Empowering the business: Watson Analytics and the future of the data scientist
A Q&A with Eric Sall, Vice President of Business Analytics, IBM
Published October 08, 2014
Big Blue’s Watson artificial intelligence system rose to fame after winning the "Jeopardy!" TV game show in 2011. Today, IBM aims to use Watson to make data analytics accessible to everyone throughout the enterprise.
According to IBM, the first release of Watson Analytics—a cloud-based analytics service based on the Watson cognitive computing platform—includes a “freemium” version of its cloud-based service designed to run on desktop and mobile devices. Watson Analytics offers a full range of self-service analytics, including access to easy-to-use data refinement and data warehousing services that make it simpler for business users to acquire, prepare, analyze, and visualize data.
Insights Magazine sat down with Eric Sall, Vice President of Business Analytics at IBM, to learn how businesses can harness Watson Analytics and a predictive approach to make better decisions quickly and more easily. We drill down into the security behind the platform and the four key benefits it offers customers.
Insights Magazine: What are the main pain points for businesses today in terms of analytics?
Eric Sall: Analytics has been around for quite a while, and businesspeople really do understand now how valuable it is to apply analytics to all types of business problems, because when you think about it, a businessperson is making decisions all day long. And any of those decisions might be made better by insight into the business, by accessing the relevant data and using it to make a more informed decision. But for most businesspeople, it’s just way too hard today to use analytics for your everyday decisions. There are a number of problems—one is the issue of getting the relevant data that you need to make a better decision, whether you’re a marketing person or a salesperson or an HR person. Another problem is that the analytical tools typically are aimed at business-analyst types or data scientists, and they assume a lot of knowledge around how to do analysis, what's the best technique to apply this certain thing, and building models and such to do the analysis you’re looking for. Then there’s also issues having to do with the time it takes to do all this, as well. When you’re a businessperson rushing to a meeting or trying to make a decision you have to make in order to move the process forward, you have very limited time.
IM: How can Watson Analytics change the way businesses operate and grow?
Sall: The purpose of Watson Analytics is to reinvent that whole analytical experience and to help a businessperson, not a specialist, but to help a businessperson inform their decisions with better information and better insight. With Watson Analytics we’ve helped you acquire the data, helped you cleanse the data—that’s an important problem because most data has issues in it. For example, in the ‘State’ field, if you’ve got a customer file in the State field, it might be ‘NY’ for one record and ‘New York’ in another record, and that really means the same thing, right? And you’ve got to do that kind of cleaning-up of data in order to get meaningful analysis to it. So, we’ve made that easier.
Get insights delivered to your inbox every week. Subscribe to our free newsletter.
Then, we also automatically take your data and surface what matters most. For example, let’s take an online retailer. What they might want to know is why certain types of product lines are selling in one region and not in other. What’s driving those sales? Is it the demographics of the user? Is it the product assortment? Is it the channel that they came in by, etc.? Those are the kinds of things that a marketing person needs to understand to better understand how they can target some of their efforts.
IM: What do these kinds of analytics tools, that put the power in the hands of the business, mean to the future of data scientists?
Sall: Of course the data scientists and business analysts add a ton of value, and they will continue to add a ton of value because they are important for certain types of analysis. But the fact is, there’s a big shortage of data scientists. In fact, a lot of the universities have started programs to train either Bachelor’s to Master’s degrees in data science to fill that gap. So, one issue is the scarcity of those people to begin with. The other issue has to do with just the notion that analytics can be informing every decision you make, not just the special ones that you’re trying to do when you’re doing, say, annual planning or something like that. You really want to be able to make those better decisions. If you’re a marketing person planning a campaign, you’re making a lot of decisions about the messaging you use, the lists that you use, the format, the placement on a website—all these things are decision that you’re making and they have an impact on the results. If you can make those decisions in a better way, you can improve the business results on an ongoing basis.
[Another] example of using something like Watson Analytics is to apply predictive analytics against your pipeline and try to figure out which deals you’re going to win and which you’re going to lose based on an understanding of what the factors are that drive that. Is the tenure of the sales rep driving a win or a loss? Are you winning more in one product line but losing in another product line? Are you having trouble in one region of the country? Are you winning when you discount more heavily? These are all factors that go into why you might win or lose a deal. [Once you have this information], then you can … drill into the situation of that deal and try to take some corrective action earlier so that you can increase the chances that you will win that deal.
