Financial institutions and retailers are adopting Big Data analytics to generate greater business intelligence and gain competitive advantage. The objective is to provide real-time experience to customers. However, except Google and Amazon, most companies are yet to see the daylight of Big Data analytics success. According to a recent study by McKinsey, three-quarters of the analytics leaders from major companies interviewed admitted to have realized less than one percent improvement in revenues and/or cost with Big Data implementation. According to another recent survey 75 percent of big data/IT projects in the broader industry are incomplete.
What are the reasons behind these failed Big Data implementations? Can companies overcome these barriers to reap higher ROIs? How can enterprises declutter for greater wellness of Big Data and realize greater benefits?
Here are some of the reasons leading to minimum or no realizations of benefits from Big Data implementations:
- The inability of the analytics team to proceed past the pilot stage of the project due to data-relates issues.
- Improper data processing due to non-compliant, inconsistent, inaccurate and untimely data.
- Unverified data due to an unmanageable plethora of sources ranging from databases to large flat files.
- Extremely long time taken to ensure data quality and value throughout the data processing lifecycle.
- Validation of data is performed manually at every stage.
- Time spent in creating, executing and maintaining a suite of scripts and algorithms that did not integrate well with the tools used at various stages of the lifecycle.
- Lack of uniformity in the design and creation of various scripts used to run the system.
- Lack of a holistic assurance approach to data lifecycle management.
Here’s what companies can do to establish data wellness:
- Implement a data assurance strategy to ensure the quality of data that was being pulled into the system.
- Ensure the data assurance strategy is comprehensive and easy to implement.
- The assurance process should be based on a Test strategy that expedites the detection and weeding out of bad data throughout the data lifecycle.
- The Test strategy should dissociate itself from the technology aspect and focus solely on creating value.
- The testing methodology should be able to accurately identify, verify and validate all key activities in the data lifecycle.
- The testing methodology should be measurable, repeatable and should address various technical and process-related challenges.
According to an estimate IDC, eight zettabytes of data will exist by 2015. It is, but obvious, that not all of it is relevant to an enterprise. By decluttering and focusing on what is relevant, reliable and verified, an enterprise can improve its data quality. Enterprises implementing Big Data projects need a comprehensive Data Assurance strategy for the entire Big Data lifecycle.
What are your views to data wellness? Share your comments in the section below.