The believability of a claim should not depend on the authority or influence of the person that makes it (e.g., I believe your rankings are wrong or, this blog, car, or car seat is better than that one). Instead, its believability should be contained in the evidence itself (for great examples, see Consumer Reports Annual Car Ratings or Stiftung Warentest – children’s car seats’ reliability and usability. So we asked the question:
How does one measure blog influence?
As you will read, challenges for MBA program rankings, car reliability ratings, and car seat reliability ratings vary,
BUT the methodological challenges are quite similar.
For instance, influence is difficult to measure, as is car reliability. To illustrate, ideas written about in corporate blogs are distributed or spread far and wide based on:
1. the blogger’s or reporting entity’s influence, as well as
2. their readers’ openness to being influenced (e.g., Consumer Reports’ approximately 7 million magazine and website subscribers).
According to Aral and Walker (2012), both influence, and susceptibility to influence in the blogger’s network or a magazine’s subscribers affect the pattern of contagion. This also affects someone’s importance to the propagation of ideas and / or behaviours in the population. The authors (see complete citation below) also discovered that men are more influential, as are people over thirty. Even more interesting is that influential people are less susceptible to influence that those with lower levels of influence.
Agarwal, Liu, Tang & Yu (2008) first introduced the problem of identifying influential bloggers. They proposed the influence flow method to find the influential bloggers in a web community. Among other measures (e.g., inlinks and outlinks, length of blog post) they used one particularly interesting proxy measure for influence, called:
activity generation – a blog post’s ability to generate activity is measured by the number of comments. Few comments suggest limited interest in the work, thus it is non-influential.
Akritidis, Katsaros & Bozanis (2011) tried to measure influence using a proxy similar to activity generation, which they called:
proximal impact – a blogger’s influence on the wider community including their readers. This is also assessed through comments received on a post.
The number of comments and how they could be important is not just of interest to bloggers. For instance, Schultes, Dorner & Lehner (February/March 2013) investigated viewer comments on YouTube. The researchers were interested in how those might influence the spread of a video clip across the larger web community. Their findings indicated some clear patterns whereby influential content got spread faster and wider than other content. While we may guess this, confirming such ideas is what matters.
Although some consider the number of comments mainly a vanity metric, tracking your readers’ engagement and reactions to your content has its place. Of course, it is not just the number of comments, but also their depth and breadth. Put differently, a comment stating, “Go get ’em, tiger!” is unlikely to add much to the discussion. A comment of 100 words offering a URL to additional material always does better.
This image illustrates that blog influence also results in social sharing, as discussed below.
We use the length of a post, as well as the length of a comment as heuristic measures to check whether comments are influential in comparison to the blog entry. Agarwal, Liu, Tang & Yu (2008) found that blog post length is positively correlated with the number of comments, which means longer blog posts are more likely to attract people’s attention.
As well, an author that tries to reply to reader comments is more likely to get that person to comment again in the future (if you care about the algorithms, find them here).
Sharing of content
One way to get a better idea is to measure whether users take action. In the context of blog influence, this means whether users actually share content with their friends. In turn, they help spread the message and may help it go viral this way, which could be the start for a word-of-mouth campaign.
Of course, just because people do share content does not automatically result in others clicking on a tweeted link. Nor does it get us to drop everything and run out to buy the recommended item. Nevertheless, people only share content with their social networks (e.g., friends on Facebook) that they like (if you care about the algorithms, find them here).
Akritidis, L., Katsaros, D. & Bozanis, P. (September 2011). Identifying the productive and influential bloggers in a community. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions, 41(5), pp. 759-764. doi: 10.1109/WI-IAT.2009.18. Retrieved April 10, 2014 from http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5284916&isnumber=5284878
Aral, S., Walker, D. (July 2012). Identifying influential and susceptible members of social networks. Science, 337(6092), pp. 337-341. doi: 10.1126/science.1215842. Retrieved April 10, 2014 from http://www.sciencemag.org/content/337/6092/337.abstract
Jimoh, R. G., Awotunde, J. B., Enikuomehin, O. A. (March 2014). Identifying influential bloggers on the web. Computing, Information Systems, Development Informatics & Allied Research Journal, 5(1). Retrieved April 7, 2014 from http://iiste.org/Journals/index.php/CIS/article/view/12009
Agarwal, N., Liu, H., Tang, L. & Yu, P. S. (February 2008). Identifying influential bloggers in a community. In Proceedings of the 1st International Conference on Web Search and Data Mining (WSDM08), pp 207-218, February 11-12, 2008. Stanford, California. Retrieved April 7, 2014 from http://www.public.asu.edu/~huanliu/InfluenBloggers/Identifying_influential_bloggers.html
Schultes, P., Dorner, V., & Lehner, F. (February/March 2013). Leave a comment! An in-depth analysis of user comments on YouTube. Paper presented at 11th International Conference on Wirtschaftsinformatik, 27 February – 01 March 2013, Leipzig, Germany. Published in the Proceedings, pp 659-673. Retrieved April 7, 2014 from http://www.wi2013.de/proceedings/WI2013%20-%20Track%205%20-%20Schultes.pdf
Shola, SivaNaga Prasad (January, 2012). Identifying influential bloggers. Master Thesis, Paper 322. San Jose, CA: San Jose State University. Retrieved April, 7, 2014 from http://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=1321&context=etd_projects
Rankings – cars, children’s safety seats, MBA programs or corporate blogs – are by nature subjective things. They reflect opinions of the designers of such formulas. For instance, most ranking procedures follow these steps:
1. Use several component scores.
2. Rate each component, using survey, collected data (e.g., from crawled blogs and MBA student surveys).
3. Convert component scores to a common scale, such as 0, 50 to 100 (latter top score, 50 middle).
4. Determine the relative importance of each component and calculate the aggregate score as the weighted sum of these components (e.g., factor x gets .20 weight and factor y is given .25 weight when added together).
5. Express the aggregate score in the desired scale (0 to 100 for the DrKPI BlogRank).
Accordingly, comparing rankings is a risky business because each focuses on different matters (e.g., Financial Times, Business Week). Even Consumer Reports’ annual Car Reliability Ratings are subjective to some degree (e.g., weighting, subscribers’ surveys, etc.):
“We provide reliability information in several forms. For used-car buyers we give Ratings for 17 different trouble areas over ten model years, so you can see a model’s individual strengths and weaknesses. We also provide a Used Car Verdict for each model year that sums up its overall reliability. The verdicts are weighted to emphasize areas such as the engine-major, transmission-major, cooling system, and drive system, which can be more serious and expensive to repair.”
As the above illustrates, the verdicts are weighted and the latter are chosen by Consumer Reports (although we can agree that engine-major and drive system are important). Even using the same weight for each area as we do is subjective (i.e. it assumes each measure component is as important as the other).
Source: DrKPI: Best blog buzz
Have you got an example of blog content that just went viral?
What actionable metrics do you use to see whether your blog has influence?
Which key performance indicators (KPIs) do you use to improve your blog’s influence?
Thanks again for sharing your insights – I always appreciate your very helpful feedback.
His latest book, Social Media Audits: Achieving deep impact without sacrificing the bottom line was published in April 2014 by Chandos Publishing / Elsevier – blog readers => grab your 25 percent discount with free shipping now.
This post is also available in: Englisch