Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

From Big Data to Botched Data: 5 Steps to Total Big Data Failure

Sri Narayanan | Oct. 7, 2014
How to avoid blunders in modernising your business intelligence efforts.

There is an acute shortage of skilled data analysis employees in the labour market and the situation is likely to worsen. But more than data crunching and creating fancy algorithms, a failure to deploy the real 'experts' in your industry to a Big Data project will certainly doom the effort.

Process Any And All Data Indiscriminately

If you fail to plan, you plan to fail. There's just too much unsorted data to process in most Big Data projects. Failing to run a data audit so you know what bits you actually want to process will most certainly cost money, loss of time and resources that will detract from the project's primary goal. JC Penney transitioned from a traditional BI environment managing structured data to a huge Big Data effort involving large datasets of unstructured information, processed through fancy predictive analytics and algorithms. The results lacked context and relevance, and the project failed.

Selecting wrong use cases

Seduced by the potential of Big Data solutions, many companies aim too high for their initial efforts. They study successful deployments from other organisations and attempt to apply the same use case scenarios to their companies without the relevant skill sets or experience. Others use the same cases from their previously installed traditional BI programmes, and wonder why they don't see any benefit when migrating to a Big Data initiative.


Go For Broke
JC Penney's first foray into predictive analytics was a massive undertaking that was just too ambitious, too soon. It was expensive and high-risk. No one knew how to make meaningful decisions with the information gathered.   Often a narrow focus and a smaller project will reap far more significant results.  

Modernising BI: Doing Big Data The Right Way
If you haven't already noticed, a huge shift is taking place in the corporate world. Companies are moving from traditional BI to predictive analytics with high volume, unstructured data to help their businesses extract insights and drive them forward.

Modernising BI to a more open Big Data effort is a journey you really can't take alone. It's fraught with landmines and requires careful consideration. As the JC Penney example illustrates, any company can and will get this wrong.  HP Business Intelligence Modernization Services identifies three major service needs your company will require to get your Big Data projects on the right track:

- Discovery environments: "Data lakes," data visualization tools and services enabling rapid, enterprise-wide data sharing and discovery collaboration.

- Analytics solutions: Addressing specific analytics needs to run the business better.

- Hybrid data management: Enabling enterprises to deliver production-grade analytics integrated into business processes and systems that leverage relevant data.


Previous Page  1  2  3  Next Page 

Sign up for Computerworld eNewsletters.