Predictive Analytics

Insights from SAP Insider BI2014 – BI for the Business User?

Those of you who follow me (and my blog posts) might have scratched your heads a bit earlier this week when you saw me Tweeting from the SAP Insider BI event. After all, I generally write for an audience of business users and let’s face it, it is the IT department that buys and uses BI tools. Well, I was there because of all the exciting new advances in BI tools and technology that have the potential to assist you, the business user in gaining insights and intelligence about your business. But more importantly I was there because so few of you are screaming for these new tools. As Steve Lucas, President of SAP Platform Solutions, said on stage, “When was the last time a business user came up to you and asked for more middleware?” Bingo! As business users, you don’t want more (or any) middleware; you want answers to your questions and solutions to your problems. The problem is, many of you have become so frustrated waiting (too long) for simple things, like a new piece of data added to a report, that you’ve stopped asking.

Case in point: You also might have noticed that you only heard from me during the opening keynote. Why? Did I just breeze in and out? Did I lose interest? Was there nothing else of importance to comment on? None of the above. After that first keynote I was never able to actually connect through the network available in the other meeting rooms. In the “Happiest Place on Earth” (yes, we were at a Disney resort), it didn’t make me very happy to have no access to Twitter, HootSuite or any of the other platforms I would have used to communicate. But actually saying “never” is a bit misleading. The reality was, after the first few (or dozen) attempts, I gave up trying. And that’s exactly what happens to the business user in search of the next bit of “intelligence” needed to make a serious impact on the business. After a certain period of frustration (minutes in my case, years in yours) you just assume it is too difficult and it is not worth asking.

While some of this new technology is new enough to be bleeding edge, and not everything that gets presented and discussed is generally available, I can tell you: Now is time to start asking for more. Yes, there will be a price tag, but when did you ever get something for free that really added value? And if it adds enough value, it should pay for itself in a reasonable period of time. But in order to find that potential value-add, you might have to start thinking out of the box. In fact, maybe you should re-evaluate problems that were deemed unsolvable in the past.

That is certainly the case when it comes to predictive analytics. Everyone would love to forecast, predict and plan accordingly, but hardly anyone is really doing it today. I grew up (professionally) in the world of manufacturing. Everyone knows there is one universal truth about the forecast: It is always wrong. It’s just a question of how wrong it is. Of course there are all sorts of different forecasting tools and models to use, but the problem is, they are usually very complex and only as good as the input provided. And how does the typical businessperson know which variables really matter? Are some truly indicative of cause and effect, or is it just a correlation? Do you really understand the difference? What if you didn’t have to?

What if you had a tool that allowed you to just throw a lot of variables at the problem? What if you had a tool that was smart enough to pick out the ones that matter and come back with a prediction with a 99% confidence level? That’s exactly what SAP InfiniteInsight is supposed to do. Never heard of it? I’m not surprised. First of all, the name itself doesn’t exactly speak volumes about what it does. I am a firm believer that the name of the product should tell you (or at least hint at) what the product does. There is really nothing in InfiniteInsight that even implies prediction.

Also, the product comes to SAP through acquisition. The acquired company is KXEN, formerly a privately-held San Francisco-based company. Its specialty:  the automation of (at least many aspects of) prediction so that business users can make forward-looking decisions. So does this mean this is a self-service function and the business user doesn’t have to wait for the technical staff to do something? Not really. Those variables the business user wants to throw at the model? Well, someone has to know where they are and how they are represented in the grand scheme (or schema).  But that should be relatively easy for the technical staff and tech-savvy business analysts might also be able to construct the model. So we’re not talking days or weeks; we should be talking hours. This makes the whole process that much more dynamic. How many times have businesses stuck with a forecasting model that was known to be flawed or no longer reflected the real business environment (after all, things change over time) just because it was too difficult and time-consuming to change?

Production forecasts, sales and operations planning and financial planning are just the tip of the iceberg here though. Getting accurate predictive analytics will become contagious. Get good results in one area and you will start to think of all the other ways you would like to “predict.”

And how many times have massive volumes of data sat largely dormant because there weren’t tools to handles such volumes in real time? Those days are also gone. While the Internet of Things is all the rage right now and spells huge opportunities in many different disciplines, this is not a new problem for manufacturers. Many have been collecting these massive volumes of data from sensors on the shop floor for years. And while these sensors might have the ability to shut down an automated production line as temperature or viscosity or any number of other measures start to drift out of tolerance, how many are able to accurately predict when that might happen?

Since preventive maintenance also causes a line to shut down production, timing is critical. Shut it down too early and you lose more production. Wait too long and you pay an even greater price. And yet without the ability to not only collect, but also process and analyze colossal volumes of data in seconds, that data is massively underutilized. Moving data off storage devices and into memory is the key to improving speed and in this case SAP will turn to HANA.

And finally, a third SAP product will play a key role in drawing the business user out of his or her complacency. All the predictive capabilities and all the speed and processing power in the world will be meaningless to the business user if the tools can’t present the results in a meaningful way. That’s the problem SAP Lumira is intended to solve.

Here’s another product name that doesn’t tell you what it does. No it doesn’t treat rheumatoid arthritis or diabetic nerve pain. Does it illuminate data? Hard to say. But even though it doesn’t tell you what it does, the name has a couple things going for it: it’s short and it doesn’t say the wrong thing; it makes you ask what it is. Here’s what SAP says it does:

“Gather and quickly make sense of your data with SAP Lumira, our easy-to-use, data visualization software. In just a few clicks, you can combine and visualize data from multiple sources – presenting both big picture and granular insights in a single view. Use a drag-and-drop interface to create beautiful visualizations, explore data, and share insights with your team.”

Does this mean as a business user you can do it yourself? My guess is, you will still need some help from IT or that tech-savvy super user. Because you still need to know how your enterprise data is organized. But if you have data in spreadsheets now (and who doesn’t?) you can download it for free and try it out for yourself. You’ll educate yourself on the different possibilities and perhaps even come up with those new (out of the box?) ideas.

The bottom line: you don’t have to be a BI expert or a techie to gather more intelligence and insight today. And you don’t have to be a psychic to predict the future. But if you sit back and think all this stuff is too difficult and time-consuming, you better hope your competitors aren’t figuring out that it isn’t.

Tagged , , , , , , , , , , ,