Social Media Index

I write this post with some trepidation, because we have come up with a new list and lists I have learned are controversial. Some weeks ago Jonny Bentwood who runs tech analyst relations at Edelman’s London office and I were debating influence and how the Facebook phenomena has changed things. When people talked about on-line influence in the past they were often referring to bloggers and Technorati scores, though obviously influence was always more complicated than that. But now with the increasing mass adoption of Twitter and Facebook and favourites listings like Digg and Del.icio.us things have moved on. Bloggers Twitter and Facebookers Dig. Many of us are multi-platform users and so our online ‘footprint’ is much more dispersed.

So we thought we would have a bash at measuring it. Underlying this attempt is the presumption that a cumulative and comparative score can be assigned for an individual’s use of the various social media platforms that are available. That is no small presumption and the assigning of mechanical scores is a blunt instrument to say the least. How do you compare the influence of a Hugh MacLeod cartoon to a Robert Scoble tech review? Technorati say this about their methodology: “On the World Live Web, bloggers frequently link to and comment on other blogs, creating the type of immediate connection one would have in a conversation. Technorati tracks these links, and thus the relative relevance of blogs, photos, videos etc”. So putting numbers to these things and assuming a level of influence from them is not exactly new.

What we have tried to do is add some new ways of measuring influence via platforms like Twitter and Facebook to blog scores. This is definitely adding apples to oranges we admit. So for example, we are placing a score for Facebook depending on the number of friends someone has. For Twitter, it is the number of friends, followers and updates. And if that is not insulting enough, we are then coming to a comparative weighting of someone’s Facebook score against their Twitter and blogging score. And the most sinful step is of course the final one where we have added those scores together and come up with a total Social Media Index. Which is an A list or a league table by another name I suppose. But we are not claiming it is definitive (how can it be with as many value judgements as I just confessed too) and we’re not entirely sure if the thesis itself will stand up. What we hope to do by this is to contribute to the debate and (and if possible) come to a community view of how you look at this. Why? Well, for us it is part of our business and I make no apology about that. Measurement is central to what we do and social media is massively powerful and we need to understand it to do our jobs. Many people do not like this and for them there is no good way of describing this.

So stage one; we created a Top 30 bloggers list from CNET Blog 100, Times Top 50 Business Blogs, Power 150 Top Marketing Blogs, Friendly Ghost Top PR Blogs and Technobabble 2.0 Top Analyst Blogs. A familiar enough grouping, but if you look along the top line you will see the new platforms we intend to combine with this listing.

Top 30 Blog – SMI weighting: 100% blogs

  Name blog multi format mini updates business cards visual favourites social media index

1

TechCrunch

98

0

0

0

0 0

98

2 Search Engine Watch 98

0

0

0

0 0 98
3 Boing Boing 98

0

0

0

0 0 98

4

GigaOM

97

0

0

0

0 0

97

5

Micro Persuasion

96

0

0

0

0 0

96

6

Scobeleizer

96

0

0

0

0 0

96

7

Scripting News

96

0

0

0

0 0

96

8 John Battelle’s Searchblog 96

0

0

0

0 0 96
9 Techdirt 96

0

0

0

0 0 96
10 Pronet Advertising 96

0

0

0

0 0 96

11

Gaping Void

95

0

0

0

0 0

95

12

Marketing Pilgrim

95

0

0

0

0 0

95

13

Online Marketing Blog

95

0

0

0

0 0 95
14 Copyblogger 95

0

0

0

0 0 95

15

Buzz Machine

94

0

0

0

0 0

94

16 SEOmoz Blog 94

0

0

0

0 0 94
17 Jonathan’s Blog 94

0

0

0

0 0 94
18 The Blog Herald 94

0

0

0

0 0 94
19 direct2dell.com 93

0

0

0

0 0 93
20 Romenesko 93

0

0

0

0 0 93
21 PaidContent.org 92

0

0

0

0 0 92
22 Secret Diary of Steve Jobs 92

0

0

0

0 0 92

23

Web Strategy by Jeremiah

91

0

0

0

0 0 91

24

Adrants

91

0

0

0

0 0

91

25 Seth Godin 91

0

0

0

0 0 91
26 Logic+Emotion 91

0

0

0

0 0 91
27 tompeters! 91

0

0

0

0 0 91
28 GrokDotCom 91

0

0

0

0 0 91

29

PSFK

89

0

0

0

0 0

89

30 adfreak 89

0

0

0

0 0 89

This next chart is the same 30 people, but now we have added individual scores for their use of non-blogging platforms and in the final column come to a total Social Media Index or score for them. Here’s how we did that:

Each person has been given a score out of 10 based upon 6 criteria:

  • Blog – analysed Google Rank, inbound links, subscribers, alexa rank, content focus, frequency of updates, number of comments
  • Multi-format – analysed Facebook – number of friends
  • Mini-updates – analysed Twitter – number of friends, followers and updates
  • Business cards – analysed LinkedIn – number of contacts
  • Visual – analysed Flickr – number of photos uploaded from the person/s or about the person/s
  • Favourites – analysed Digg, del.icio.us

Each score out of 10 was the given the following weighting across the categories : Blog – 30%; Multi-format – 20%; Mini-updates – 25%, Business cards – 7%, Visual – 3%; Favourites – 15% which created a total score for each category. The sum of each of these numbers created an individual’s Social Media Index. Clear as mud. And about as appetising for some I suspect, because the weighting system is massively subjective. I repeat, it is our first stab at it and we are interested in your take.

