BuzzLogic calculates social media influence

BuzzLogic calculates social media influence

Summary: BuzzLogic hopes to solve one of the riddles of the Web--who is influential within the millions of conversations taking place in social Web, especially blogs. I have been following the progress of BuzzLogic for a few years.


BuzzLogic hopes to solve one of the riddles of the Web--who is influential within the millions of conversations taking place in social Web, especially blogs. I have been following the progress of BuzzLogic for a few years. Mitch Ratcliffe, an old friend who also blogs for ZDNet is a co-founder and continues to evangelize for the company. BuzzLogic started out mapping the connections among bloggers in a visual array, but has morphed into a potentially powerful tool for gaining insight into who wields influence, which can determine the fate of products, brands and reputations.

Delivered as an on demand software service, BuzzLogic looks at content relevance, using full text analysis; the number of messages from a certain publisher over time; traffic, including users referred by influencers and total number of pages views; and attention to messages via link analysis, including temporal data and total number of inbound links.

It indexes 7,300 mainstream media sources, social media sites and corporate site as a basis for its applying its algorithms, said Robert Schettino, BuzzLogic chief marketer.  BuzzLogic has a "reach" calculation, which shows how many sites a blog relevantly reaches on a specific subject. "By calculating influence on the fly and who merits attention, advertising support can be directed at any given moment based on relative influence," Schettino said. "It's also a way for bloggers to monetize their content."


In a nutshell, for BuzzLogic an influencer is defined as a "post or publisher generating a significant volume of relevant inbound links and comments about a particular topic or conversation, within a specific timeframe."

Beta tester David Churbuck, vice president, of Global Web Marketing for Lenovo, provides a press release quote, saying that BuzzLogic "promises to do for brand communications and customer satisfaction, public relations and press relations what Web metrics have done for online advertising.”
BuzzLogic does appear to fill a gap in extracting useful statistics out of the growing body of social media (mostly blogs), including mapping which influencers are influencing other influencers in near real time.

Pricing for the service will be "disruptive," Schettino told me, starting at $500 per month to track 20 conversations (queries), 25 influencers deep. The product is currently going through a second phase of beta testing. In addition, the company plans to publish APIs that will hook into external business intelligence applications.

BuzzLogic is not alone in pursuing influencer metrics--companies such as Nielsen BuzzMetrics and Cymphony are also measuring influence by trolling the social media space. The real test for BuzzLogic will be the degree of accuracy and currency produced by its algorithms and index. "As we look at posts, we are providing some context to what makes a site or individual influential over time. The system architecture has to be sensitive and changes over time," said Todd Parsons,  executive vice president of products. Spam could also be a problem. "We are cross checking blogs to verify and parsing out a few things," said Parsons said. "Cues from blogs indicate they are part of a trust network and we limit the index to those, so we get a higher quality and large index."

If you want to understand better the role of influence and attention in shaping opinion, read this post from Mitch. Here is a sample:

Influence, like gravity, is the mechanism that moves the networked market. The mass of planets account for gravity, the work of people produces influence. How people take raw data, in the form of attention or preferences, and make value isn't fully explained, but it is measurable today and will continue to be more accessible to our analysis in the future. 

Topic: Social Enterprise

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • "Linked to" != "Influential"

    This kind of stuff is reason #298 that "Web 2.0" and "Blogosphere" people have little connection to reality. To beleive that just because a lot of people link to something or comment on it does NOT mean that it is influential, it merely indicates that people found it noteworthy. There is a world of difference.

    What is more "influential"? A foreign policy blog with three regular readers (Tony Blair, George W Bush, and Vladimir Putin) who never link to it or comment on it? Or a blog post containing pictures of Tom Cruise's kid with 2,000 "She's so cute!" comments and 10,000 inbound links? In fact, I would say that a memo sent from a senior editor of the [i]New York Times[/i] or [i]The National Review[/i] or any other major publication to their staff is 1,000 times more "influential" than any blog, since it will shape the direction that massively consumed content will take, despite being read by only a few dozen people.

    Sorry, but in a world where even founding members of Wikipedia have doubts about "truth by common agreement," the idea of determining "influence" based upon links is absurd. Britany Spears is popular, it does not make her important.

    Justin James
    • Actually, Justin...

      the BuzzLogic algorithms do much more than count links and conclude the most-linked posting is influential. Spend some time on our stuff and I think you'll find we have scraped a little reality together.
      Mitch Ratcliffe
      • I am still skeptical

        Maybe the article was misleading or I read it wrong, but the following text from it is where I derived my statements from:

        [i]"Delivered as an on demand software service, BuzzLogic looks at content relevance, using full text analysis; the number of messages from a certain publisher over time; traffic, including users referred by influencers and total number of pages views; and attention to messages via link analysis, including temporal data and total number of inbound links."[/i]

        It certainly does look like who is actually reading the blogs (which is impossible to tell) is less important than who is linking to it. Again, I beleive that whether or not something is linked to or not, or who links to it has less relevance than who is actually reading it. The real question is, regarding the algoritm, is "who is considered an 'influencer'?" I can think of tons of people who are quite "influential" but certainly not by the "blogosphere"'s standards, and vice versa.

