5 common mistakes in Big Data Analytics
Big data projects like any project requires understanding the possibilities, priorities, and planning. Current trends in promoting software solution does not automatically lead to value generation. Software execute the analysis needed but what analysis are needed?
For that one needs to know the goal, map the business data collection processes, see what internally or externally exists but all in light of certain questions. In what follows, some of the key mistakes with regards to big data analytics are elaborated. Their recognition can help establishing an effective approach in use of big data.
1. The Haystack Syndrome
Late Dr. Eli Goldratt in his book published in early 1990s, entitled Haystack Syndrome – The: Sifting Information Out of the Data Ocean, beautifully explained ahead of its time that a "must" for every manager concerned with meeting the challenges of the 21st century is to see the differences between data, information, and decision-making in a new light, the goal – creating value, making money! The quality of analytics today hangs all into the goal. When Big Data buzz word was not hovering around every mangers head, Eli mentioned all starts with the structure of organization and the understanding what we want to do with the data, posing the right question and underpinning then right the information architecture rather than searching for software. Over quarter of century later, the Haystack Syndrome in the new light of real Big Data suggests failures are due to having no objective, wrong-formulated questions, followed by wrong architecture and structure in the organization. Therefore, one cannot achieve the goal of generating value through Big Data.
2. Crowd mentality
While the word "big data" has already reached top of every search engine yet only a few people really understand them. Without having the “right” people around, following the crowd just because everybody is doing it, brings about more chance of failure than any success. The need for going for big data must be studied very carefully and patiently – there is no plug and play recipe around, no one size fit all is available. And surely, it may not fit every organization, but again, it is about understanding the reason why it should be done in the first place. With understanding of the vision and/or the goal comes the proper planning that helps identify the right people, the right process, and the right technology for big data.
3. Big Data Silhouette
Almost all Big Data appraisals papers suggest a picture where Data Warehouses can nicely and easily be connected; in theory yes but in practice not. This has created in reality a situation where many businesses attempt to organize and build data hubs for various functions. However, a Big Data Silhouette suggests a solid Data pool where the total picture can be solidly seen, where actually, the walls between functions / business units are removed and data flows through for more value generation.
4. Stakeholders-wide Engagement & Change Management
As much any other project, big data value creation requires a good change and project management system in place. Often as any change project, too much of attentions is paid to the technical sides and the human capital in a project is overlooked. Applications of Business Analytics tools, collections of right data, constantly mentoring risks and security issues involved, all must be touched by trained people, they need to be directed and led towards how the new organization operates. Failure to see the importance of change and project management results in unbalanced emphasis on various aspects of a project and lack of acceptance of all stakeholders.
5. Ethical, Legal, Regulatory, or Privacy Compliances
Big Data Analytics bring about various formats of accountabilities. Failure could be due to not recognizing these in design phase by ignoring at the ethical, legal, regulatory, or privacy compliances issues. The massive data collected in various formats and interlinked can easily infringe the peoples and organizations’ right. Disagreements of the laws with certain aspects of a Big Data project may cause splitting and isolating certain "activities” which then may hurt the core to how a company uses its data. The problem can exacerbated when different projects using the very same “activities” – therefore the strategic priorities would be jeopardized.