And these results are vaguely embarrassing for us too, because the conspiracy minded will have noticed that Edelman’s own Steve Rubel of Micropersuasion fame is (drum roll) number one in this (surprise, surprise) Edelman league table. And we know how well that will go down. But the fact is he is so prolific across all platforms it is what it is. Personally I think he needs to do some client work some time.

Top 30 Blogs – SMI tiered weighting
Weighting: Blog – 30%; Multi-format – 20%; Mini-updates – 25%, Business cards – 7%, Visual – 3%; Favourites – 15%

  Name blog multi format mini updates business cards visual favourites social media index

1

Micro Persuasion

29 17 25 7 2 13 93
2

TechCrunch

29 20 22 0 3 13 87
3

Scobeleizer

29 20 25 0 3 9 86

4

GigaOM

29 20 13 7 3 13 85
5

Gaping Void

28 20 22 3 2 5 80
6

Scripting News

29 16 25 4 2 4 79
7

Web Strategy by Jeremiah

27 15 24 6 3 2 77
8

PSFK

27 16 14 0 3 12 72
9

Marketing Pilgrim

28 8 15 6 2 7 67
10

Adrants

27 5 19 4 1 8 64
11

Online Marketing Blog

28 15 12 0 2 5 62
12

John Battelle’s Searchblog

29 17 0 7 3 5 61
13

Buzz Machine

28 20 0 5 2 5 61
14

Copyblogger

28 13 3 0 0 8 53
15

Seth Godin

27 5 0 0 3 9 45
16

Techdirt

29 0 0 0 3 11 42
17 SEOmoz Blog 28

0

0

5 3 5 41
18 Search Engine Watch 29

0

0

0

0 9 38
19 Boing Boing 29

0

0

0

3 5 38
20 Jonathan’s Blog 28 1

0

0

2 4 36
21 PaidContent.org 27

0

0

0

3 4 34
22 Secret Diary of Steve Jobs 27

0

0

0

0 6 34
23 adfreak 27

0

0

0

0 6 32
24 The Blog Herald 28

0

0

0

0 3 31
25 direct2dell.com 28

0

0

0

0 3 31
26 Pronet Advertising 29

0

0

0

0 2 31
27 Romenesko 28

0

0

0

0 3 31
28 Logic+Emotion 27

0

0

0

1 2 30
29 tompeters! 27

0

0

0

0 3 30
30 GrokDotCom 27

0

0

0

0 3 30

But this list just orders in a new way, a list of top bloggers. We then looked wider a-field to try and come up with a more pure Social Media Index where we have added top scorers not restricted to the blogging top 30. Same methodology, just a wider catchment group. Yes and Steve is still number one.

Top 30 social media index
Weighting: Blog – 30%; Multi-format – 20%; Mini-updates – 25%, Business cards – 7%, Visual – 3%; Favourites – 15%

  Name blog multi format mini updates business cards visual favourites social media index

1

Micro Persuasion

29

17

25

7

2

13

93

2

TechCrunch

29

20

22

0

3

13

87

3

Scobeleizer

29

20

25

0

3

9

86

4

GigaOM

29

20

13

7

3

13

85

5

Gaping Void

28

20

22

3

2

5

80

6

Scripting News

29

16

25

4

2

4

79

7

Web Strategy by Jeremiah

27

15

24

6

3

2

77

8

/Message

25

20

23

0

3

6

76

9

POP! PR Jots

21

19

23

6

2

3

72

10

PSFK

27

16

14

0

3

12

72

11

James Governor’s Monkchips

26

12

21

5

1

6

71

12

hyku | blog

23

15

20

5

2

3

68

13

Ars Technica

18

16

25

2

3

4

67

14

Marketing Pilgrim

28

8

15

6

2

7

67

15

Marketing Nirvana

23

16

15

6

2

4

65

16

a shel of my former self

24

15

15

5

2

4

65

17

Adrants

27

5

19

4

1

8

64

18

People over Process

23

5

23

5

1

5

63

19

What’s Next Blog

25

5

23

1

2

6

63

20

Strumpette

24

15

21

0

1

2

62

21

Online Marketing Blog

28

15

12

0

2

5

62

22

John Battelle’s Searchblog

29

17

0

7

3

5

61

23

Buzz Machine

28

20

0

5

2

5

61

24

Tecosystems

24

4

17

5

3

6

59

25

The Groundswell

26

11

12

6

0

4

58

26

Marketing Begins at Home

20

12

18

4

3

1

58

27

Common Sense PR

20

15

15

3

0

4

57

28

PR meets the WWW

19

12

15

4

1

5

56

29

Russell Davies

25

0

22

0

3

6

56

30

Drew’s Marketing Minute

26

5

17

6

0

1

56

So what does this last list mean? The overwhelming majority of new entrants to this more ‘pure’ Social Media Index are individuals which is probably not surprising given that corporates or even collectives don’t really use Twitter or Facebook . . . people do. Obviously each platform has different primary functions and some are much more personal (Facebook) than others. But bloggers quite openly use Twitter and Facebook and MySpace to market their blog posts and many blogs these days have widgets cross marketing the individual’s Facebook or Twitter profiles. And the personal and the professional was a line blurred for many of us years ago.

