12 Things You Must Never Do In Big Data


al-naqviUnless you are living under a rock as big as the Grand Canyon, chances are that by now you would have heard that something major is brewing in the technology world. Whether you read about it in Harvard Business Review or the Wall Street Journal, whether your CEO asked about it or your technology department pushed you to take a look at Big Data, most likely you would either be exploring or implementing Big Data. Big Data is truly “big” as it can lead to creating the most powerful competitive advantage for your firm – however, before jumping into Big Data you must know what not to do. No matter what, you must never do the following 12 things in Big Data:

ONE: Jump into Big Data without a corresponding business strategy

What makes Big Data “big” is your ability to gain an unparalleled competitive advantage, to get never-seen-before analytical insights about your stakeholders, to develop the next generation of strategies that will truly redefine how business is conducted, and to shape a new future for human civilization – and all of that means you must have a business strategy to pursue the Big Data projects. No, no, no … it is not the Christopher Columbus journey… Big Data is more like the mission to the Moon – complex but requires flawless planning. If your Big Data project team is not talking about your business strategy, something is wrong. The first step in developing a real use case for Big Data is to become intimately familiar with the overall strategy of your company– and the second step is to pioneer a new one with Big Data.

TWO: Start the project without understanding Value

Yes… dollars and cents matter… even in the Big Data world. If you cannot show the Return on Investment (ROI) for your project – you haven’t done your homework. Time to go back to the drawing board! Do not pursue a Big Data project that cannot clearly demonstrate its ROI. Demonstrating ROI goes back to the first point: where is the business strategy? Big Data can positively impact the shareholder value from increasing revenues, decreasing costs, reducing the cost of capital, mitigating risks, increasing capital efficiency, and helping you develop real options. It is your job to show your Board, CEO and CFO how your Big Data project will create value.

THREE: Call any large-scale data project Big Data project

A new gimmick in the technology world is to call any project that requires managing large-scale data a Big Data project. Wait a second! We have been managing large-scale data for decades – long before Big Data technologies became available. Then why all of a sudden are we calling other large-scale data repository projects as Big Data projects? It could be because we are being misled by our advisors or because we feel we can add more glamour to our project by calling it a Big Data project. It is important for business people to realize that the Big Data technology and use cases are very different from traditional and regular data collection and management. Not every large-scale data project requires Big Data technologies. You don’t need a large Caterpillar machine to seal your 12X10 feet driveway – a trip to Home Depot and a call to your neighborhood handyman will do just fine. Implementing Big Data is expensive and if you are doing it where ordinary technology could have done the job – you are missing the point.

FOUR: Ignore the Big Data cross-functional impact

Big Data is not like ERP, CRM, or SRM. It is not necessarily designed for specific departmental or functional focus. In fact, an infinite number of Big Data applications are possible. Why limit it to one department or function. Remember, the new power of business strategy will come from a new level of cross-functional visibility and collaboration. Can your regulators see how excellent your customer service is? Can your customer input be part of your HR management strategy? Can your shareholders have a direct say in corporate governance, thus making corporate raiding unnecessary or at least difficult? Big Data will define the new strategic frontier and it will not happen in isolation and with functional compartmentalization. Make your Big Data use cases flow across various functions, and beyond.

FIVE: Force fit traditional technologies with Big Data

“We will build a seamless bridge between the relational databases and the graph databases and then do the same on the vending side.” said an architect who was working on a Big Data project. The Big Data reference architecture continues to be a source of mass confusion. First, as we all know, there is no such thing as a seamless bridge. Second, if we are building so many bridges between the traditional technologies and the Big Data technology, perhaps we need to revisit our use cases. Many companies are trying to force fit their traditional technologies with the Big Data framework. This can lead to many problems – as many companies and government agencies are finding out. Don’t depend on your traditional architects to architect your Big Data project. They are still operating in the traditional technology world. Big Data is different. Bring in Big Data architecture specialists.

SIX: Select a product suite without understanding the details

This is a twofold problem. First, many suppliers of traditional technologies (e.g. ETL, analytics, and business intelligence systems providers) are claiming to support Big Data. While legally speaking they may be saying the truth, but it is nothing more than putting a new wrapper around an old box. Real Big Data technologies are different than the traditional technologies and their usage patterns are also different. Not all products can support Big Data projects. Second, within the actual Big Data technologies, the pace of change is enormous. You must truly understand your current and future requirements to develop the right technology stack for your Big Data project.

SEVEN: Ignore the importance of the support structure

Implementing Big Data without having the support structure such as Data Quality, Data Governance, Metadata Management, Master Data Management etc. in place is like going high altitude skiing without your ski gear. Building an integrated data management framework is a prerequisite and it must never be ignored.

EIGHT: Approach it as a technology (only) project

Enough of technology guys and girls telling you how to go about bringing this great technology in-house, now it is your turn to lead. Make it a business priority and a business project. Don’t think of it as technology – think of it as a business innovation.  Big Data must be led by business leaders and sponsors must include CEO and CFO of the firm.

NINE: Deploy a traditional “Tech plus Biz” team

“Bring in the technology team and have the business team work with them”… still trapped in the 20th century technology development process? This was the old way of doing things. When it comes to Big Data, in addition to your traditional business plus tech team, you would need some rather unique specialist skills. For example, your project may need the following skills: neuroscientists, statisticians/mathematicians, linguistics experts, data scientists, financial experts, and social/business/consumer psychologists. For real strategic leap and powerful results, build a solid team for success.

TEN: Have a function focused rather than enterprise focused approach

We have made this mistake a hundred times and we are repeating it again. When new technologies come, we tend to bring them in piecemeal. Marketing gets it… then finance experiments with it… then Human Resources tries it… and you know the drill. When we follow that approach we waste a lot of resources and time. First, we failto develop enterprise wide capabilities. Second, we acquire different products for different projects and that wastes resources. Third, we fail to benefit from the cross-functional benefits of Big Data. Fourth, we increase the total lifecycle costs. Why wait four years when someone will come in and tell us to rip out all the individual implementations and replace them with an enterprise platform. Do the right thing now: approach Big Data as an enterprise platform.

ELEVEN: Let your consulting firms do all the thinking for you

The Big Data field is in early stages. It is still very esoteric and only a small number of people truly understand it. The concept and knowledge gap is enormous – and that includes the consulting firms. So while your consulting firm may have published a 200 page report on what your counterparts are thinking and doing in Big Data, it may not have the right skillset to define and develop your Big Data vision and strategy. I know that when it comes to the Big Data resource shortage, everyone talks about the Hadoop skillset. But the real issue is developing the Big Data vision and strategy. Finding the Hadoop resources will be a lot easier than finding the right people who can help you develop your strategy. You must take the lead and bring in a wide group of experts to develop your Big Data strategy. Don’t just limit yourself to a single consulting or technology firm.

TWELVE: Limit your thinking

When I say Big Data is “big” – trust me, it truly is. Don’t limit yourself by approaching Big Data for some mediocre, non-strategic application. The potential of Big Data to redefine your competitive advantage is as big as your imagination. If you and your audience are not saying “wow” when your Big Data use cases are discussed – there is something majorly wrong. Big Data is not just a strategy – it is “the” strategy in today’s world – and if this strategic orientation is not beaming out of your projects, chances are you are limiting your creativity and imagination.

Avoiding the above 12 critical points of failure in Big Data can assure you success. Big Data is “big” but it can also be a big waste of your time and resources if you don’t do it right. The bottomline is that strategy matters.

Al Naqvi



No Comments Yet.

Leave a comment