The myth of automated blog monitoring

Posted on by

Share the post "The myth of automated blog monitoring"

We were recently approached by a global enterprise tech company that’s interested in tracking discussions of its brands in blogs and social media specifically in the EMEA region. This is a surprisingly common request. Of course I always give them the spiel about the global nature of the blogosphere and the futility of trying to divide it up along conventional geographic boundaries – but this kind of thinking isn’t really compatible with the practicalities of working with a multi-national company. If Joe is head of EMEA PR, then Joe is only interested in EMEA blogs, and it’s my job to figure out a way of delivering or they take the business elsewhere.

The difficulty is that there’s no kind of identifying tag which lets you easily and automatically ascertain the geography of a blog – you can’t simply perform a search for blogs only from the UK, or Asia or Western Europe. The only way to figure it out is to manually examine each blog and hope the author has explicitly stated a location in their profile, or at least left some sort of clue in one of their posts.

It’s tricky and time consuming, and it really needs to be done manually. So many companies in this space are trying to develop fully automated monitoring systems, but I think when you look at the requirement to segregate blogs geographically (combined with the technical problem of separating genuine blog coverage from spam-blogs and false positives) the only way to guarantee high quality results is to underpin whatever automatic blog monitoring tools you may be using with a strong element of human intelligence.

Some people in the industry don’t like the sound of this, because it essentially means that you need to have a member of staff taking the time to read each piece of blog coverage to decide whether it’s genuinely relevant to your client. This, of course, costs more than a software solution which could do the job automatically at the push of a button – but the truth is that no such system exists, and anybody who tries to tell you they’ve solved the problem is most likely trying to sell you snake oil.

  • http://www.attentio.com/blog Simon McDermott

    I don’t really agree with this, there are techniques to start this process. You certainly need quality checks and there is a “seeding” component to start the crawling. We have successfully brought together 2million + European blogs and quality checks show good results. Simon

  • http://www.spectrum-analysis.com Mark Westaby

    Be very careful how you read this piece. It’s absolutely true that the web knows no geographical boundaries and it is also correct to point out that there’s no such thing as a ‘European-only’ blog. BUT, it is entirely possible that ‘Joe’ is responsible for blogs emanating from EMEA and that he therefore has to do what he can to determine the source, which isn’t unreasonable. It is also possible with today’s technology to be 90% to 95% accurate regarding the geography of the source, which is pretty good.

    Perhaps more importantly is that automated analysis of blogs is now extremely good and much, much better than human analysis. A combination of automated analysis with good practice and common sense would result in a perfectly robust and cost-effective outcome for a client.