There are of course many platforms that we did not include in this, like MySpace, Jaiku and Pownce and of course this list is very English-language centric and includes none of the local social sites which dominate in countries like Korea and Germany. We may well do this in draft two depending on what people think, but for now we thought we had enough in for the basis of this to be discussed.

Obviously Jonny Bentwood and myself will be delighted to get your feedback. If people think it is worth trying to come to some sort of wider community consensus on this we can look at putting that to a wiki or even hosting some sort of debate at a later date.

Methodology

a) Blogs

Google PageRank

Google PageRank is a link analysis algorithm that interprets web links and assigns a numerical weighting to each site. High-quality sites receive a higher PageRank. The ranking uses the actual PageRank as part of its algorithm.

Bloglines Subscribers

Bloglines displays the amount of subscribers each blog has to its feed(s). Subscriber ranges were determined and each range was assigned a number that was used as part of the algorithm.

Technorati Ranking

Technorati ranking relates the authority of a particular blog (via the number of sites pointing to it). The more link sources referencing your blog, the higher the Technorati ranking. Similar to the Bloglines Subscribers value, and each range was assigned a number that was used as part of the algorithm.

Content/Frequency/Comments

Scores with strict criteria were assigned to content focus, frequency of posts and number of comments. The combined score was used as part of the algorithm.

Alexa Ranking

Alexa ranks web pages based on usage of millions of people. It is a combined measure of page views and users (reach). As a first step, Alexa computes the reach and number of page views for all sites on the Web on a daily basis. The main Alexa traffic rank is based on the geometric mean of these two quantities averaged over time (so that the rank of a site reflects both the number of users who visit that site as well as the number of pages on the site viewed by those users). Ranks closer to 0 have been assigned a high number that was used as part of the algorithm.

b) Multi-Format

Facebook Ranking

As a multi-format tool, Facebook was selected to identify a persons influence/popularity. Other formats such as MySpace could be used at a latter date. The number of friends was determined and each range was assigned a number that was used as part of the algorithm.

c) Mini-Updates

Twitter Friends and Followers Ranking

As a multi-format tool, Twitter was selected to identify a persons influence/popularity. Other formats such as Pownce could be used at a latter date. The number of friends and followers were combined to determine a total friends and followers score. Each range was assigned a number that was used as part of the algorithm.

Twitter Updates Ranking

The number of twitter updates was determined and each range was assigned a number which was combined with the friends and followers score to give a total that was used as part of the algorithm.

d) Business Cards

LinkedIn Ranking

For Business Cards, LinkedIn was selected to identify a persons influence/popularity. Other formats such as Plaxo could be used at a latter date. The number of connections was determined and each range was assigned a number that was used as part of the algorithm.

e) Visual

Flickr Ranking

For visual tools, Flickr was selected to identify a persons influence/popularity. Other formats such as YouTube could be used at a latter date. The number of pictures about uploaded about an individual or about that person was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.

f) Favourites

Digg Score

The number of Digg’s an individual has had was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.

Del.icio.us Own Library

The number of pages in an individuals own del.icio.us library was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.

Del.icio.us Others Library

The number of pages of an individuals own postings that have been saved in other users’ del.icio.us library was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.

Moving forward other tools such as Reddit will be included to make the scoring more complete.

g) Flexible weighting

Each specific social media outlet listed above was given a standard score out of 10. Using a flexible weighting scale it is possible to vary the importance of the different tools available and consequently establish different total scores of individuals web influence. For the table listed above the following weightings were used:

Blogs 30%
Multi-Format (e.g. Facebook) 20%
Mini-Updates (e.g. Twitter) 25%
Business Cards (e.g. LinkedIn) 7%
Visual (e.g. Flickr) 3%
Favourites (e.g. Digg, del.icio.us) 15%

The weighting scale listed above was created through discussions with several key new media gurus. I do not anticipate this weighting scale to be the final standard and welcome everyone’s views as to what this should be.

Future copies of the Social Media Index will allow you to assign your own subjective weightings to the index to establish your own score.

David Brain

149 Comments

  1. Need time to absorb this, but I like the idea in general. I like the flexible weighting. Not sure I’m sold on the categories though–they seem a bit too application-specific and arbitrary. But the ability to weigh them compensates for this quite well. I might recommend different headings for the categories, focused on what you’re actually measuring in that category. For example, it’s not business cards, but “size of rolodex” or something like that… Again, just quick thoughts here.

