Best practices for evaluating and selecting content analytics tools Sponsored by IBM Speaker: Seth Grimes, a Business Intelligence and Decision Systems expert Moderated by Jonathan Gourlay, Site and News Editor for SearchContentManagement.com Jonathan Gourlay: Hello I am Jonathan Gourlay, Site and News Editor for SearchContentManagement.com at TechTarget and I would like to welcome you to today s expert podcast on Best Practices for Evaluating and Selecting Content Analytics tools. I am joined today by Seth Grimes, a Business Intelligence and Decision Systems Expert and founding chair of the Text Analytics Summit. He is also the founder and principal consultant at Alta Plana Corporation, which consults on business analytics strategies. He will tell us about how companies might go about evaluating and choosing content analytics tools to unlock the value of unstructured data to best meet business needs, and we will find out some pitfalls to watch out for along the way. Hi Seth, and welcome to TechTarget and SearchContentManagement.com s podcast booth! Seth Grimes: Well, thanks Jonathan for inviting me to participate. Jonathan Gourlay: It s great to have you here today. Well let s get started. Once the company realizes there is a lot of potential informational value locked in its unstructured data, what s the first step to take in setting up a content analytics strategy? Seth Grimes: There definitely is loads of business value locked in so-called unstructured sources. That might be social platforms or enterprise sources, such as e- mail or online sources like the web. That s information about customers, about competitors, about the market. The first step, though, really is to try to align your text analytics, your content analytics, with the business problems at hand. What are the business problems you are trying to solve? Are you, for instance, a telecom trying to reduce customer churn? Are you a hospitality industry, hotel, restaurant trying to boost customer satisfaction? Are you doing manufacturing or you want to increase product quality, and thereby decrease the number of returns, and so on, and boost your sales while you re at it? First step in trying to get value out of so-called unstructured content is to
really figure out what business problem you want to solve and then figure out the most direct way to get at the information that you need. What are the sources that are most important to you? Would it be surveys? Would it be e-mail? Would it be contact center notes? Would it Twitter or Facebook pages? You have to figure out the sources that are the highest value for you in providing the unstructured information that you need to solve your business problems. Jonathan Gourlay: That s interesting. I was wondering if you could tell us whether there are strict differences in the available solutions to be considered, whether a company wants to use mined data for marketing purposes, say, rather than for something like legal e-discovery purposes. Seth Grimes: We really have a market now that has solutions that are tailored to a broad set of business needs, that is solutions that are tailored to media and publishing needs, ones that are not in an industry vertical like that but rather handle so-called horizontal needs, business functions like customer relationship management, customer experience management, customer service and support. We have capabilities from certain vendors that target fraud, risk, compliance types of areas. There are other capabilities that are great for life sciences. Fortunately, there is a broad set of choices out there. You can find solutions that are tailored to particular business sectors or to particular business needs, or you can go out and license tool kits from which you can build your own solutions, build your own customization for the business problems that you have at hand. We are very fortunate in having all of the huge variety of solutions out there. Jonathan Gourlay: Why is analytics more important than a solution that provides say strong search capabilities? Seth Grimes: That is a key question, and I will answer first by saying that s its not either or, one can have, using content analytics solution, both the analytics that finds insight, discovers knowledge in source information and research capabilities that can be used to access that knowledge that s been unlocked. The problem with search, as it s been traditionally delivered by search vendors, whether that s on the web or in the enterprise search domain, the problem is that it relies on keywords. And, it relies on the users having some preconception of what he or she is looking for. That is, you got to know how to ask for the information that you are looking for. Search is not a good technology for knowledge discovery, for finding connections. That s what knowledge is really all about. It s about connections, concept, topics, themes. And, you really need to go beyond search in order to get at all of that really rich information to make connections, to essentially mine all of the data that s out there on the web and in the enterprise. So it s really not an either or. Text analytics, content analytics works hand in hand with search to provide a rich set of knowledge to recommend the access with search solutions. Jonathan Gourlay: Is text analytics the same thing as content analytics? Seth Grimes: Yeah. Maybe you caught me jumping back and forth between text analytics and content analytics. I actually see text analytics as a subset of content
