Asi Sharabi’s Private Selections, “No Man’s Blog” recently posted his feelings about what is lacking in the field of social media analytics/monitoring apps. I have spent the last year undergoing a similar exercise but for different reasons. We’ve both come to the same conclusion. He calls it frustrating. I call it market opportunity. In the end, because he has taken the time to air his feelings, it provides validation for the effort of those “nice people” who work at these companies and first hand customer feedback.

I’d like to run down his list to summarize his valid points and talk to how our product, is addressing them. Here we go…

---

Asi Sharabi’s Private Selections, “No Man’s Blog” recently posted his feelings about what is lacking in the field of social media analytics/monitoring apps. I have spent the last year undergoing a similar exercise but for different reasons. We’ve both come to the same conclusion. He calls it frustrating. I call it market opportunity. In the end, because he has taken the time to air his feelings, it provides validation for the effort of those “nice people” who work at these companies and first hand customer feedback.

I’d like to run down his list to summarize his valid points and talk to how our product, is addressing them. Here we go…

1. Shortcomings of the Underlying Technology: Asi is right. Keyword technology is the farthest thing from accurate. That’s why we are using Natural Language Processing and other approaches that attempt to derive meaning and context from data. Some vendors are beginning to use computational linguistics based technologies which is better than plain ol’ keywords. We take it a step further and filter out all of the irrelevant conversations based on sophisticated algorithms and intelligence we derive from your existing online presence (web site, blog, social networks, etc.).
2. Unreliable Data. Again, collecting all the data is one thing but sorting through it and manually extracting intelligence from it is extremely time-consuming. Like some existing vendors, we filter out the spam and noise and also look for other metrics to determine relevance. With out tool, you’ll have the peace of mind that we do the dirty work and present you with what you should be addressing, good, bad or indifferent.
3. Sentiment Analysis is Flawed. Asi points out correctly that keyword based sentiment analysis is almost not worth paying any attention to. There are simply too many nuances in our language especially when it comes to emotions! Let’s face it. People interject sarcasm, irony and subtlety. We are addressing these problems with manual linguistic intervention so you don’t have to. We are also taking a different approach to sentiment so that you can glean much more information from it beyond positive, neutral and negative.
4. Region Specific Data: This is one of the most important areas for global brands and it also involves multiple languages. We pull information about geography beyond the domain based on data from website analytics, traffic analysis, etc. using foreign language modules.
5. Influence Analysis is Flawed. There has been a lot of discourse on this subject and the jury is still out for some. We believe that influence can be measured if it comes from multiple sources leaving room for error. I can’t elaborate on the depth of our approach for competitive reasons. I’ll just have the technology’s accuracy speak for itself.
6. Unreliable Data + Partial Data = Unreliable Stats and Visualization. Yup. Garbage in, garbage out. Some of these tools are pretty “flashy” (literally) when it comes to data visualization but if the underlying data is flawed, what’s it worth? We take a straightforward and simplistic approach so you can easily understand our visual representations but we put most of the effort in the back end for accuracy.
7. Time Consuming. Our goal is to do provide the analytics and analysis for you. Moreover, we offer it to you in language you can understand. If you want to drill down into the underlying data and free form search, you can, too.
8. Price. With the availability of open source, the price of developing technology has dropped and the speed to market has gone up. Offering a SaaS (software as a service) product, is the way to go so that you can offer lower pricing. Our pricing will reflect the usage needs and therefore be more accessible to those with smaller organizations and offer high customized functionality for power users at a higher cost. We offer a fair, affordable and cost effective solution.

I have to say that I agree with everything that Asi has to say, except for one thing. Google has made little attempt at semantic search. It’s very hard to do and Google’s purpose is to provide relevant site results and works on constantly improving algorithms due to the large volume of new content that proliferates in the web. I’m sure it’s on their radar screen but we’ll just have to wait and see. We’re not waiting…neither should you!

Thanks again to Asi for speaking his mind…

Tags: Social Media Measurement

This entry was posted on Wednesday, August 19th, 2009 at 12:00 AM by Wendy Troupe and is filed under Social Media ROI, Wendy Troupe's Perspectives. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.