It used to be that businesses could never learn enough about their customers or about business process shortcomings. Now, businesses can learn virtually anything they want, and the harder challenge is deciding what parts are important. In an article for InformationWeek, Rosemary Radich shares four steps to distinguish important data from fluff and use it to transform the company:
- Identify your end goal.
- Identify the good, the irrelevant, and the unusable.
- Data is only as good as the people interpreting it.
- Speak in visuals.
Order from Chaos
Before you go trying to transform anything, you need to have an end goal to your analytics in mind. Executives must be on the same page about what is important to the business, and the explicit aim of data collection should be to yield insights that improve that area of operations.
As for deciding which data is really viable, Radich says this:
Not all data is created equally and knowing the source of your data is critical to the success of any resulting analysis or actions. The sheer quantity and accessibility of data is important, but data quality should always be the primary consideration. Working with skilled data scientists to sift through the sea of information and pinpoint the most accurate and relevant data is crucial to ensuring your analysis stays on task to achieve your goal.
Unsurprisingly, making good use of analytics ultimately depends upon the people collecting the data. Just like weather forecasters make intuitive sense of data that would look like gibberish to other people, data scientists need to be able to intuitively identify patterns in data models. Hire for a diversity of experiences and perspectives in this regard.
When going to present the findings of analytics, use charts and graphs as punctuation marks on the major insights you want to deliver. Have the visuals tell a story that will make the next course of action clear—even though the work that went into developing that insight was anything but clear.
You can view the original article here: http://www.informationweek.com/big-data/big-data-analytics/weathering-the-data-storm—4-steps-to-analytics-success/a/d-id/1329236