Why the IT department can't ignore big data

Here's an interview with IBM's VP of Emerging Technology, Rode Smith. In the interview, Smith pointed out the advantages of the IT department that accepts large data projects, and the risks faced by IT departments that reject large data projects.

The momentum of large data has been severely frustrated in many IT departments. Many people may have the idea that "it's enough now." We have plenty of SQL data storage and all the business intelligence tools we need for our business needs. So what are you talking about?

Yes, so far everyone has heard of Hadoop and the insights gained by digging semi-structured data. In many cases, however, this value orientation does not work for ordinary IT staff. This intransigence, I believe, is part of the reason that the share of IT spending is gradually skewed from the CIO to the chief marketing officer and the business unit manager.

I interviewed Rode Smith, an IBM emerging technology vice president, last week. So far, Smith has been engaged in large data practice and product development in IBM for four years. I asked Smith why the big numbers were strongly resisted in some cases.

Smith quotes a big data-Twitter project. In Twitter, a big customer with 150,000 business partners in the financial sector wants to monitor Twitter for information about new potential partners. For example, if a new restaurant sends a tweet saying they won an award, the company sends a congratulatory message to the restaurant and will use it as an opportunity to establish a customer relationship with the restaurant covering a loan or credit card business.

The response from the IT department is not enthusiastic, according to Smith. Smith has noted that it is not willing to consider this Twitter application as a real value application:

It is hard to believe that the IT staff will do something to get you access to Twitter data or Facebook data. Even if you do, these large data applications are short-lived. They may be useful for some particular activity that you are engaged in, or help with the introduction of a new product, but these applications may be discarded after use. So this leads to IT staff not being happy to say, well, we'll work with you because we know that the big data applications we create and your use will continue to create value. They will no longer touch or explore these large data applications in the same way.

It is not surprising, then, that many of IBM's big data projects originate from the business unit. "Business units or chief marketing officers don't want to wait 18 months to get their application research and development work into the to-do list. At the same time, IT staff are also trying to assess whether the application proposed by the business unit has real business value. ”

Last month, I attended the O ' Reilly strata meeting on Big data and saw a similar situation. Many customers are using large data tools to analyze the behavior of Web users to optimize the usability of Web applications, shorten the path of online transactions, or deliver appropriate Web ads at the right time. But these customers often outsource web applications that analyze and generate click Stream data, rather than resorting to internal IT departments.

The deployment of large data projects is a fast-changing goal, whether it is willing to accept it or not. I asked Smith how many of his clients had been through the exploratory phase, and that large data analysis was considered something they had to deploy. I was surprised by Smith's answer, which he estimated accounted for 70% of the percentage of clients. In the past 4-5 months, Smith said, when he called clients, "they would introduce me to their big data architect."

There is no doubt that the result came from the circles Smith had contacted. However, a recent study by Interxion, a British firm, has come to a stark reversal: only 7% of companies now see big data as a priority development project. Of the 750 respondents, 62% said the big Data project would be a priority development in the next three years.

Mr. Smith told me that both E-commerce and physical retailers are taking large data as a way of analyzing customers and eventually creating a recommendation engine. With this recommendation engine, businesses can immediately recommend products that are of interest to their customers in the store. The most exciting thing for Smith is that big data is expected to be applied to the health sector. In a recent research project launched by the Seton Heart Institute, large data analysis will analyze the electronic medical records of patients suffering from congestive heart failure and help doctors predict which patients are likely to be hospitalized again.

So far, I have experienced the bursting of many bubbles, such as the dotcom bubble. But I don't think big data will be one of them. Yes, in the development of large data analysis there will be some botched attempts, and these attempts may not create value. But with professional knowledge and software upgrading, it's clear that big data analysis will bring more and more rewards in marketing, as well as in the areas of health, manufacturing, finance, media, and any vertical areas you can think of.

Many IT departments will accept the pioneering and iterative nature of large data analysis, but some IT departments will see large data analysis as short-lived fashion and avoid them. I bet the latter will be abolished because of the declining importance of the company. (Grid Network Gundy compilation)

(Responsible editor: The good of the Legacy)