The world is creating new digital information at escalating rates. The bulk of this information growth isn’t generated in a planned and structured way as is done by scholars and scientists. Rather, it’s a byproduct of the frenetic digital activity of the world’s population, as well as the enormous digital output of the world’s sensors. Although much of this vast information is stored and accessible, it remains largely untapped. However, a growing number of businesses are analyzing this information on large scales in order to find useful and profitable patterns, trends, and associations. This activity is called big data. Here are two examples of how business uses big data:
Insurance companies base their premiums on how likely an individual will file claims. Individuals deemed likely to file a lot of claims will pay more than those deemed to file less. Since it is (or was) impossible to know this on an individual basis, relatively crude statistical predictive methods are used. For example, because young men in their teens and twenties get into lots of car accidents, then everyone in this demographic will pay high premiums. This is unfair to the exceptional drivers in this group.
Some insurers offer a “pay how you drive” option that uses a telematic device installed on the insured’s car. The device relays an individual’s driving information such as speed, acceleration, time of day, distances, and driving locations back to the insurance company. This information flow is ongoing. An individualized risk profile is generated on the basis of this data.
For example, high accelerations, high speeds, long distances, night driving, and car use in high crime areas indicate a high risk driver. The individualized risk profile serves as a more fair basis for the insured’s premium rates. Storing and analyzing the ongoing data streams of millions of customers is a classic big data problem.
Businesses have access to reams of customer data in the form of email correspondence, customer service interactions, newsletter sign ups, and purchasing histories. This data, when analyzed on a large scale, allows a business to make offers based on individual behavior. It allows them to identify the habits and characteristics of people who are most likely to become loyal repeat customers. It also provides insight into how to convert one-time purchasers into repeat customers.
Big data analysis of social media interactions reveals on a huge scale, the what, where, how, and why behind consumer behavior as well as their feelings and attitudes towards brands and products. If you want to know why some products are taking off while others are dropping off, and which types of products will be future hits, it’s all in the data.
The Role of Cloud Computing in Big Data
The downside to big data analysis from the business’s point of view is that it requires a big investment in large servers, IT staff, and people with other specialized skills. It also requires a large time investment in procuring, installing, configuring, and testing the technology. This is a big risk in something that may not provide sufficient payback for a particular business. If only there were some way to do a proof of concept test of big data analysis before putting in the big bucks. As it turns out, there is.
Cloud computing gives instant access to infrastructure well suited for big data analysis. A business can simply rent the required resources from the cloud service provider. Some projects may require more resources than others. This is readily accommodated by renting as much as needed at any point in time. Cloud computing is also well suited to the business that wishes to extend or augment its on-premise big data systems.
For more information about cloud computing and how it can help your business, contact us today.