely to have considerably more than 50% more social media capital than someone with a score of 60. It seems likely that relatively few users fully understand the nature of a logarithmic scale, or are even aware of the difference that scale makes to scoring, which is potentially problematic and could make the scores less intuitive than they appear to be.
Clout v. Klout
A fundamental issue with all of these approaches is that they base their scores predominantly on social media interactions and data, while much of a person's social capital is more likely to be based on their actions and relationships in the real world. Not considering real-world reach might mean that someone's influence score seems laughably high, or indeed unreasonably low. For example, a journalist writing for the Financial Times might personally have relatively little social media capital, but has the potential to exert greater influence in their role with an authoritative media outlet.
In response to criticism on ignoring the effects of real-world influence, the platforms have tried to include elements of real-world social capital by allowing people to award others extra Kred or Klout, though these are necessarily limited in their impact to prevent the scores being “gamed” too easily.
In an innovative approach to including real-world data, in August 2012 Klout began incorporating an individual's Wikipedia page, its page rank and the ratio of inbound to outbound links in their scores. The more popular their page, the higher the score. This resulted in Barack Obama's score rising, and Justin Bieber's falling. This inventive proxy for potential real-world social capital seems intuitively effective, if the President of the United States now ranks higher than a teen pop sensation.
Other wild-card accounts whose true social capital is probably marginal at best were also downgraded significantly. Astrology/horology mouthpiece @Xstrology, ranked at 95 in June 2012, had fallen to 81 by the end of the year and @uberfacts, which broadcasts interesting trivia, fell from 94 to 83 in the same period. By this measure, Klout and PeerIndex (which rank these particular accounts much lower, at 69 and 74 respectively) certainly appear to have a potential advantage over Kred, with Kred ranking both of these accounts at its maximum Global Ranking of 1,000.
Including Wikipedia as part of the score will only adjust the score of individuals who have Wikipedia pages and most people or organizations won't. For example, in the UK PR community, only five of the top 10 in PRWeek's UK Power Book 500 2012 have Wikipedia pages, despite their considerable social capital within the industry. These tools are trying to get better at taking real-world status into account, but it is an enormous task and they are still in the early stages.
The data
To gain an understanding of which individuals score the highest, I looked at all tweets from English-language Twitter accounts with Klout scores of 80+ during a 24-hour period.48 Klout was only selected due to the ease of accessibility of this data via the DataSift platform.49 The 55,000 tweets identified came from 4,301 different accounts, and analysis of a random sample of 100 of them showed that 87 were “Verified” by Twitter. Verification is used by Twitter to establish the authenticity of accounts and is generally applied to well-known individuals and organizations. Of the remaining thirteen, ten were relatively well-known individuals, brands and outlets, including Elvis Costello, audio brand Dolby and Metal Hammer magazine.
So, to state the obvious, the people and organizations ranked as having more social media capital were mostly those who are likely to have more social capital in the real world. Unsurprising, but it doesn't mean that their scores are without meaning, as they do provide evidence of who is investing most successfully in engaging with their publics on social media.
As social media develops, it seems likely that almost every notable person or organization will have to become part of the online community. After all, recent additions to Twitter include UK Prime Minister David Cameron and the Pope. This could mean that ultimately the online world will mirror the offline one, with differences reflecting relative levels of investment and tactical advantage. When this happens, the scores will tell us who is using social media best to interact and engage and, possibly, who is managing to stay ahead of social rankings and game the system. However, as we all become more Twitter-literate and social online there may come a time when there will be a limited amount of this distinction left for social media scoring to illuminate.
Social capital for PR
So, can these scores be useful to public relations professionals? To my mind the answer is a “yes”, albeit a very cautious one. They can provide you with some helpful data for influencer outreach and community development, using the scores and their assessment of relevant influencers on a topic to help start and/or refine an influencer list. They could also prove valuable in crisis communications and reputation management, to help identify members of the online community who talk about you and who may have the most potential to damage or enhance your brand's reputation within social media.
Another more recent development is the use of such scores and analysis in an internal communications context. Here such measures of relevant social capital could be used to identify the people within an organization who have expertise in a particular area, who are most likely to engage on a topic in internal discussions or have the widest internal networks.
In all these situations, social ranking tools can be useful. However, they are relatively new and require, like much new technology, consistent analysis and sense-checking. For example, when using the scores to assess levels of social media capital within your team, don't overlook less active but well-connected individuals. When building lists for influencer outreach, keep in mind that relevance is everything: throw out wild-cards, only compare scores across individuals who are broadly similar, and remember the importance of assessing an individual’s real-world social capital.
The value of social capital
Social capital is purported to be a measure of an individual's potential to exert influence. Assessment of an individual's social capital can be a powerful tool to help PR professionals identify relevant people or organizations with whom to connect.
Online social capital ranking tools are far from perfect; they are still very young, filling a niche created by the growth of social media, in itself relatively new. Social capital scoring is useful: it can enrich and expedite the process of finding and ranking relevant influencers for outreach and reputation management. We should use them, but we should do so with an active awareness of their limitations, and a healthy degree of human judgement regarding each individual's real-world “clout”.
Biography
Adam Parker MCIPR (@AdParker) is chief executive of RealWire, the UK media intelligence company, and is the architect of its Lissted application – a tool for discovering influencers on Twitter. Adam also contributed the chapter “Media Relations Modernised” to Share This, on the importance of putting Twitter at the heart of influencer relations. He is a chartered accountant and previously spent nine years with PwC in its audit, corporate finance and consulting practices. He occasionally blogs at www.showmenumbers.com.
Notes
37ONS Guide to Social Capital: http://cipr.co/YR65AP
38R. Putnam, Bowling Alone: The Collapse and Revival of American Community, Simon & Schuster, 2002, ISBN 0743203046
39N. Lin, Building a Network Theory of Social Capital: http://cipr.co/XlDtc9
40Influence definition: http://cipr.co/11PNbq5
41Michael Conroy, Mapping Social Media Influence: http://cipr.co/VGK3ja
42White Whale definition: http://cipr.co/XJYnmY
43Altimeter Group, The Rise of Digital Influence: http://cipr.co/WXxHhK
44Klout score: http://cipr.co/YM1N9X
45Kred rules: http://cipr.co/VMogGU
46PeerIndex FAQ: http://cipr.co/W6RHjH
47Sean Carlos, Can Social Influence Be Distilled Into A Score? Part 2 – Potential Pitfalls: http://cipr.co/WAxcxs
48Period analyzed was 15:00 17 December 2012 – 15:00 18 December 2012
49DataSift: http://cipr.co/XhEEwW