9 Ways to Get Business Value from Big Data
Many use cases are available today as safe starting points for leveraging big data.
- By Philip Russom, Ph.D.
- August 27, 2013
In the 2013 TDWI survey on managing big data, 89 percent of respondents reported that big data is an opportunity, up from 70 percent in the 2011 Big Data Analytics survey. According to respondents in both surveys, the primary path to seizing the opportunities of big data is through advanced forms of analytics. In addition, other paths to business value from big data include data exploration, capturing big data that streams in real time, and integrating new sources of big data with older enterprise sources.
Although the consensus is that big data presents new opportunities for a business, few organizations are wringing much business value from big data today. In an effort to prime the pump, I offer nine established use cases that you should consider for your programs in big data and analytics.
#1: The primary path to business value is through analytics. Note that this involves advanced forms of analytics such as those based on data mining, statistical analysis, natural language processing, and extreme SQL. Unlike reporting and OLAP, these enable data exploration and discovery analytics with big data. Even so, reporting and OLAP won't go away because they are still valuable elsewhere.
#2: Explore big data to discover new business opportunities. After all, many sources of big data are new to you, and many represent new channels for interacting with your customers and partners. As with any new source, big data merits exploration. Data exploration leads to patterns and new facts your business didn't know, such as new customer base segments, customer behaviors, forms of churn, and root causes for bottom line costs.
#3: Start analyzing the big data you've already hoarded. Yes, it's true: many firms have "squirreled away" large datasets because they sensed business value yet didn't know how to get value out of big data. Depending on your industry, you probably have large datasets of Web site logs, which can be "sessionized" and analyzed to understand Web site visitor behavior. Likewise, quality assurance data from manufacturing leads to more reliable products and better leverage with suppliers, and RFID data can solve the mysteries of product movement through supply chains.
#4: Focus on analyzing the type of big data that's valuable to your industry. The type and content of big data can vary by industry and thus have different value propositions for each industry. This includes call detail records (CDRs) in telecommunications, RFID in retail, manufacturing, and other product-oriented industries as well as sensor data from robots in manufacturing (especially automotive and consumer electronics).
#5: Make a leap of faith into unstructured big data. This information is largely text expressing human language, which is very different from the relational data you work with most, so you'll need new tools for on natural language processing, search, and text analytics. These can provide visibility into text-laden business processes, such as the claims process in insurance, medical records in healthcare, call-center and help-desk applications in any industry, and sentiment analysis in customer-oriented businesses.
#6: Expand your existing customer analytics with social media data. Customers can influence each other by commenting on brands, reviewing products, reacting to marketing campaigns, and revealing shared interests. Social big data can come from social media Web sites as well as from your own channels that enable customers to voice opinions and facts. It's possible to use predictive analytics to discover patterns and anticipate product or service issues. You might likewise measure share of voice, brand reputation, sentiment drivers, and new customer segments.
#7: Complete your customer views by integrating big data into them. Big data (when integrated with older enterprise sources) can broaden 360-degree views of customers and other business entities (products, suppliers, partners), from hundreds of attributes to thousands. The added granular detail leads to more accurate customer base segmentation, direct marketing, and other customer analytics.
#8: Improve older analytic applications by incorporating big data. Big data can enlarge and broaden data samples for older analytic applications. This is especially the case with analytic technologies that depend on large samples -- such as statistics or data mining – when applied to fraud detection, risk management, or actuarial calculations.
#9: Accelerate the business into real-time operation by analyzing streaming big data. Applications for real-time monitoring and analysis have been around for many years in businesses that offer an energy utility, communication network, or any grid, service, or facility that demands 24x7 operation. More recently, a wider range of organizations are tapping streaming big data for applications ranging from surveillance (cyber security, situational awareness, fraud detection) to logistics (truck or rail freight, mobile asset management, just-in-time inventory). Big data analytics is still mostly executed in batch and offline today, but it will move into real time as users and technologies mature.
For more information on the business value of big data, please register for these related, upcoming TDWI Webinars:
Delivering on the Promise of Real-Time Business Analytics, September 12, 2013 at 12 noon ET, 9am PT
Big Data Analytics: Getting Business Value from Big Data via Advanced Analytics, September 26, 2013 at 12 noon ET, 9am PT