IM: What are some ways Watson Analytics will change data processing and analytics for companies?
Sall: One thing is the idea of making [the use of analytics] more of an essential everyday activity for more types of businesspeople to make better decisions. Another thing is to increase the familiarity with analytical tools. Again, a lot of times people today, when they want to do analysis, they go to a specialist and they get their help. The dynamic there is, you go to that person, they help you, and they give you your answer. But as a businessperson, you may not really understand what magic happened there and how to think about that answer. By putting it in the hands of regular businesspeople, it can spur the right kind of thinking. It’s not just simply getting an answer, but it’s also about exploring an issue. Seeing other related things and using it to get a deeper understanding of not only the decision you’re trying to make at that moment, but how to drive success in your business and have a longer-term understanding of the various driving factors.
We asked Eric Sall what value Watson Analytics can bring to organizations. He offered his top four key benefits:
- Self-service: “As a business user, you can be completely self-enabled to conceive a need, to get the data, to analyze it, and to communicate the results all by yourself.”
- Unified experience and usability: “You don’t need to be an expert; the product takes you by the hand and helps you on your analytical journey.”
- Bring to bear different analytical tools: “Whether you want to do some reporting on data, some predictive analysis, visualization, or interact with the data in a cognitive way by asking questions on the data and just focusing on your business problem, Watson Analytics has all those abilities. It brings together what would typically, in other analytical offerings, be all separate products that you’d have to learn and use differently and most likely you wouldn’t have access to all of those things. With Watson Analytics, it’s more of a unified capability.”
- Stories: “The stories are like templates to start you off on your analytics process. It could be things like campaign management for a marketing person; win-loss analysis for sales; employer retention for HR. These are pre-created examples that you can use just to learn what you might do, or use as a template for your own analysis if your business problem is similar.”
X Close
Another dynamic going on here is this notion of self-service in general. I think a lot of us have changed the way we interact with technology, largely because of what's happened in our personal lives. Things like your smartphone or your iPad have trained us all to expect an immediate answer to everything. I’m sure you’ve been in the situation where you’re on the couch and you’re watching a movie, and you’re, like,’ who is that actor?’ And you pull out your phone and look up IMDB and find out the name. Then people come to work and that situation is totally different? At work, you’re much more limited in your ability to get at answers because you’re dependent on other people or you don’t have access to information. So, what’s going on is people are demanding that same level of self-control and self-service at work that they’re getting used to at home. That’s really the need that we’re feeding into with Watson Analytics.
IM: When people talk about cloud, there’s always a question of security. Can you speak to how IBM addresses those security concerns?
Sall: Watson Analytics is multi-tenant, which means that that data that you upload or connect to in your use of Watson Analytics is not visible to anyone else. It’s in your own instance of the software, which might be separate to you as an individual or it might be a group thing depending on how you want to configure it. But the notion is that the data that you upload is only available to you, it's not something that other people could access. The other thing that's important to understand is that with data sources, especially enterprise data sources, sometimes an individual user, even within a company, is authorized only for part of that data. For example, they can see customers’ names, but not their credit card numbers. Watson Analytics will obey any of those kind of access control rules for the data source you are using, so you’re not going to have any rights that you wouldn’t have using the application directly.
IM: This release includes what you are calling a ‘freemium’ version. Can you elaborate on this and what it means for clients?
Sall: We felt like a big part of Watson Analytics was removing a lot of the barriers—usability; getting and cleansing data; learning different analytical techniques; and the cost barrier. This is not just a temporary strategy; what we intend to do with Watson Analytics is have this freemium offering that’s really quite capable. It allows you to access the data, to analyze it, to visualize, and to produce results. For many people, some of their problems will be solved completely by that freemium version, but we will have offerings that are additions for more capability, and the kinds of things that you’ll get in the bigger additions would be larger work space area, things like live connections to other cloud-based sources, and maybe some additional functionality as well. So, it's not a trial version that we're offering, it's really a very functional thing. But then there will be the ability to trade up to get, as I was saying, more work space, more data source connections, more features, and also support.
Find out more about IBM Watson Analytics.