These days, everyone is hearing about Big Data, but few companies know the best ways to implement it. In this guest post, IT instructor Michael Dorf has some advice for businesses.
Everywhere we look today, the talk is about Big Data and how it’s affecting enterprises in every conceivable industry. From automotive to zoos, data is growing fast and handling it is the prerogative of the IT administrator. Metrics at all levels are being tracked. No longer are databases full of simple text, now they’re filled with consumer preferences, purchase histories, graphics and other media, social connections, and more. The amount of data grows and businesses wanting to stay on top of the market must learn to not only store, but analyze and make all that Big Data useful in tangible ways.
Today’s holy grail for data control is to be able to ask any question and research data sets in any way we wish without limits as to size, scope or availability. Those who are able to do this see a transformative effect on business models and income streams as new and better strategies become possible. Most business executives understand that Big Data means bigger income.
Here are some ways that you can transform Big Data into bigger business.
Quite often, businesses make the mistake of jumping on the Big Data bandwagon without a smart plan or strategy because they haven’t looked at the possibilities beyond esoteric lists of wants drawn from non-specific literature. It’s great when an article, magazine, book or the presenter at a seminar tells you that you need to harness Big Data, but the vague, non-tailored information they include for doing so will not necessarily apply to your business.
When this inefficient, dreamy approach is used, it usually results in a lot of money being thrown at the idea without much direction or result. This nearly always means that the business will fail to harness its data in any significant way and will therefore see little or no return on investment (ROI) from it.
Studying how your specific data as well as the ways others in your industry are using it is paramount. If you do not learn your needs and aren’t able to quantify them with expected gains, then a Big Data plan will almost assuredly fail.
Setting up clear experimentation based on ideas that come from your needs and your industry’s expectations is a solid foundation for knowledge. Conduct hypothesis-driven experimentation internally and, if necessary, through contractors knowledgeable about the job. This will create prioritized funding that ties both your business and your IT specifics together. The efficiency also pays off at the ledger, of course, and should start with the “easy kills” first and work its way into the more complex and costly solutions over time.
Big Data is now your business
Whether you make widgets or sell services, you should make Big Data the center of your business model. In today’s environment, whatever the industry, data is the driver. Your design, planning, manufacturing, selling, and investing will all feed into and come out of your data center. That center is where you track sales, manage purchases and inventory, hold payroll, and look for future growth investments. Once this is realized at a very basic level in your corporate DNA, then you’ll be ready to really harvest the rewards of Big Data.
At every level of your organization, access to and manipulation of the data will drive things. Better decisions from management will come thanks to solid analysis, streamlined manufacturing will come thanks to the efficiencies better data input can give, and sales will increase thanks to better customer interaction via the analysis of purchase histories and expectations.
Of course, the incorporation of Big Data and this new culture is demanding. Insight and leadership will be the driving force behind it and the “trickle down” of this new DNA from management to the rest of the company will likely be slow, but well worth the effort. Once the benefits become obvious, resistance will fade and adoption will accelerate.
IT needs to be able to keep up
Much of this innovation in your business will hinge on your information technology department as well as upper-level management. On average, data doubles every two and a half years and if your IT planning does not take this (or more) into account, you can expect that it will run into constraints sooner rather than later. Many of the previously standard IT practices for data collection, storage, and management no longer apply. Architectures are very different and data storage is often not always done in-house.
Much of the change requirements will be driven by the type of analysis to be done rather than the data being stored. Anyone can throw more hard drives into the server bank and store more data. That’s been done for years. What’s changed is that the data must be much more accessible and in multiple ways and by multiple users. Often in scenarios where those users are not only accessing it simultaneously, but in ways that cross-reference one another while doing so. Data is no longer linear, it’s now a 3-dimensional web and must be treated as such by both storage and access machines.
Often, the way data is handled will change over time as business needs morph into new paradigms. Perhaps yesterday, customer purchase histories and contact information was all that was important, but today minute changes in purchasing and indications of customers look at your competitors (or even data about your competitors directly) is being integrated while supply-chain corroboration is also added as another metric. All of this needs to be accessible in real-time by sales staff as they try to negotiate new sales.
Pulling together a Big Data system from scratch is no easy task. Usually, however, it’s not being done from scratch, but is instead being done from existing frameworks that may or may not be conducive to the new way of doing things. This adds complexity to the task. Big Data is coming of age and those who aren’t on board will be left behind.
About the author: Michael Dorf is a seasoned software architect and instructor with a M.S. in Software Engineering and a dozen years of industry experience. He is a co-founder of LearnComputer (learncomputer.com), an IT/Open Source training school based in San Francisco Bay Area. Our Big Data Overview training course is designed for IT managers who need a fast track to Big Data solutions available on the market today.