3招帮你争辩你的大数据

haosunzhe 2015-05-16

展开全文

3 Steps to Help You Wrangle Your Big Data

Here's a sobering thought: 65 percent C-Suite executives surveyed concede they risk becoming irrelevant or uncompetitive unless they embrace big data. At least that's the conclusion of a study by CapGemini and EMC.
Pratibha Vuppuluri – founder and CEO of on-demand research and analysis firm, KeyInsite – agrees.
“We’ve been seeing some of our clients waking up and talking about making sense of the data they’ve collected,” she said. “A large part of the wave we’re noticing regarding analytics trends is that those companies with agile, lean, nimble platforms have higher ROI and conversion.”
To give some guidance to executives who understand the urgency – and opportunity – that big data presents, but are still struggling to dip their toes in, Vuppuluri offered three recommendations about how to take the next step.
Take Small Steps1. Understand the data triplicate Big data has scary connotations. For those who don’t step back to take in the bigger picture, the sheer amount of data a business has to deal with can be overwhelming.
However, Vuppuluri has a different view:
“Don’t harp on volumes of data. The biggest issue is to identify and develop models to help you analyze the data. Big data has always been there. It’s nothing new – it’s stored, and people try to make sense of it.”

One way to make sense of it all is to build a model to help you answer your big questions – something that requires a particular skill set that businesses haven’t yet figured out how to source, she said.
The skill set?

A data scientist to help with programming, and building out a statistical model
A data analyst to interpret the data
A domain expert who understands the data trends and patterns of your business and its competitors to help pinpoint your analysis

“It’s an elusive profile that people are having an issue finding,” said Vuppuluri. “Someone who is a modeler isn’t a domain expert, and vice versa.”
The solution, she added, is not to hire just one of these professionals, but a team that incorporates all three – a fitting recommendation for the CEO of a company which provides just that.
“Data science has become so pervasive,” she said. “Using it should make sense. Usually it’s a team that’s needed to make it happen.”
2. Create a data-driven cultureIn order for your employees to understand the importance of making decisions based on data, Vuppuluri suggested fostering a data-driven culture by getting employees in the habit of regularly reviewing some simple metrics.
“Create metrics and track your data on a daily basis,” she said. “If you do this, you can find out, for example, why certain divisions in your business are performing lower than average and then take action to deal with it."
She added that regularly reviewing as few as two to three metrics “creates a sense of discipline and insight that does not require an outside analyst.”
For example, if a company is tracking social media performance, such as number of followers or impressions, people will want to open the dashboard to find out progress. It’s just human nature, said Vuppuluri.
“It’s a human element,” she said. “We all strive toward finding out which grade we get. If I’ve performed better than yesterday, it’s a positive thing. If not it gives me a sense of wanting to do something.”
3. Start with the data you already haveVuppuluri noted that when people think about data sources, they tend to consider only outside data sources. A better approach, she advised, is following a process that instills cost discipline and provides a sustainable approach to analysis, such as the following:

Identify your existing data sources
Investigate how clean the data is
Find out how much of the question you can answer with your current sources

Once you’ve followed this process, only then should you identify sources that can append that data, if necessary, she said.

What you have in-house is a wealth of information. Leverage it. Squeeze the data, identify whitespace and see what you’re getting. Then, build a model to fill in that white space. If you don’t do this, your cost structure increases, and you won’t have the resources you need to achieve your goals."