    The real question is how do you get access to all the data you use to measure influence, e.g., facebook friends lists that may be closed unless you yourself are a friend. How are you getting access to this data? Self-reporting, primary research or third-party data?

  2. cool beans. how do i get on your blogroll? i need to improve my stats. and perhaps you could join the friends of redmonk FB group and twitter/monkchips while you’re at it… 😉

    Seriously this is an interesting approach, and obviously its nice to be well represented on a list like this, given there is some, if rough, quantitative elements to this. But key to this is its really about what you put in rather than what you get out. Your metrics appear to reward those that make contributions to social networks, rather than just sucking it up.

    I suspect counting del.icio.us via’s is a surprisingly effective influence counter.

    Well done for supporting Jonny on this, and encouraging an open debate

  3. I understand the methodology and of course it is arguable but the bigger question in my mind is: who is being influenced? Do we know for instance whether there is a clear impact on general buyer behaviours? Do we know much or indeed anything about influence reach? What can we say about niche markets? What about what I term ‘the groupie effect?’

    The reason I ask this is that in all the years I worked as a tech journalist, I always knew that the people who were most influenced were the PRs, not buyers. Why? Because readers’ letters were usually faked on most titles including those with ABCs running to 00,000s. Whether that is changing now we have blogs is another matter.

  4. “The number of pages of an individuals own postings that have been saved in other users’ del.icio.us library was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.”

    interesting – how’d you do this, if i can ask?

  5. James,

    yes these stats are weighted towards givers rather than takers . . influencers rather than the influenced I guess.

  6. Dennis,

    yes all good questions. Seems to me this is a bit like peeling an onion in that there is always more to know. My own belief is that it is almost impossible to see where “influence” drives sales, which I guess is part of what you are saying.

  7. Quote from Todd Van Hoosear

    The real question is how do you get access to all the data you use to measure influence, e.g., facebook friends lists that may be closed unless you yourself are a friend. How are you getting access to this data? Self-reporting, primary research or third-party data?

    Allow me to answer this for David as I created methodology behind this.

    We can only measure the data that is publically available with all data being sourced through our own primary resaerch (i.e. no self reporting or 3rd party data).

    With regard to Facebook, I have only included numbers for those people who have made their friend list public. In my opinion if someone aims to use the different social media tools available to promote their voice, they shouldn’t be restricting people from hearing what they have to say.

    Again let me reiterate, all this is open to discussion – if the consensus is to change this, then I will happily adapt the methodology.

  8. Taking the liberty to answer another question…

    Quote from Stephen O’Grady

    “The number of pages of an individuals own postings that have been saved in other users’ del.icio.us library was identified and assigned a number. Each range was assigned a number that was used as part of the algorithm.”
    interesting – how’d you do this, if i can ask?

    It’s a little complicated but bear with me. To take each point in turn…
    del.icio.us – in your case I checked your library: http://del.icio.us/sogrady and then added this total to a search within del.icio.us for other saved posts that include your url. This gave a total del.icio.us figure.
    This score was assigned a number depending on the range it was part of (i.e. between 10 and 20, 21 and 30 etc).
    A similar process was done for Digg and this score was added to your del.icio.us score which gave you a ‘favourites’ score.
    The next process was to standardise each score so that they can be used as comparative measures depending on what weighting scale was used.
    Rinse and repeat for other parts of the SMI.
    Hope this makes sense but happy to clarify in greater detail should you wish.

  9. Not wishing to diss anyone’s profession I’ve consistently found that when it comes to spend influence, the most powerful voices are those of customers. Analysts and journalists are very low down the pecking order.

    The other problem ‘we’ all face right now is that blog influence may be deep in certain places but it sure as heck ain’t wide. Although I detect some change, I suspect we’re still talking to ourselves a lot of the time.

  10. Dennis,

    I know what you mean but in some cases and on some points the blogs are on-ramps to mainstream media . . . though mostly they are not. Actually, one point on this exercise might be to add what mainstream media coverage a blog or it’s author gets. Jeff Jarvis and Steve Rubel are high on that and I guess therefore influential in the old world as well as the new.

  11. ah, i see. i’d thought you’d magically solved the whole domain problem in tracking del.icio.us mentions. in other words, for all of us, that’s just a partial list in that it doesn’t track specific del.icio.us or digg mentions for items on the respective domains, so that specific pieces like this:

    http://redmonk.com/sogrady/2007/07/05/google_apps_migration/

    aren’t necessarily tracked – just the parent domain. and even there, it could suffer from deltas between, say, redmonk.com/sogrady and http://www.redmonk.com/sogrady.

    but i don’t have a solution to that problem either, and it certainly applies to all of us equally, so kudos.

    v interesting list.

  12. In my experience Facebook is replacing LinkedIn as the new business card. So, having your phone number and email address available on Facebook should count as being a business card.

  13. I can see that for you Robert and for the people you have as friends, but many of mine are still in the old world and I’m stuck with both. . . that said, Facebook is taking increasing amounts of my email traffic.

