Sentiment Extraction Algorithm vs. Social Media Analytics Vendors

I've been toying around with various social media analytics vendors over the past years and been finding a lot of these tools often not too useful. Even the expensive social media analytics tools and reputation management analyzers are not really as good as I expect them to be. After heaving a brief discussion with Marshall Sponder, who is one of the most well-known social media analysts, I have to agree with him that a. organizations (too often) don't care about numbers & b. in depth analysis takes enormous resources,which is something startups or medium sized companies cannot provide.

Regarding a. it's unfortunate, but a lot of organizations are built around positive news. Nobody wants to hear the negative or even neutral findings, which often come out of a social media analytics report. Obviously this has to do with responsibilities, hierarchies and other company politics, which is probably material for a PHD thesis and not for a post on a mediocre blog (such as this one). At the end inconvenient news and numbers are too often thrown away. Be it the incapability of re-acting or just the complexity of the results.

The other problem I've noticed with social media analytics is the "deep web" and the limited resources of social media analytics vendors. One of the largest issues here is the variety of web pages and the different technologies of the search engines. Regarding the deep web....it is incredible complex to crawl as well as parse websites (unless there is RSS and Co.).

That all being said... what's more & more important are sentiment extraction algorithms of the big search engines. Search engines analyze reviews of products/services themselves, weigh these results and then display the results (hopefully) accordingly in the SERPs. This means that social media analytics system have barely a chance to identify the reach of a review, which has been posted.

I am sure all the above said will be something, that companies such as Reputationobserver and other startups are working on. We'll probably see over the next few years if we come closer of understanding the always evolving social & "old school" world wide web.