Underperforming Platforms and Web Journal – Week of April 1st- 8th, 2012

Again, a lot of work preparing for San Francisco next week and a few other proposals – they seem to take a lot of time, if they are going to be any good, and decided that’s where my time needed to go (along with my students at Rutgers University School of the Arts and the upcoming Measurement Course I’m creating for University of California, Irvine (summer courses are not listed yet, BTW), that is set to go live this summer.

Problems with the Tools for Listening

There’s also this, looking at using social listening for understanding the mindset of the Voter and testing that out with looking at the conversations in the smallest state in the Union, Rhode Island, by hearing from an expert on them (and the people who attended).  It’s out of that discussion I’ll begin this post.

Most of the platforms I have worked with turn out not to be very good at finding out what people think … about anything. Some view this as a UX-Design problem, but I think it goes further past it.  The solution, based seems, on the face of it, to focus on specific implementations of tool sets – because the self serve solutions simply don’t cut it.

I maintained for a while that, with unstructured data, especially, you need to first define what your listening for (and what the query looks like) before actually listening – otherwise, you entering a noisy mall and trying to understand what people are saying – an almost fruitless and extremely frustrating task – but one that is very common.   I tried to see what I could do with Rhode Island, but I didn’t know what the issues really were (and what bills or initiatives in Congress are tied to them) so, before doing effective listening, I had to do my own research, just so I’d know enough to be able to listen, and what to listen for.

But when I tried to do Listening – this is what I ended up seeing – the platforms on the market today, esp the most popular ones, could very well lead to this kind of architecture (seen below).

 ….

It’s amazing just how “bad” this could become – awful, and it has to do with the way the platforms have been created and sold (the marketing part).  Perhaps the Boolean Query and the logic structure is also to blame, most platforms are simply not able to move past it - limiting how effective they can be, ultimately.

The interesting thing is an alternative that does almost the same thing would cost more 10 times less to build and to maintain on a monthly basis, and be much more effective, in my opinion (see below).


 

 …

I think the issue is streamlining the application layer – until emerging technologies mature sufficiently, it may be that a custom application that is optimized will be a better solution than a self serve platform – depending on what is being asked for (and how much an organization is capable of defining what it wants to know) – that’s all I’ll say on this subject for now – but if you want to hear more, and your in San Francisco next week, come hear me speak.

Web Journal

Analytics

  • Was speaking with some of the people behind and supporting SAP HANA recently, so the news SAP is going after SAS with Predictive Analytics wasn’t surprising to me.  I think the big deal is “in memory systems” that can do real time Analytics, and in real time, on large datasets.  Big data, as I pointed out a few weeks ago, is not a new construct, many large corporations have been doing it for a while, but not so much in real time, or with such large datasets.
  • I think this is very interesting  Social + News Monitoring, together for the first time care of NewsCred and DataSift.  I know that I tried to do this kind of thing manually a few years ago, using some of the platforms I alluded to earlier in this post, and concluded, it’s a waste of time – the platforms on the market, till now, can not do a good job at matching up buzz by story – they were not designed to.
  • More hype about the uses of Big Data in the future.
  • Bit.ly’s early warning systemNot sure about this one, but .. a new Reputation Monitoring service uses sentiment analysis technology from Lexalytics to pinpoint potential brand threats in Bit.ly linked content as well as identifying opportunities for brands to shape discussions around breaking topics.
  • I personally, haven’t been that interested in Pinterest, but try telling that to 36,000 people who signed for analytics around Pinterest the other day.
  • Readers know I’m a big fan of Gary Angel and the great work they do at Semphonic, and hope to touch base in SF next week.  Gary just released a post on a framework to do reporting that I think brings some degree of automation into the space that has been missing.  Gary’s no fan of reporting as it’s been done, even his own (I share the frustration with both the tools and the looseness of the reporting requirements).  I think what is being offered here is a way out of part of the problem.  Gary doesn’t say that what he has presented is the total solution – we both know it’s not – but it’s a good part of it and well worth doing (and investing money into doing it well).
  • It was only a matter of time before some of the older players in the SMM space were acquired or vanished, that’s just happened (acquired) with Cymfony.  As noted ..WPP Group, which owns about 24 percent of Visible and all of Cymfony, will own 49 percent of the merged company.    You know how I feel about Communications groups owning Analytics firms, but then again, I guess this kind of thing is inevitable.
  • I think this post on B2B Social Media Analytics was well thought out and I agree with it. I noted over 2 years ago there was little real evidence (or case studies) that Social Media was working well in a B2B environment – and the writer confirms what I suspected all along – much of the hype about B2B is really B2C stuff that should not be counted.   The author concludes that there is value in SMM and B2B, but it’s a long term play – nothing that will generate much immediately, or even in the medium term – chew on that one.     I almost chuckled after reading this….
    • “….. Increasing the sales funnel: I am not convinced that social media analytics is going to do much by way of generating new B2B sales leads. Unlike the consumer space where the customer relationship is shallower, the enterprise decision process and associated buying indicators are incomparable. How can I compare the thought process of a consumer buying a $500 iPad with an enterprise buying a $50,000 database system? As social media adoption matures, and it gets more tightly integrated into the business, this will likely change. However, its a bit of a chicken and egg situation at the moment.”
  • I’ll be attending the Sentiment Analysis Symposium in NYC next month – and this guy will be speaking at it.  Noting that what is being done here was something I discussed with Adaptive Semantics 3 years ago, or more – they now run the Huffington Post/AOL back end for news information processing.

Social Media

Interesting Stuff

That’s it for me today- a week worth of posting in 2 hours and 1600 words – try to share more between the the hard prep sessions I’m finding myself needing to do, these days.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>