        After reviewing the BuzzLogic Web site (I understand of course, that the nitty-gritty details of your proprietary business logic must remain under wraps, of course!) that the system does something like figure out what topics a particular author or site is considered "influential" in, and use that to weight how important their "gestures" are when applied to the customer. For example, if Blogger A gets an average of 50 comments, 200 inbound links, and 100 TrackBacks on a post about "astrophysics", any links/comments/TrackBacks they give to the customer's site will be considered very good if the site is about "astrophysics" but nearly worthless if it is about "foreign policy."

        Again, I posit that while this information is still quite useful and often indicates a strong correlation to "influence", it still does not have any meaning.

        My experience has been that the better written the item, the less feedback it receives, positive or negative. Why? Because people tend to link to or comment upon things they agree with less often than things they do not agree with, unless the topic is inflamatory. Someone could be the world's leading expert on car seats for children, and write huge amounts on the topic, and never once receive a comment and only get a handful of links, because there is nothing controversial about car seats. If one of his readers is a major CEO who decides to give away a free car seat that he recommends with every bank account opened or minivan sold, he is incredibly influential, despite the complete and utter lack of measurable data.

        Likewise, a racist who is incredibly outspoken on their blog will get a ton of comments and links (both good and bad), and many of them from real experts (like, "look at this bone head!") and your system will mark him as being "influential" even though there are only 3 people on the planet who actually care to read what he says or who will actually take him seriously.

        [i]The Economist[/i] has a readership well below that of [i]Maxim[/i], but I would consider the fact that the readership of the former holds much more money and wealth than the readership of the latter to be a better indicator of "influence". There are tons of examples out there, but "who pays attention to you" is much more important than "who talks about you." The former is impossible to accurate guage, and the BuzzLogic system (and any system using data culled from Web sites) measures the latter.

        Of course, if you care to give me more insight into the logic rules used for BuzzLogic, I would be more than happy to reevaluate my skepticism and take back my statements if needed. But the information posted in the article and on the BuzzLogic Web site lead me to beleive that the system is using a flawed metric.

        Justin James
        • How would you measure influence

          Hi Justin,

          I am enjoying your critique. I was wondering if you had thought about how a company be able to measure the influence of a blogger, other than links?
          • Gladly

            John -

            I am happy that you like what you're reading. The best way to measure influence is to produce a matrix of who the readers are and the actions they take pre- and post- reading. For example, if Reader A agrees with you, reads you, and does not change their actions at all, you had zero influence. Indeed, it is hard to influence people who already agree with you; at best, you reinforce their beleifs, as measured by them quoting you or using your particular version of the logic. If Reader B diasgrees, reads you, then agrees, you are extremely influential.

            In other words, influence of *anyone* (regardless of "blogger" or not) is impossible to measure by strictly Internet activities!

            [b]Let's get real folks.[/b]

            Nothing terribly important strictly occured strictly online. The Internet is a great tool to create leverage and reduce friction, but if nothing happens, nothing has happened. Who gives a hoot how "influential" a blogger who reviews stereos is, if none of his readers have the money or desire to buy stereos? He convince the entire East Coast that Brand X stereos are the best, but if no one ever buys a Brand X stereo, it does not really matter, does it?

            That is the crucial flaw in the "Web 2.0" and "blogosphere" group's self-congratulatory thinking: ideas without actions is [i]irrelevant[/i] just as faith without works is dead.

            It does not matter how many people read a blogger, link to a blogger, whatever Internet-only-action to a blogger if no action or result is produced in reality. Blogging is the sound of the tree falling without the tree actually falling, so it really does not matter if anyone heard it.

            For example, I "blog" (I prefer to think of it as writing articles and indistry analysis) over at TechRepublic. If one person uses the code I post in a project and it saves them 10 hours of work, I am infinitely more influential than a blogger that someone read, said "hmm, that is interesting" and never paid mind to it again.

            Justin James
          • What about search marketing

            Hi Justin,

            What you say makes sense, in that it would be great to measure real actions on the part of people once they have read your material.

            But to me when someone links to a website, makes a comment or clicks through to a website. That's a measureable event. In fact what if you can track a user from a blog to your website, who then purchases something. Surely that must count for something.

            On the subject of links, if someone links to you, and enough people do those links will eventually influence your ranking in search engines for keyword terms. So what you do on your blog may influence what happens on Google, Yahoo, MSN and Ask. If you achieve higher rankings on those search engines, and that sends you more traffic over time. And the people who come to your site take some action, didn't blogging just have an effect?

            Again, I don't disagree that it would be great to measure the influence of a blogger in the real world. But I also think there are ways you can measure some effects from blogging.
  • clarification

    Dan -- to clarify, we search across more than 7000 main stream media sources. The number of social media and corporate sites indexed numbers in the tens of millions.
    • MSM content

      from where are you sourcing your MSM content?