In my previous blog, I wrote about the Internet of Things (IoT) and the opportunities afforded to organizations that embrace working within the “long tail”— the space where for many organizations true revolution in data insights will happen. In short, working within the long tail allows you to recognize data patterns at the edge of your enterprise that can empower your company with new insights for improving your business—and your bottom line.
However, recognizing these opportunities are one thing, operationalizing them at scale is another, a challenge I call the path to production.
Web Business Agility
For web companies, their requirements to manage and analyze data at Internet-scale led to the invention of Hadoop and big data analytics techniques. In many instances web companies are able to operationalize new insights quickly, with proprietary infrastructures for testing changes to their sites and then promoting ones that prove to be valuable. As early and vocal advocates, they convinced IT and business owners in other industries to adopt big data technologies. I believe, however, that the relative simplicity of their path to production set unrealistic expectations on what can be done in traditional brick-and-mortar industries.
I’m not trying to be critical. Web companies pioneered the technologies and analytic approaches that will be incredibly valuable to brink-and-mortar organizations as they virtualize and capture vast amounts of data about their physical world using IoT architectures. However, it’s important to recognize that web businesses are fully instrumented. What do I mean? Everything a customer does with a web company’s product or web site is known. Moreover, the web company’s engineers and analysts are likely users of the products they sell, so there is a much deeper understanding of the business than what might be typical in physical-world companies.
Web companies also have infrastructure in place so that data scientists with access to relevant data can develop a new insight in the morning, put it into limited release for testing in the afternoon, and, if it proves out, promote it to full production by the end of the day. The potential impact to the business is significant, even for something as simple as changing the location or color of a button on the home page. The difference is that web companies can test the impacts of various changes and promote them with an agility that physical-world companies simply cannot match.
Complexities of physical-world companies
Compared to web-based businesses, the reality of operating in the physical world is much more complex. Imagine, for example, that you wanted to track IoT data from a transportation company in charge of busing. You could put sensors on buses to capture data about driving behavior, location, mechanical status, bus performance and riders. Armed with this incoming data you could now predict trouble spots and adjust routes in real-time by joining sensor data with historical rider data, real-time traffic and weather information.
But consider this: will your union contract allow you to make changes on the fly? Can your operations support rerouting drivers, expanding service to accommodate a spike in riders on a particular route, or reducing it when traffic is less than planned? It’s not that simple, is it? You can see the complexities inherent in a physical-world setting, and how it can get a bit more complicated than just changing the location of a button on a homepage.
In the end, it’s really about overcoming constraints. First, realize that there is significant value for IoT data for both web and physical companies. Yet, we must also be pragmatic enough to understand that there are often real world social, political, economic and perhaps even cultural obstacles to overcome in effectively using IoT data to make better business decisions.
The opportunity is ripe for taking, but strong leadership will be needed to affect change. Are you up for the challenge?