Enterprises are already thinking about an IoT-intensive future. For example, Verizon's chief data scientist spoke at the recent IoT World event about the big data analytics platform that they are building to capitalize on the IoT. Kshitij Kumar, who blogged about the speech, mentions how Verizon is understanding driver behavior from the billions of miles of data collected from "connected" vehicles. There are some interesting observations such as the homogeneity of driving behavior across model years in some makes versus others. But, as the author notes, it is unclear "how these observations could be converted into revenue."
Therein lies one issue. We may amass a treasure trove of data, but we may not be able to do more than play with it, as one would with a toy. What we really need is an ability to find patterns in the data that provide insights that illuminate the path to fruitful action.
Will data scientists lead us out of darkness? According to the blog post, Ashok Srivastava, Verizonís chief data scientist "suggests that we need to leverage the resources out there and think broadly to build a data science team." But is data science all that is needed, and is the lack of data science talent the first thing that enterprises need to worry about?
The science is sound, no doubt. It is founded upon statistics, machine learning, and other proven disciplines. But what about the data itself? Thatís a whole different story. Itís a bit like Eliza Doolittle waiting for Professor Higgins to help her speak "proper English." We may need to teach our data to speak "proper business-ese." There will be lots and lots of it!
Data: lots of chocolate for me to eat
The worldwide data projections in the Goldman Sachs report shows data growing from under 10,000 exabytes in 2014 to four times that by 2020. Much of this growth will likely to be generated by sources that lie outside the edge of the enterprise, including the "things" like smart devices, sensors, and other gadgets that are already starting to pop up around us.
Surely, the volume, velocity, and variety of data that we talk about are not something to worry about? After all, the industry has been busy with all kinds of technology to throw at the problem.
Sure, our computers scream along at light speed, and our software can reduce everything that's thrown at us into little pieces and delegate the work to commodity hardware that crunches data effortlessly. But our enterprises include people too. We are a bit slower. Quite a bit. We are also somewhat unpredictable and unruly, unlike the computers we create. We bring mixed-up objectives and agendas, emotions, attachments to the past, and all the rest of the pleasure and pain associated with being human.
To make data science work, to set the technology humming, and to render the data enterprise-worthy, we need to somehow overcome human limitations to make wise decisions and take timely actions In the face of all this complexity. Fortunately, we have been working on a couple of basic disciplines.
Help us control ourselves. Help us think clearly We have been talking about data governance for a few years now. At its core, it's nothing more than making us all behave in an effective manner to treat data as the corporate asset that it is, so that we get the business performance and value gains that are possible.
We also need a way to think through the complexity of the problem space and the solutions that donít seem to stop coming at us from vendors and technology innovators. Information architecture and planning are disciplines that can bring some method to the madness. Data governance and information architecture/planning are not new. Yes, as Led Zeppelin sang, the song remains the same, but we just need to learn to sing like Robert Plant. It requires discipline and hard work, just as it does to put the enterprise into data. Without good data, data science will be nothing more than a hack.
Do you think enterprises are ready for the data deluge? I'd love to hear your thoughts. Please share in the comments field after the infographic.