This is Piplzchoice's Typepad Profile.
Join Typepad and start following Piplzchoice's activity
Join Now!
Already a member? Sign In
Piplzchoice
Recent Activity
Interesting undertaking Ray. I wonder how you would classify social media context analysis? There are text analytics companies that use social media content to produce research and intelligence of different types.
Heather, This is an excellent summary, however keywords approach to meaning of WoM and transactional measurements, while relatively easy to implement, are only good enough to scrape the surface. If you go beyond these methods, you discover that WoM is an incredible source of Market and Customer intelligence that challenges institutional orthodoxies and clearly points to specific actions.
1 reply
Good post Allen. The challenge of making Collective Intelligence truly useful to SCRM is in interpreting "raw" data, that we use to express our opinions, into structured information an enterprise use to digest knowledge. There are a plethora of technological and cultural issues to be contended with, as an enterprise culture is extremely self-centered in it's processes. There are quite a few companies around that try to address the technology challenges, and my company (http://www.amplifiedanalytics.com) is one of them. However the cultural challenges are much harder to deal with and I have not seen a lot of new methodologies and processes that would change the way an enterprise "consumes" Collective Intelligence. I you do have a different experience, I surely would love to learn.
Rob, I am not sure the connection I've made is the one you had in mind, but I want you to look at what we are doing at Amplified Analytics to extract product reputation metrics from customer reviews. I think you would find it interesting. At this point we are focusing on CE and computer products reputation scores.
1 reply
"At best, about one in ten customers will leave a product review on a website." - I think you greatly overestimate this. I am focusing on CE products reputation analysis and there are average about 35 customer per CE product. Using your assumption a product only sells 350 units on average, that is economically unsustainable. Nevertheless your point is a great one and I absolutely agree with your line of reasoning. The number of reviews is growing exponentially as people find them useful in purchase process and advertisers utilize them for acceleration visit to purchase conversion. I am aware of CE manufacturers who attempt to use the reviews for a purpose you suggest, but find it too expensive and difficult to process manually. There are many software programs available for text analysis, but the cost, complexity of implementation and lack of methodology make investment in them a somewhat risky proposition. We are developing algorithms to rate products reputation for Functionality, Reliability and Support based on customer reviews, and currently have a database of over 13 products being monitored. As we convert unstructured text into metrics, the analysis of the resulting information supports the process you suggest quite easily.
Piplzchoice is now following The Typepad Team
Jan 28, 2010