  14. I think this index will spark a great discussion; thanks for putting it together. A few thoughts:

    1. Often you are dealing with the same audience across the non-blogging platforms. In other words, the same people who subscribe to Scoble’s blog also get his twitter feed and are his friend on facebook. This means that his footprint is not necessarily wider, just arguably deeper – which may or may not be meaningful.

    2. I fear you are falling into the trap of equating reach with influence. There are several fundamental questions that this index fails to address:
    Who are these people reaching?
    What is their content addressing?
    What are they influencing?

    Romenesko is very influential in the journalist community, but no so much in terms of fashion. I might trust Steve Rubel when it comes to new media, but his restaurant recommendation likely doesn’t mean much to me. Any one of these people may get me to look up a product on Amazon, but if the user reviews there are terrible, will I still buy the product?

    The list above is loaded with a bunch of tech and communications media platforms, but I don’t think they are driving the majority of decisions about what to buy, who to vote for or what to go and see.

    Many of these comparisons are apples and oranges. Subsequent indexes may be better informed by defining communities and the type of influence being examined.

    3. The index is biased toward individuals as organizations are less likely to be active in social networking based applications (then again, social media in general is biased toward individuals who are free from the constraints of organizations). Notice how Direct2Dell, Jonathan’s Blog and Romenesko (Poynter) drop further and further on each index. Yet the index fails to account for something like Dell’s Community Forum or their Second Life presence.

    The index certainly calls attention to the fact that blogs are only the tip of the iceberg in terms of social media, but I feel the idea of influence requires a deeper understand of what drives decision making at the user level – let’s not give too much credit to those talking the loudest, or most often. Participation in the conversation is certainly meaningful – but in what ways?

    (full disclosure: I work for Edelman in Chicago)

  15. The third and more diverse Social Media ranking seems to be less likely to be skewed by a huge score in one area, for instance, a huge number of LinkedIn contacts. Given the networking skills behind MicroPersuasion, it’s not surprising that it tops each list. (You may have the most physical business cards, too!)

    Overlap would be interesting to try to calculate. Of the blogs and sites listed, I know that I personally would count double or triple on some (LinkedIn, Blog subscriber and Twitter, for instance) and would be counted on more than one on the list. A measure of unique individuals would be helpful, as would an average of the number of methods by which a user connects with a site.

    Another key metric that could be useful would be feed subscription and aggregation data definable by service or tags.

    As mentioned, adding sites like MySpace, del.icio.us, Gather and Squidoo may have a significant impact on any of these index calculations, as well as raise questions of definition–Is a MySpace or Yahoo! 360 blog a blog or an element of the platform? Where would a Twitter read on a blog widget fall? (Although Neilsen’s new metrics may make this easier to quantify.)

    Most important, who are those influenced? Defining the scope of influence that the scores represent in the above social media index group would, I suspect yield numerous profiles of PR, marketing and tech folks. To what extent are we talking with each other? (I’ve noticed quite a bit of buzz within this community about these lists.)

    Where might a blog like TMZ, which is arguably quite influential in entertainment, fall?

    The readers of top blogs in the social media index may not be an attractive target segment for consumer products marketers, for instance, but the same marketers might be extremely interested in readers of the TMZ blog.

    To the extent that social media can be defined as an interest group, as well as describe the modes of communications those in the group use, this methodology could be especially valuable applied to various influence segments and vertical markets to create vertical indices.

    My thought is that to do this in a way that reflects authentic influence will require more analysis of the characteristics of the contacts and readers, and, ideally will measure elements that difficult for spammers to exploit.

    Thanks for putting this together. It raises lots of interesting thoughts about the scope of social media and its influence on thoughts and behavior.

  16. Well make a widget or a thingie that does all this culling and parsing of my stats on all those platforms automatically for me, so I can see how I come out, then I’ll let you know : )

    It sounds horribly corporativist and rigged so far, however because so many of these social media thingies are highly subjective, twitter might be down, somebody joins jaiku or never comes back or goes to Pownce, oops, there wonderful big post and friend stats on Twitter go down the tube. Perhaps if there was a cumulative footprint over 90 days…somehow, it just isn’t feeling right, using some of the gamed and rigged internal metrics of these silly social media things to then arrive at an ostensibly more objective metric for SMI.

  17. Well, it’s a good idear and a good start. I wonder if technorati or google will change their measure standard accordingly? It seems that they should do it long long before. Did you guys apply for a patent or something?
    What about if they follow your guys’ idear, and make a little change, and use another different name instead of SMI?

  18. Pingback: PR 2.0
  19. Pingback: GREENblog
  20. Jeff, you make some very good points.

    I agree that we may often be measuring a deeper rather than a wider score when the same person reads Scoble on Twitter and on his blog. It’s similar in the off-line world too when we talk about readership reach for magazines and newspapers and you sometimes end up with a figure that is higher than the population of the country in which they are published. Possibly, repetition is useful, but I suspect that this is one of the areas where measuring in this way comes up against a hard stop as I fear we will never be able to detect multiple “readers” of Scoble (for example) across platform.

    I also agree that reach is not influence. Perhaps that would be the next step, though it would be good to hear how you think we might measure it.

