Three key to making big data yield high ROI

According to IDC, the world's largest data technology and services market will grow at a 31.7% annual compound growth rate (CAGR), which is expected to reach $23.8 billion trillion in the 2016 big data market. As more and more companies enter large data areas, the immediate question for companies is: How do you make large data operational?

More specifically, c-lever managers want to know quickly whether the organization can get "nuggets" buried in large data repositories. Large data workers do not want to be in a position where "gold" is difficult to obtain. (Historically, geologists estimate that 80% of gold prospectors did nothing for the California Gold Rush of 1849). There are three ways to avoid the consequences of this tragedy. (Note: If possible, 1th and 2nd should be considered before implementing large data.) )

1. Understand your Purpose

Matt Ariker, chief operating officer of the Consumer Market Analysis Center at McKinsey, the IT consulting and research firm, wrote in a blog post, "... The promise of big data is so tempting that people are increasingly looking for Pb-or EB-level gold insights to grow and beat rivals. This process allows many marketing to jump from one problem to another and then to the next. Great insights do exist in large numbers, and companies that use large data well are moving ahead of their competitors. But one big reason is that before you start, want all of this data to do what they have a very clear goal. ”

Ariker urged companies to start their big data-mining journey with "purposeful thinking" (Destination think). In other words, the marketing department and any other department that plans to use large data should sit down with it and identify key performance indicators (KPIs) that are expected to improve from large data, such as building market share in communities with low service levels. Large data problems should be carefully designed with a clear intent to discover valuable insights that can be paid immediately. Actions and see if predefined KPIs are compliant.

2, check your data samples

A recent study published by Princeton University's Center for Information Technology Policy (CITP) and the University of North Carolina at Chapel Hill (University of Carolina at Chapel Hill, UNC) warned marketers that they were looking for inquiries from social media outlets ( such as Twitter and Facebook) collect consumer data to give absolute confidence.

The draft report "Large data: traps, methods, and concepts in an emerging field" (Big Data:pitfalls, Methods and concepts for a emergent field), by Professor UNC and Princeton CITP companion Zeynep Tufekci , she challenged large data methodologies that relied heavily on social media insights, and compared with Drosophila fruit flies, which are easy to use and can be nurtured in laboratories and produce immediate results. The problem with fruit flies, Tufekci says, is that it does not necessarily represent real-life scenarios, so it can distort research. Similarly, the results from the Twitter survey, a typical marketing audience participation only 10%, or Facebook, only represents a certain segment of the market, it may distort the results.

The lesson here is that you don't get reliable action based on the fact that you can't accurately reflect the market's data. You need to verify that the big data really represents your entire market.

3, find ways to create rapid success

Install a new large data solution to see its actual ROI within 12 months, one way is to harvest some of the most readily available fruit. Most people who already have experience in operational marketing know that when the project is stacked up, publicity, product launches, and branding activities are the first, and then exploratory research, which makes it possible to make large data available as early as possible.

There are also a growing number of cloud-based analytics providers that can help companies gain experience and confidence in finding bullion in large data ores. Some options include IBM's Infosphere Biginsights,google BigQuery, and Microsoft's Windows Azure.

The marketing department seems to understand the urgency of allowing large data to operate. This was reflected in a June 2013 study by The Economist think-tank (EIU), where marketers listed "the ability to use large data analysis to obtain predictive results" as their highest priority.

Conclusion

Now that the test of investment has set in, marketing and IT departments must find ways to reduce the time it will take to create business insights, to keep fidgety c-lever managers happy and to meet their expectations.