If we look back over the past two or three decades, the growth in digital data has been driven by the spread of transactional IT systems. Every function in an organization today is likely to be supported by their own specialized IT applications.
The arrival of Management Information Systems (MIS) meant that data volumes were doubled as we took a second copy of the transactional data - one data copy for the transactional system, and an additional copy for the MIS reporting tool.
"In Order To Drive Innovation, Data Analytics By Definition Needs To Tell You Something That You Did Not Already Know. Or Even Better, Contradict Something That You Thought You Did Know"
But now the Internet of Things is driving exponential data growth, and this new phase of data growth is not limited by the number of IT users and the numbers of IT applications. The new sources of raw data have driven the development of new database technologies, such as Hadoop, and new data analytics tools.
IT departments now need to be able to scale data and data processing at high speed; speeds that strain the conventional processes involved in running in-house datacenters. And new Open Source data analytics tools are being rapidly developed to complement the existing data analytics tools that are available in order to achieve rapid scalability for data, and providing hosting for data analytics tools (open source and proprietary) the most logical option available is cloud computing.
It is unlikely that large organizations will place 100 percent of their IT data processing in the cloud, but the majority is likely to transition to cloud computing, especially in the case of data analytics.
The first challenge when deploying data analytics is the business case. Until you have meaningful output from a data analytics platform it is hard to say where there may be potential benefits. Should the data analytics investment be focused on manufacturing quality? Or spare parts inventories? Or customer satisfaction issues? All IT investments require a “leap of faith” to some degree in that the benefits come after the investment and are always dependent on more than IT successfully delivering, but data analytics investments are especially difficult to justify.
The second challenge is the application of proprietary knowledge to the outputs of data analytics platforms. Specialist knowledge is still required to correctly interpret the output. For example, a data analytics platform may indicate that there is vibration in a gearbox at a certain frequency, but does this mean that the whole gearbox should be replaced? Or just a single part replaced? Or is the vibration within acceptable limits and nothing needs to be replaced? So in order to make effective decisions, you need more than the output of a data analytics platform. Specialist knowledge is required. One day the combination of data analytics and Artificial Intelligence (AI) may be widely available, but not just yet. As a general approach, I would focus data analytics on known business issues (areas with opportunities for improvement) and areas where there is a high level of specialist, proprietary knowledge available in-house.
In order to drive innovation, data analytics by definition needs to tell you something that you did not already know. Or even better, contradict something that you thought you did know, therefore breaking a misconception. I know of one company that thought it fully understood the failure characteristics of its products and thought that its spare parts inventory was fully optimized. But they decided to run a small data analytics pilot to check out the tools and concepts. Consolidating data from several sources into a single data analytics platform proved that the failure characteristics were not understood and that the spare parts inventory was far from optimized. Innovation is often driven by the combination of several different data sources feeding a data analytics tool, with surprising results offering a new perspective.
In terms of growth, the objective is to acquire data points (leading to management decisions) that are not available to competitors or at least, not available to competitors as quickly. Speed can be a major advantage, but of course the corporate culture needs to allow for rapid “data-based” decision making. This means that the output from data analytics platforms needs to be put in the right “decision making hands,” along with the authority to make decisions, if the benefits of data analytics are to be fully realized.
I think that the coming years will be focused on Big Data or the Internet of Things, smart cars, smart buildings. All types of products will be communicating digitally back to the company that manufactured them. Every product that requires a significant capital investment will be digitally connected: planes, cars, building HVAC units, mining equipment, fleets of leased equipment, and so on. Big Data will require “Big Data analytics” tools to analyze the data, and both the data and the data analytics tools will be hosted in the cloud.
Smartphones have been driving voice recognition. Hands-free operation is ideally suited to voice recognition, but I expect that PCs and laptops will follow the smartphones and become increasingly voice driven.
AI will also be a developing area. There is probably more AI in operation today than we are generally aware of. But as the price points come down, AI technology will transition from the “exotic” towards “mainstream” and will become part of most organization’s IT portfolio.
Looking into the IT crystal ball is tricky, but here a few guesses:
• Plan on using more Open Source software
• A pendulum moving from on premise towards the cloud
• More IT investment outside of the “core ERP” platforms, with “core ERP” increasingly becoming an “IT commodity” item that takes a smaller slice of the IT investment pie
• Cloud platforms that for the first time ever will allow IT to move faster than the business
• Increasing business focus on speed and agility
The mining industry has already transitioned from being people-intensive to capital-intensive. All mining equipment will eventually be “smart enabled” and is well on its way, allowing communication both to the equipment manufacturers and the mining companies. Reliable networks will be needed to ensure connectivity. Data analytics will produce the information to allow mining operations to be fully optimized. Safety will continue to be a high priority for the industry, and the desired approach for underground mining is to have all the human operators safely above ground. The human operators will be remotely driving some very large underground drones. And machine-to-machine communications will be introduced alongside the hierarchic command and control systems.