If we got a dollar every time someone told us that they’re democratizing big data, we’d have a hundred dollar bills, and that’s just so far this week.
So when an IBM representative sent a note saying that it was making Watson Analytics available to everyday business users to let them ask questions in plain language and get big data informed answers back in short order, our hearts didn’t patter too much. Even when they said we could try it for free.
After all, we’ve sat side by side, live and in person, with vendors who make big, big data analytics promises. And when we’ve said “Show me, turn me into a data scientist, big data analyst, heck, even novice user,” they haven’t been able to do it.
And maybe it’s us, but more than likely, here’s the deal -- in most cases, these vendors don’t include people like us (non-data workers) as citizens of their democratic, big data republics. And that’s fine, as long as we define that from the start.
That being said, yesterday IBM unveiled a freemium self-service data analytics tool that may actually provide us (you and me, and the guy in the cube down the hall) with the ability to ask questions in natural language and analyze huge volumes of data from which we can obtain actionable insights.
If you haven’t already heard, yesterday Big Blue announced Watson Analytics.
Can Watson Analytics Really Help Us Get the Same Insights as a Data Analyst?
We know that an IBM spokesperson would tell us yes. After all, that’s what yesterday’s hoopla was all about. So we sought out a few analysts and knowledgeable outsiders to ask if IBM’s SaaS solution could actually deliver on its promise.
Most of them said yes, but they (unintentionally and without prompting) qualified their answers with statements like, “Watson in the Cloud will allow a business analyst to get the same results as a junior data scientist.”
This left us, once again, with the idea that it wasn’t for ordinary Jacks and Jills who have no analytics training.
But John Myers, managing research director at Enterprise Management Associates (EMA) said that Watson Analytics is indeed for ordinary workers, like a retail store manager who wants to know why his purple shirts aren’t selling and what he might do to change that.
“Maybe he’d need help ingesting the data,” Myers qualified, “but after that Watson Analytics could guide him through the process by asking questions.”
“It’s truly getting down the path of democratizing data,” added Myers. “It’s both cognitive (meaning that it can think and “find paths”) and analytical, and can bring back the best answer via complex processes that the end user will never have to see.”
Big Data Insights for Everyone?
Watson Analytics, according to Alstair Rennie, GM business analytics, IBM Software Group, helps you upload data securely and guides you through sophisticated insight and visualization tools without needing to know the nuances of predictive analytics and statistics.
But that’s a simplistic view of big data and predictive analytics, say experts who provide sophisticated tools for data scientists and are working to democratize big data as well.
“Watson Analytics is a novel approach to bringing simple data sets and natural language questions together for common business use cases,” said Ben Werther the CEO and founder of Platfora. “It is a piece of the puzzle, but doesn't directly attack the problems of big data analytics -- i.e. making sense of massive datasets across transactions, customer interactions and machine data and giving business analysts visual tools that are native to this scale to amplify their understanding," he added.
Arijit Sengupta, CEO and founder of analytics provider, BeyondCORE, said that the biggest problem with Watson technology is that it's a "magical model" – it tells you the answer, but it doesn’t explain the reasons behind the answer so you can’t understand the "why" behind the answer.
He likened it to the “Expert Systems” and artificial intelligence that was hyped in the 1970s and was supposed to take over the world.
“There was significant evidence that expert systems were often much more accurate than human experts. However, no one could understand the reasons behind the answers given by the expert systems and thus would not act on their recommendations,” said Sengupta.
“People don’t bet their careers on magic models,” he added. “The ‘why’ behind the ‘what’ is more important, but IBM focuses on the what.”
Not There Yet, But Getting Closer
But Boris Evelson, a vice president at Forrester Research, said that Watson Analytics is a step in the right direction. He viewed the news from a historical perspective. He explained in an email:
We used to base all analytics on relational DBMS. They still serve their purpose, but they are limiting, rigid, inflexible. So we now have new types of DBMS like columnar, index, associative, in-memory, etc. They provide a lot more agility and flexibility.
The point and click, drag and drop user interface (UI) used to be the ultimate "user friendly" technology. That's passe now in favor of Natural Language Processing UI.”
We also, somewhat sarcastically, asked Evelson if the Watson announcement was a blow to vendors who have been working like dogs to make (big) data analytics easier.
“There's still more than enough room for many vendors to grow in this market segment,” he said, pointing out that Forrester’s latest research shows that only 12 percent of enterprise raw data is being converted into info, that more than half of analytics (often as much as 80 percent) in any large enterprise are based on home grown apps using spreadsheets and that there's a significant disconnect between business and IT.
Biggest IBM News in a Decade?
IBM, in its press release surrounding the Watson Analytics news, said that this was its biggest announcement in a decade.
“Is it a game changer? Is the news that big a deal?” we asked Myers.
“For IBM it’s a big day” he said, explaining that initially IBM was a company that delivers “a box of boxes” (hardware and software) that needs to be installed and configured by, or for, its customers.
Later, IBM began to bring in much of its revenue as a services firm (IBM Global Services) where IBM put the stuff in the “boxes” and even the stuff in other vendors’ boxes together for you, taught you to use them, supported them and so on. This is the relationship that IBM had with its market today.
So yesterday’s announcement was groundbreaking for IBM, in that Watson Analytics is delivered via an entirely new model. IBM owns the box of boxes, it puts them together on its cloud, and you pay to use the services it provides. The beauty is that all the end user needs to do is to log in and they’re good to go. You can be up and running in a matter of hours or days.
It’s a far cry from today where are RFPs, on prem or hosted data centers and long implementation and integration cycles are still the standard.
There’s also the freemium, crowdsourcing model to consider where individual workers choose their own solutions and IT doesn’t get involved until something becomes “a hit” or a check has to be written. Imagine all of the months of meetings, bad choices and user acceptance problems that that avoids.
My Data On Your Cloud?
Assuming that everything works as promised and users are as delighted as IBM suggests they will be, the big question is whether companies will be willing to give up their data. But “if you can’t trust IBM with it, who can you trust?” asked Myers.
Sengupta argued that the fact that it’s IBM doesn’t matter much at all.
“Nothing is ever gone on the Internet so the idea that you can upload to the cloud, analyze and delete is disingenuous. If the customer did not trust their data on the cloud, why would they care that they could have the data on the cloud for a period of time and then delete it?”
He also warned that IBM Watson learns based on the data and asked,
What assurance can we have that none of the client data is retained anywhere in the IBM system? What is needed is a solution that works on the Cloud but can be used on premise if the client so desires. Such a solution should bake privacy into the core architecture so that there is no tradeoff between the benefits of big data and the downside of privacy violation. With the right software design, we can have both: invaluable insights with zero privacy risk.”
Is Watson Analytics All That?
As Myers said, for IBM it certainly is.
But what about for the rest of us, is the democratization of big data really here? Based on what we’ve seen not only from IBM but also from Microsoft, Tableau, Alteryx, Platfora and many, many others it’s on its way. Soon we will all be data workers.