    Thanks for your comments and interest . .and watch this space for the next version.

  21. Susan Heywood.

    Wow, all great thoughts thank-you. Difficlut to argue with any of them really. i definitely think that next iteration needs to offer a more segmented approach and I am stumped on how to siphon out the overlap issue.

    Thanks for this . . .all taken into consideration for next version.

  22. Profky Neva.

    Great thought on the ninety day (or period of time) footprint. That might even out some of the bumps.

    Thanks for that.

  23. Where is the human element? This study confirms that many are so engrossed in the tools and popularity contests versus developing relationships. My virtual “rolodex” is not be pages long because I believe in the power of direct communication — Please note this is not anti-technology because I am a geek. Awareness does not mean anything. Action is everything. Is the content useful and acted upon?

  24. Lauren I think that is at the heart of this but what we have tried to do is take one big value judegement (what is the social media influence of person X) and break it down into smaller chunks which are also (I admit) value judgements. And you are right, this dos pose problems in terms of awareness versus influence and on that I have no answer but I suspect no-one does. If you are focused on the purity of the relationship then this sort of output model will never satisfy I suspect. Any suggestions on measuring relationships? there is a body of work out there looking at that by the way, but not about on-line relationships.

  25. This so ties into some thinking I’ve being doing on niche influencers. For me, it gives a baseline on how to compare influencers within a category. None of the people listed above would feature on much of the stuff that I do, but I could use the process to look at the most influential in knitting or cooking or beauty etc.

    Overlaying it needs to be a way of measuring the impact, if you are using it in a marketing sense. looking at reference since the network before and after any effort is necessary – which would then have to feed into the influence measure.

  26. Yes we are of the view that it becomes more useful the more niche or sectoraly it is used as the comparisons become more valid.

  27. Hi Rachael

    You are dead right. Once the methodology is agreed (which is the hard part) the SMI can be applied to any group of people.

    One thing I am continually struggling to find an answer to is how to measure the importance of the friends you have (for example in twitter) compared to the number.

    For example, I would be far more influential if I only had 2 friends listed but they were CEOs of Fortune 100 companies compared to 100’s but all university mates.

    Google Rank provides this in one way but unfortunately only in one medium – perhaps the next step is to analyse the profile of the people in a list and score them too. This raises even more concerns but maybe this suggestion should be thrown into the ring with all the other comments so that the community can help forge a methodology that everyone (or most people) can agree with.

  28. IMO, this is a very impressive first step in what will surely be one of the most ambitious initiatives that the web has seen to date. However, I would suggest that for the next iteration, you guys might want to consider reaching out to academic researchers, some of whom have been involved in social network analysis for many years (esp. the collective dynamics group at Columbia University in New York).

    I would probably state FWIW that it might be problematic to conflate, as others have written above, awareness with influence. Furthermore, how do you account for the effects of the content itself? _Regular_ blog reading is, according to most of the studies we’ve produced, by no means a mainstream activity, although there appears to be a tremendous variance in terms of what kind of content generates the most apparent awareness (tech bloggers come to mind, but we could get much more specific).

    Similarly, despite all appearances to the contrary, the direct reach of bloggers tends to be pretty limited, even with pingbacks etc, possibly unless the blogger also appears in a newspaper or on television. Then again, a VC may make decisions to invest in a given company based in part upon blogger opinions that are not widely read, sparking the company’s growth into a powerhouse (i.e. low awareness yielding high influence).

    In other words, what I am trying to get across is that there are a spectacularly large number of variables to capture well before you consider off-line activity. Moreover it is very difficult to quantify the effects of content (partly because of emotional factors) unless it is tied to a very specific action or “success event” such as the purchase of a product.

    The points above aside, I think that your approach is in general quite sound and marks a major first step in what may be a very long journey indeed.

    Best,
    David

  29. David (July 19),

    Thank-you very much for such a long and thoughtful piece. You are right on so much of this and the semantics of the issue are important too. I think inviting academics to the next stage s a great idea and I have noted the suggested names. Thanks again.

  30. Pingback: isedb.com
  31. Pingback: Anonymous
  32. Pingback: PR2Peer
  33. I’m curious about the Digg score…does the “number of Digg’s an individual has” refer to the number of times a person’s submitted links have been dugg, or to the number of items they’ve voted on themselves, or the number of items they’ve submitted to the service, regardless of whether or not they land on the front page? Or none of the above?

  34. Hi Jason

    Regarding Digg, the numbers are calculated by analysing how often a particular domain has had it’s pages ‘digged’. (i.e. this applies to every page – not just the front page).

    In addition, Digg works by only allowing one user to digg a specific page.

    The Digg score does not include how many times a user has submitted a digg.

    Hope this clarifies things for you.

