Blogs

Tomato sauce and analytics: Variety is a must!

Content Marketing Manager, IBM Analytics Platform, IBM

In 2004, Malcolm Gladwell—author of Blink and The Tipping Point, among other books—once presented a TED talk in which he gave what is still one of my favorite descriptions of analytics and expanding one’s viewpoint. Gladwell said that American market researcher and psychophysicist Howard Moskowitz was convinced that different versions of the same base product must be created to satisfy the needs of various groups. Perhaps this concept seems obvious now, but in Moskowitz’s time, the idea was revolutionary. As Moskowitz once said, “You had been looking for the perfect Pepsi. You’re wrong. You should be looking for the perfect Pepsis.” And although he was never asked to create the perfect Pepsis, he struck gold when, on entering the world of tomato sauce, he gave the world a new option: extra-chunky.

http://www.ibmbigdatahub.com/sites/default/files/variety_embed.jpgAs I sit here drinking my Pepsi Wild Cherry, it’s hard to ignore two things: that someone at PepsiCo eventually took Howard’s original advice, and that the variety that Howard championed now extends even to the analytics landscape. And that’s a very good thing, because when you choose analytics solutions, you’ll likely have many diverse needs, all of which either need to be addressed now or will someday need to be addressed. Moreover, because even a single platform can offer considerable variety, getting the variety you need all at once has never been simpler. But you must also take care that the variety offered is the right variety.

For example, you might need to address several of the following needs, or perhaps even all of them: the need for a Hadoop-based analytics solution, the need for business intelligence and performance management, the need for predictive and prescriptive analytics, the need for real-time insights and the need for analytics that business users can easily understand. Let’s look at each in turn.

A Hadoop-based analytics solution

Enterprise-grade solutions based on the open-source software framework Hadoop are increasingly popular. However, when you add a Hadoop-based solution, be sure to focus on certain particularly important capabilities—among them a strong SQL on Hadoop capability, with high performance and scalability, and full integration and support for R. Other capabilities, such as the ability to perform text analytics or the ability to apply machine learning algorithms, can come as bonuses, and you should consider them as well.

Business intelligence and performance management

You might also need the comprehensive view of the business that business intelligence and performance management provide. Historical metrics play a particularly important role in understanding a company’s health at both broad and granular levels, and this capability should be supported across your organization—with global deployments if the need exists—and scaled cost-effectively. Moreover, you should consider which solution is ideal for you: on-premises, cloud or hybrid.

But the solution that you need might well go beyond a simple historical dashboard, and that’s where capabilities such as dynamic interaction can help. By allowing users to transition from viewing information to quickly and easily conducting deeper analysis, such capabilities provide a well-rounded solution.

Predictive and prescriptive analytics

You must also look ahead to capitalize on coming opportunities and to prepare for looming threats, and predictive and prescriptive software packages can help you achieve this goal as well. But you must look for capabilities such as the ability to perform analytics regardless of whether data is at rest or in motion, as well as the ability to automate data preparation. These two features in particular simplify analytics and make it inclusive.

Your chosen solution should also have a scalable architecture and should not require proprietary databases, thus simplifying access to your chosen data source. And you’ll want to make sure that your solution can analyze unstructured as well as structured data. Having this array of capabilities at your disposal can prepare you to face the future.

Real-time insights

Considering how quickly data is created, you have likely already noted that insights must be captured and acted on at once. Streaming analytics solutions can help make this happen, but many other resources can also help. Tools such as geospatial and time-series accelerators can come as welcome bonuses to your solution, and contextual awareness and other such capabilities can help you take into account cultures, languages and other contextual information—all while helping create a self-learning, self-correcting system.

Of course, you must make sure that the solution that you choose is enterprise-ready in every area, including integration, privacy, support and total cost of ownership. Fortunately, you have a wealth of information at your fingertips. Seek the opinions of industry analysts to help you evaluate a solution’s enterprise-readiness.

Analytics that business users can easily understand

According to the IBM Institute for Business Value, “Speed-driven organizations accelerate data analysis not only by using more advanced analytics, but by using them more extensively and broadly across the organization. This pervasive use of advanced analysis methods differentiates Front Runners.” In other words, analytics must be democratized, allowing people in all parts of the business to derive insights.

Thus, when choosing analytics on behalf of business users, minimize the burden that users must bear. Capabilities such as a natural language user interface can assist users by allowing them to query the data much as they would ask a question of another person in a conversation. In addition, data quality scores—which provide numeric indications of how good certain data is—can help users feel confident in their insights.

Whether for tomato sauce, soft drinks or analytics, having choices can help you meet your needs. IBM has taken a broad and deep approach with its analytics platform by offering solutions to address the five needs discussed here, among many others. Learn more about the full range of analytics options offered by IBM. And while you investigate, why not grab a soft drink and sit down to enjoy something topped off with your favorite tomato sauce—all while joining Howard Moskowitz in toasting the wonders of endless variety.