Social Media Monitoring with ScoutLabs – Interview

During my previous posts i have shown some examples of Sentiment Analysis using Twitter. I came across a Sentiment Analysis product named ScoutLabs which is able to give insights on what customers are saying on the Web about a product or service and decided to get an interview from ScoutLabs CEO Jennifer Zeszut :

– Please tell us about ScoutLabs and how companies may benefit from using it.

Scout Labs is a powerful, web-based application that finds signals in the noise of social media to help teams build better products and stronger customer relationships.

Scout Labs is a product company, not an agency. We provide cutting-edge technology and a collaborative platform for companies and their agents to listen to customers and engage with them out across the Internet. With Scout Labs, our users:

* Know when to tune in and what’s most important to pay attention to
* Hear what customers love and hate about brands
* Reach out to influential customers to build relationships
* Engage in proactive customer service
* Let the voice of the customer inspire new product and marketing ideas

Scout Labs has grown significantly since it was founded in 2006. With offices in San Francisco an…

During my previous posts i have shown some examples of Sentiment Analysis using Twitter. I came across a Sentiment Analysis product named ScoutLabs which is able to give insights on what customers are saying on the Web about a product or service and decided to get an interview from ScoutLabs CEO Jennifer Zeszut :

– Please tell us about ScoutLabs and how companies may benefit from using it.

Scout Labs is a powerful, web-based application that finds signals in the noise of social media to help teams build better products and stronger customer relationships.

Scout Labs is a product company, not an agency. We provide cutting-edge technology and a collaborative platform for companies and their agents to listen to customers and engage with them out across the Internet. With Scout Labs, our users:

* Know when to tune in and what’s most important to pay attention to
* Hear what customers love and hate about brands
* Reach out to influential customers to build relationships
* Engage in proactive customer service
* Let the voice of the customer inspire new product and marketing ideas

Scout Labs has grown significantly since it was founded in 2006. With offices in San Francisco and users all over the world, the company currently employs over 20 professionals. Our CEO and product team guide the application with insight from the world of marketing, brand management and product management, but the majority of Scout Labs employees are senior engineers with expertise in search technology, high-performance systems, natural language processing, machine learning, web crawling and data visualization.

– How is ScoutLabs different from other sentiment analysis solutions?

Many sentiment analysis solutions are really human-powered, which is great, if you’ve got a big budget and or a lot of time. Ours is automated, and we process millions of posts per day and score it for sentiment as it happens, with an accuracy rate (agrees with humans) 73% of the time and it will get better, because as users (across our system) change sentiment values in our system, we aggregate that data and use it as labeled data to improve our algorithms even further. We also can “backfill” or back-score 3 months of previous data with its sentiment scores in 20 minutes to (at most) 24 hours.

– What kind of information (product names, company names, areas, city names) can ScoutLabs identify in user conversations “Out of the Box” ?

Scout labs lets you track anything you like. It is purely search based (not a database of x fixed company names). We see searches on everything! Company names, product names, people, industries, “the price of rubber”, green energy, styrofoam, “tricked out shoe”, “canceling cable”, “favorite hotel in bali”– you name it. For some of these, sentiment doesn’t make a ton of sense, but we’ll score it for you nonetheless (just in case).

– One interesting functionality is to be able to identify associated keywords by products. For example consider two different brands of running shoes such as Nike and Adidas. “Great design” might be associated with the first brand while the other brand is perceived as being “comfortable”. Is ScoutLabs able to identify automatically such information between two or more similar brands-products?

We do offer an analysis of the top conversations associated with any search. This comes stright from the conversations themselves. They are the frequent words that are emerging from the conversations for the time period. And yes! Very often we find powerful adjectives that describe a brand (sometimes good, but not always). For example, when i did a search for dora the explorer (the popular cartoon character for the pre-school set), the words that emerged was “skank” “sexy”. Huh? Sure enough, parents are outraged by a mattel announcement about how they are going to have dora grow up, move to new york and get fashionable (with a short skirt). This is a true story: http://www.Scoutlabs.Com/2009/03/12/scandalous-doll-drama-scout-labs-style/

– How much should one expect to pay for such a service?