This week in New York City, IBM unveiled its much-anticipated Watson Analytics platform. Importantly, Tuesday’s event provided a cohesive message around this technology — something many of us (industry analysts, press, and end users alike) had been previously piecing together from the separate bits that were available. The IBM team generated a lot of buzz throughout the day, with a collection of demos, walkthroughs, and various Q&A sessions.
These efforts, and the technology itself, seem poised to launch an interesting new commentary on the future of the analytics market both for IBM specifically and for the industry in general.
But, first things first. Let’s look at what was actually announced. 1
The first big announcement concerned the way users can interact with Watson. Fundamental to the user experience is Watson Analytics’s ability to understand natural language. Users can ask questions to explore their data or build predictive models by asking questions in the same way that we might Google a question or ask someone “what are our revenues by region?” Additionally, Watson Analytics is packaged in an impressive user interface and design (which shouldn’t come as a huge surprise given IBM’s monstrous $100 million experience centric investment in this realm). Combined together, these traits make for an access point for non-technical users to be brought into the analytics fold. This creates the potential to broaden the role of data driven decisions throughout a company.
From a high-level view, Watson Analytics has many possibilities. Chief among them, however, is IBM’s hope that it can greatly augment analysts’ ability to do their jobs. Using Watson’s cognitive capabilities, the offering may be able to lend a very intelligent hand along each step of the data analytics process. Things like cleaning data sets, identifying relevant data points, building predictive models, or creating visualizations can be enhanced and made much easier when machine learning on your side.
For instance, a data analyst could upload a 10,000-row spreadsheet with missing, incomplete, or redundant data and Watson Analytics can not only automatically produce a completed and cleaned data source, but will also identify which rows are actually relevant to the analysis at hand. All in all, a process that might previously have taken an entire day can now be completed in minutes. If this sort of capability successfully makes its way from demo to a production environment, IBM will have made great strides in affecting the nature of enterprise BI.
Cloud-Based, Freemium Access Model
IBM’s second big announcement was that Watson Analytics would be delivered as a cloud-based ‘freemium’ offering, which is something new for Big Blue. This means that individuals will have access to core capabilities for free, with the option of paying for advanced pieces such as more storage space, or additional data connectors. More importantly, it means that prospective users can try it out without constraints around system compatibility, implementation times, or costs, thus presenting an opportunity to ‘try it before you buy it.” In many ways, this is IBM’s acknowledgement of how technology decisions are being made in enterprises today. Today’s world is one in which the business case comes first. Technologies are tested by line of business users, and once proven, the burden falls on the IT department to make them work. In that context, IBM’s focus on making it as easy as possible for line of business users to experience their product is a smart move.
Truth be told, the cognitive backend of Watson Analytics only utilizes a small fraction of the potential of Watson’s problem solving prowess. (Cleaning data sheets seems comparatively simple to helping doctors find cures for cancer). This is one area where the freemium pricing model comes into play: beyond the storage and data connections mentioned above, users will have the opportunity to tap into greater stores of processing power as well – for a price.
Watson Analytics and the Future of BI
IBM’s chosen interface and delivery model make a much more important point about where the market is headed. The promise of cognitive analytics, at scale, in a user-friendly package could be truly revolutionary. Cognitive computing’s integration into analytics solutions offers the chance to conduct analysis in ways that previously were not possible. As opposed to incremental advancements (think: stronger, faster, cheaper), when next generation advancements hit the market the market must comprehend their value before it will generate demand. IBM realizes that they must ‘show’ not ‘tell’ the benefits of Watson Analytics – a sentiment perhaps best summarized in the century old (and might I add apocryphal) words of Henry Ford, “If I had asked people what they wanted, they would have said faster horses.”
I’ve blogged previously on where I think cognitive computing will ultimately take the future of analytics as well as how the intersection of natural language processing and analytics creates new possibilities and access points for non-technical users. I see Watson Analytics as the first step towards what (to me) seems like an inevitable path. With IBM’s statement that the platform will be made generally available later this year, it’s a development that the industry should be eagerly watching.
1. For those of you who’d like a deeper dive into the specifics, or who might not be aware of the announcement, I’d recommend checking out the recording of the show or any of the write-ups from the tech news community.