    Jonny

  35. I have to say, I’m extremely disappointed in this post, and the fact that so little research has apparently gone into this project to determine whether it would be even vaguely valid. If it shows anything, it’s that the people behind it really still understand very little about social media tools, other than how they benefit them directly. I find it so disappointing, because I work not only on the PR side of things, but also the tech end through running around 2 dozen sites of my own on top of my clients predominantly being webmasters and online entrepreneurs (several at the top of their fields, including online communities). Here are some of the fundamental faults with this idea (the same shared by pretty much every other ranking system in existence for blogs):

    1. Facebook – Using Facebook doesn’t demonstrate influence in the slightest. Anyway can start an account, and add random users as friends. It doesn’t mean they actively use the account or exercise any influence over them. On top of it, Facebook isn’t valid for all niches by a long shot, and with Myspace cracking down heavily on auto-friend-adder bots, it leaves Facebook and the other social networks more open to the social media spammers (who can generate tens of thousands of friends quickly, but with no real interest in what they have to say). Facebook being more popular perhaps in Europe is fine, but that only makes it valid (but even that not really) if you’re talking about influence with a European audience, and nothing wider.

    2. Twitter – Twitter’s really such a joke that I’m absolutely shocked to see it here. It’s actually discouraged quite aggressively in the blogging community, as a big contributer “blog clutter.”

    3. Even the mention of Digg and Del.icio.us is amusing. The same goes for things like Technorati. If the PR community actually paid attention, they’d know that these once-useful tools are rendered worthless in gauging popularity with the overwhelming abuse that grows daily. You have paid Diggs, favorite and bookmark exchanges to improve rankings, etc. It’s sad, but it’s a reality in Internet marketing and SEO efforts, and it’s time to stop ignoring it.

    4. Where are you pulling your inbound link count from?? Google, Yahoo!, and MSN can’t even come close to the same estimates (often by the thousands), so there’s no real authority source to pull this info from. Even if you went by Google’s standards because of their market share, you run into the issue that they choose to ignore and devalue certain backlinks for the purposes of their own ranking algorithms… which has nothing whatsoever to do with overall popularity and influence.

    5. Google Pagerank – Obviously the PR community missed the Web memo that PageRank means almost nothing unless you’re selling backlinks on a website based on their ability to pass link juice (which Google actively discourages and is trying to weed out). It’s also incredibly easy to manipulate, and changes radically based on algorithm changes at Google’s whim… not with necessarily with changing influence. You also mentioned backlinks separately in your post. If you’re intending to count them both separately, you’re double-counting the same information, as incoming links are the driving factor behind Pagerank.

    6. Content frequency tells you pretty much nothing, if you can’t vouch for content quality at the same time (and there really isn’t a way to do that… hence the issue with the entire ranking system really). Using it as a standard means that splogs have a better shot in that area than legitimate blogs.

    7. Alexa – I’m utterly amazed that there’s still a living soul that gives any value to an Alexa rank. Again, it shows how out of touch the originators are with the tech side of the equation. Alexa rankings are the most easily manipulated ranking in existence, through simple URL redirects, autosurfing, and browsing your own pages. It’s reliance on the Alexa toolbar means most traffic to a site isn’t even factored into the equation, which is why Alexa has a tendency to be dominated by sites aimed at tech geeks and early adopter-types.

    8. LinkedIn – If you’re only evaluating business-oriented blogs, this might not completely suck, but it’s also not used widely enough yet (even assuming no manipulation, which is highly unlikely) to justify using it as a yardstick to measure influence.

    The most fundamental problem that you’re missing with this social media index is that you’re making a faulty correlation between social media and a blog’s influence, while simultaneously ignoring other extremely strong factors.

    Frankly, a blog’s influence depends entirely on knowing one’s target market and what tools are likely to influence them. Your system can’t account for the fact that these tools simply don’t work all markets in even close to the same way.

    You’re also ignoring things like paid advertising (increasingly popular for driving targeted traffic to content-oriented sites like blogs). You ignore it because it doesn’t fall within your limited scope of social media. Yet it leads to more targeted traffic than a lot of these other sources, and therefore traffic more easily influenced by the blogger. It’s just one example of why you can’t equate a social media “score” to the influence a blog has.

    You also ignore a variety of other factors that can equate to influence, such as media coverage of the blog, interviews done with the blogger, the blogger’s level of communication in non-typical social media environments like forums and comments on related blogs, etc. It all matters. But it’s not all something you can measure in a hands-off, automatic way, with info plugged into an equation. I don’t blame you for not including these things, because of the sheer work that would be involved. But then please don’t claim to be offering an even slightly honest listing of blogs by their level of influence.

    All you really ended up with is yet another list touting the same “top” bloggers (some deserving; some not so much), which in turn will simply drive even more traffic and exposure their way, showing quite directly that influence can be decided by enormously more than social media.

  36. Glad to see that there is at least one other person in the world that can see this for what it is… nice work Jennifer…

    This is not a rank of influence it is a rank of how often certain individuals open their mouths. There is no quality score what-so-ever, there are no indicators as to what audience this stuff is reaching (if any) and the weighting is an absolute joke. Did any thought go into it or were numbers just randomly plucked from thin air? This is just an indicator of how much noise these people are introducing to an already overly congested media with no indication as to what effect this noise is having.

