4 Steps to Big Data Litepost by Chris Curran on April 10, 2013
Every time I breathe the words Big Data to a group of business and technology executives I brace myself for a barrage of questions…. How is it different? What’s our business case? Where will we source the data scientists to drive deeper insights? How do we acquire the technology to process mountains of information at a rapid-fire rate? The silent question lurking underneath them all: What if we fail?
I’m not discouraged by their doubt. In fact, it’s quite the contrary. Their questions mean they aren’t prematurely plunging into Big Data and rooting around for random patterns without a plan to power the business, as I’ve seen in the past. Executives are starting to recognize that Big Data is much more than data and is beyond plowing though mountains of social media feeds. Approached with business goals in mind, Big Data is about gaining valuable insights into customers, products, markets and more and using that information to make better business decisions.
As I write this blog, the debate continues to rage about whether Big Data is the biggest buzzword of all time or an effective tool for changing the way companies do business. For us, the issue is settled. Seeing is believing, and we’ve witnessed first hand how Big Data is transforming how businesses are run. Here are some examples of Big Data revelations that have generated real results that are featured in our new white paper, “Capitalizing on the Promise of Big Data.”
- A leading global bank reduced loan default calculation time from a few days to a few hours across its portfolio of over 10 million mortgages. As a result, the bank is able to identify and hedge high-risk accounts much more quickly and effectively.
- A major life insurance company recognized that customer needs and preferences are evolving and so should their insurance underwriting and distribution process. Combining Big Data with predictive analytics, the company created a “model” life insurance market that resulted in dramatic changes in how it markets, sells, underwriters and distributes its products.
- A mobile telecom company conducted sophisticated time series of customer behavior to detect unknown predictors of customer churn. Through proactive account management, the company lowered the churn rate for those customers by over 30%.
Taking a bite-sized approach to Big Data is best. Launching one or a few Big Data pilots doesn’t require a substantial investment of time and money and eliminates the fear of failure that can hold executives back. Here are some specific steps to get started with Big Data LITE:
- Learn - learn about the problem or opportunity, prioritize the key issues and form the hypotheses you want to test with the analytics
- Insights - build a model (first on whiteboard or spreadsheet, then in analysis tools you are testing), analyze and generate insights and possible interventions, identify data from inside the organization (order forms, payments, sales leads, etc.) and link it with publicly available data (from market research firms, government agencies) that weren’t previously accessible, quantify the range of potential impact on key metrics
- Test - conduct tests, see what works or doesn’t work and generate the learnings
- Enhance - enhance models and plan to operationalize the analytics or interventions in the field and roll it out
In a way, Big Data is both a sprint and a marathon. You can realize an immediate impact on your business. For example, one bank we worked with analyzed a portfolio of 30 million complex cash flow instruments across 50,000 different scenarios in less than eight hours. However, getting to a point where the vast majority of your decisions are driven by Big Data will take time. Cultural change is a gradual process. For now, get started with a pilot project and capitalize on that “first-mover” advantage.
Image shared by dullhunk
Pingback: Finding a Home for the Chief Data Science Officer — CIO Dashboard()