analytics. That s because content is not just textual in form. It also includes images and audio, and then video. So those are very rich sources of business information and, in fact, I believe that You Tube video sharing site is the fastest growing social platform out there, so all of those advanced forms of unstructured content are very important. But text is among the easiest to mine and analyze, and text is also extremely rich, the stuff that we put out on social platform. Yeah sure, we post videos and all of that kind of stuff, but we post even more text. So, text analytics is a subset of content analytics. We will see a shift over the next few years to an increasing proportion of attention being paid to those other forms of content, to audio images and video. But yes, the techniques are very similar. The business purposes are very similar, to find out insights about the customers, about the markets, about the competitors. If you are in electoral politics about the voters and their views on different issues, if you are in clinical medical care, find out the voice of the patient, text is the key source for that right now. Jonathan Gourlay: How far have vendors come in adding automated textual content analytics in business applications? Seth Grimes: They have come pretty far actually. So, I mentioned earlier that there are solutions that are tailored to particular business problems and, if we are a certain industry, then there are solutions that are tailored towards need such as human resources or customer service and support, or with CRUD and so on. So, we see solutions that are freestanding to handle particular business problems and then some of the largest enterprise application software vendors and services vendors are embedding text analytics, content analytics capabilities into their software. That means that the capabilities are directly implemented within the line of business applications, the ones that users are using already. So, they don t have to go out into separate work benches. They don t have to learn how to do analytics. It just comes naturally because it s been designed into these enterprise solutions. So, to go back to my response to an earlier question, we are very fortunate in having a lot of choices. There are standalone work benches for particular business problems and then there is the integration of text and content analytics within enterprise applications, software and services that are out there, loads of choices. Jonathan Gourlay: And, good to know. I think then what I would want to know is what I should I look for in a solution if my company has marketing, customer targeting, legal and other needs in the entire cross-section of departments in the enterprise? And, should I go for solution that promises a lot of cross-industry options within the application or one that promises ease of use? Seth Grimes: Well it s, once again, not necessarily an either/or. You are going to see greater ease of use often if you do have that kind of alignment with the particular business functions that a marketer or customer services support person or a contact center person is going to be doing. It s going to be easier to use because it s going to conform to the work practices and workflows of the applications that the person is already using. It is tempting to go out and to implement an enterprise-wide solution for just about any kind of business problems, but that s not realistic in a lot of cases. In a lot of cases, well,
you are working in marketing. You have a job to do. You want to get your job done. You want to create return on investment. You want to create return very quickly. You don t want to have to be burdened by an evaluation of tools that s going to take a year and then a heavy duty implementation. You just want to get to the answers that you need in order to better solve your business problems, do better customer support, to solve quality issues, all of that kind of stuff. In that case, the solution that s going to be best for you is one that really is departmental level. So, that s actually the recommendation I have. Try to solve some very concrete business problems, so-called low hanging fruit. Get at some good payback from your initial investment. Make it a modest investment. And, then go from there to create an enterprise-wide solution. Jonathan Gourlay: So, to follow it up just a little bit, if I am the guy who is in charge of doing research and I am convinced the company needs to take a close look at its analytics capabilities, with an eye toward picking a powerful content analytics tool, how should I get started? Do I just pick one and fly with it or is there something more sophisticated I need to take into account? Seth Grimes: Yep. Well, before answering that, let me make it clear that it s not just the people in research who are using these tools. You have people in the Line of Business department, in Marketing, in Customer Service and Support who are going out and using these tools and they can do it pretty quickly and easily, and at relatively modest cost because a lot of the solutions that are out there, including solutions that are tailored to their business needs, are available as a service. So, as a service on the cloud, if you prefer to call it that, means that you can use an application without making a large upfront investment in software. You don t have to buy the hardware to run the software. You don t have to involve your IT people who then install and configure the hardware, and then ask you all kinds of questions that essentially get in the way and take a large proportion of your budget, much more than you are prepared to pay. You business analysts, as well as researchers, can go directly out and get these solutions by using them as a service. I mentioned that there are also open source options that are often better suited to people who have a more technical background. And then, not every commercial option that is installed is going to be expensive. There is really a very wide range of options available. So, the best way to get started with all this, I think is to just do it. Find some information that can be analyzed in such a way that it will respond to particular business problems that you can identify, that will produce measurable results that you can then use to justify the investment that you have made and justify an extension in that investment in order to expand the use of text and content analytics by your department, and then ultimately in your whole organization. Jonathan Gourlay: That sounds great. It doesn t look like there is whole lot of barriers to getting started with a solution.
Seth Grimes: No. I would say there are not a whole lot of barriers. There are solutions that are very affordable and very powerful at the same time, and then a third leg is that triangle, very usable by end users who are in Line of Business departments. Jonathan Gourlay: That was Seth Grimes of Alta Plana Corporation. Seth, thank you very much for spending the time today! Seth Grimes: My pleasure, Jonathan! Jonathan Gourlay: For SearchContentManagement.com, I am Jonathan Gourlay.