    The appalling lack of understanding as to how social media works that is demonstrated here is shocking. Aside from the fact that having 4400+ friends on facebook is just ridiculous (it is about two-way interaction, does anyone in the world have the time to interact with over four thousand people?), the egotistical use of the word ‘followers’ instead of friends further demonstrates the misuse of the medium… these guys are trying to turn Facebook into blogs. They clearly do not get how New Media and Social Media are being used by the general public and instead are using the excuse of bringing much needed measurement and understanding of the internets impact as part of the marketing mix to blow their own trumpets within their existing sphere of influence.

    I would be more angry about this sorry excuse of a methodology potentially damaging the industry of reputation management but thankfully it is so poor that the vast majority of people will see it for what it is, a cheap publicity stunt by Edelman.

  37. Jennifer and Nicolas I am sorry we disappointed. You make many good points, some of which have been made by other commentators and which we are taking on-board. The key one of these I think was my use of the word “influence” when as you rightly point out much of what we were ‘measuring’ was reach. However, you both talk as if we put this up as some sort of definitive view rather than a “first bash”. In fact, I don’t think we could have been less definitive and more apologetic for the approach then we were in my post. We talk about “presumption” and adding “apples and oranges” and the “insult” of our method of comparative weighting. And we expressly said this was not about the list produced but about how we talk about multi-platform use. We even said “we are not sure this will stand up”. And we expressly asked for input and feedback (and again thanks for yours) so I hope this makes you a little less “angry”.

  38. It’s not so much a matter of being “angry” as being shocked that people in the PR industry, especially at Edelman’s level, are this far behind in understanding what these tools really are, how they really work, and how they’re really being used. There seems to be this assumption that these tools are actually being used as they may have been originally intended by their creators, but that simply isn’t the case. Unless you’re working down in the trenches with the webmasters and online entrepreneurs creating and using these kinds of platforms, and not just jumping on them based on their own buzz after the masses are already ahead of the game, chances are that you’re clueless (not you personally; speaking generally regarding members of this industry).

    The post was definitely extremely disappointing, because obviously thought went into it, and people thought they were doing it “right,” while missing a heck of a lot. The comments were / are definitely meant to be constructive, and perhaps a bit of a reality check. Support for something like this doesn’t necessarily make it a good idea… it does however demonstrate a more widespread naivety that’s a bit disturbing.

    Hopefully Edelman takes the issues to heart and either comes up with a more legitimate ranking model or drops it, so as not to become yet another false source of rankings in the blogosphere seeming to say only that “bigger is better.” At least it would all be worthwhile if at least a few readers here listen to the comments from their colleagues and take this as a stepping stone to finally educating themselves a bit more on the world of social media… and not just from the already completely “lost” PR community.

  39. Jennifer,

    We are taking the comments to heart and will be having another go in some form soon, though I personally believe this will be more in the area of getting a view of reach or footprint rather than ‘influence’ which I think is too big a claim for pretty much any mechanical method (though Peter Hirshberg of Technorati is pretty compelling on why their system can do this). And ultimately of course, the best system is going to be a massively blunt instrument and out of date as soon as published (are we not always in beta?). The “angry” quote was referring to Nicholas the subsequent comment to yours by the way.

  40. This might be helpful for you guys, as you write a number of emails that might want some flavor? Ever want an email that would self-destruct or send an email you don’t want someone to be able to print? Take back that email you wrote to your boss at 2AM

    Check out http://www.BigString.com

    It is a new free webmail program. When you send mail from your BigString account, you are protected. BigString is like an automatic shredder for your email. You can self-destruct or change an email that’s already been sent or read.

    Have a great day!
    Michael
    michaelf@bigstring.com

  41. We’re still a new site, but I dunno, we’ve felt kind of influential.

    Why don’t you skip all the bizarre metrics? You’re a PR company. Just use your own internal data of what impact mentions on the blogs you monitor have.

  42. Interesting that this kind of debate is going on in the broader corporate world but I would have to agree with a few of the skeptics that is all about “noise” (reach to put it kindly) rather than “true influence”. The very nature of social media means that an influential network is not determined by the number of friend links – as is very easy to end up with thousands of friends and spammer adder-ons – so this metric is fundamentally flawed. I would be interested in an influence metric which would assess the number of comments on a site (even if these are inflated by the blog owner replying back to posts or trackbacks). I recently posted about this on my blog because for example MIT are doing some really interesting work uwing a tool to measure the true influence of social media site visitors – which if combined with feed subscriptions, number of comments on blog and technorati rankings (although again technorati is also fundamentally flawed – ie links on one’s own site can influence one own’s rankings). At least this is provoking a broad debate though 🙂
    My thoughts are here in more detail if anyone is interested:
    http://marianina.com/blog/2007/07/02/social-networking-analysis-meets-web-analytics-meets-marketing-effectiveness/

  43. Pingback: PR2Peer
  44. Pingback: Social Media Index
  45. Pingback: Media Guerrilla
  46. The implications of this index are very intriguing though I don’t know of anyone who is interested in Alexa anymore. Google Page Rank seems to be king.

Leave a Reply

Your email address will not be published. Required fields are marked *