An Interview with Catherine Havasi, Chief Executive Officer, Luminoso
Could you provide a brief overview of the solution that Luminoso provides?
Catherine: Absolutely. Luminoso is a text analytics company. Basically, what we do is help take text and turn it into actionable insight. It can be text you get from forums, text you get from website comments, social text, etc. It can also be text from more traditional sources such as surveys or market research. Luminoso uses extracted Web data like product reviews and online conversations in social media to discover actionable knowledge.
I think that’s really important because Luminoso’s text analytics solution is effortless. You don’t need to train it. You don’t need to build ontology or taxonomy, and you don’t need to build a keyword list. This is facilitated by an advanced machine learning algorithm that Luminoso has developed.
It will automatically pull information about a particular domain. So, if you came in with a particular area – be it automobiles all the way down to something like cat food – you wouldn’t need to tune a system to that particular area. It would do it automatically from the text that you have.
It’s essential for taking something that’s just web data and moving to knowledge very quickly without having to put a person in the loop and without having to change the tuning of the system between each time you have new text you want to study. It’s also very important because it can deal with misspellings. It can deal with jargon and slang which is very prevalent on text that you scrape from the Internet.
I think it really provides people with new power to be able to tackle and get actionable insight out of their data.
Luminoso’s dashboard provides unprecedented levels of fidelity in understanding what customers are saying.
What kind of situations do your clients typically come to you with?
Catherine: The situations are primarily focused on market research and product development. Our customers often want to make sense and draw insight out of customers talking about a product, a competitor’s product or a market. Customers also look for information from general forums and could be everything from “What are the trends right now in the way people are looking at hair dye?” to ”If we made a spicy chocolate bar, where would people be interested in buying it?”
Ultimately, the objective can be anything from being able to develop products to being able to understand the market a little bit better.
Our clients come to us with varying degrees of data to interpret:
- Some have the clear cut data and they want to extract actionable insight out of it
- Some have data they want to explore, like survey open-ends or customer feedback
- Some need to collect the information they want to analyze from forums or reviews
The last scenario is where a solution like Connotate is an essential way to extract information.
How does your solution help your clients make better decisions?
Catherine: It takes what would normally have been a very qualitative analysis and makes it data driven. It allows people to be able to look at more text and more opinions when making a decision than they ever would have before. You can also pull in all sorts of interesting and fuzzy ways that we talk about things like flavor and fragrance when we’re passionate about a product.
We can talk about any aspect of a product in a wide variety of ways. If we’re passionate about a fragrance, there are many different ways we could describe that fragrance. You and I would know when someone was saying the same thing about how a product smelled. It would be very difficult for a computer that was not using Luminoso to be able to do that.
If you’re going through blogs and people are talking about the product smelling musty, you want to understand when other consumers have called in with similar complaints. You don’t want to have to train the system ahead of time. You’ll want something that just adapts, and you want to be able to take things that are very qualitative and actually change them into something that you can compare and numbers that you can have. That’s what we specialize in.
Can you tell me about the coolest client example that you can think of and tell me how you helped the client to make better business decisions?
Catherine: Sure. When people think about getting information from the web for product development, marketing, or market research they think about going to Twitter or Facebook to get information, but the honest answer is there’s really great information all over the web. There are blogs that people are looking for and people are writing that contain your specific customer type and are very focused to what you do and what your product is – be it technology, pharma, services or consumer packaged goods. Being able to actually grab that data and look at it often gets you really great insight.
My favorite story of looking for data in an unexpected place and getting a great result is always YouTube comments. YouTube really has the reputation of the epitome of the bottom of the web in some ways, but it turns out that there’s a lot of data there.
We had a customer who was working with a consumer packaged goods product and was working to refine that product, put out a new version, and they were interested in what types of consumers the product worked well for and what types of consumers the product broke for and, if it broke, how did it break.
They went online and scraped video reviews of their product, and they took the comments that people had there, and people were like, “Yeah, I understand you like it, but it didn’t work for me or this particular part of the product broke in this particular way.” They were able to understand more than they could have gotten from their limited ability. The customer had a limited ability to bring people in to conduct product tests, and it turns out, by pulling online comments, they found out how the product was breaking and not working for different people. They were able to fix that for the next version out of YouTube comments. Data is everywhere.
Forum data is an amazing way for people to get information off the web.
How does your partnership with Connotate allow you to provide better solutions to your customers?
Catherine: I think the key to doing a good analysis is having good data. People have come to us with questions and they don’t necessarily have the data they need to answer those questions – they can often find it online.
It’s very important to them that a solution exists that would allow them to just get the data they need off the internet — and get data off parts of the web that are not necessarily your traditional social media locations.
Finally, what advice would you give someone who was ready to take on a project that might be able to leverage the Connotate and Luminoso partnership?
Catherine: The quality of your data is what’s most important and when you go out there and you’re going to do a data scrape or a data pull, really think about the sources that your data comes from and think about what you’re looking for.
If you go out there and you do a very broad sort of search on something like social, you want to make sure you’re actually bringing down stuff that’s related to your product as opposed to things that maybe have the same keywords but aren’t related. You can remove all that spam, but in that case, you have a lot of data that wasn’t useful. You don’t find that problem to the same degree in reviews or specialized sources.
So think about where your data is coming from, explore sources beyond Twitter and Facebook and really get in there – the data you’re looking for is out there somewhere.
The post Connotate Partner Spotlight: Luminoso appeared first on Connotate.