5 keys to data success: Management

Mik Data Science March 11, 2016

Like any asset, data needs proper management to reap the benefits.
Like any asset, data needs proper management to reap the benefits.

Taming data starts with management.

This is the second article in our 5-part series “5 Keys to Data Success.” Last time, we talked about getting off on the right foot by focusing on good data collection. Garbage in, garbage out (forgive the cliché).

In this article, we’ll tickle your brain with 5 questions to ask yourself on the next topic: data management. Data management means everything from storing data to how you structure internal policies on who has access. It is also an area that can get very expensive very quickly. Here’s our 5 questions to be sure you ask about your data management strategy:

1. How much data do I expect to collect?

The answer varies a lot. But being realistic about the scope of your initiative and the quantity of data is critical. Do you need warehousing, Hadoop, or just some in-house servers? Remember, Big Data is about more than the size on a disk.

2. What’s my data lifecycle look like?

How long do I plan to store my data? Indefinitely? Only the last 6 months? You may need to anticipate significant upgrades and maintenance on the storage if you plan to keep data around for a long time. This will help you answer #1, as well.

3. Who needs to use this data?

Data should be treated like any valuable asset. Consider who should really be allowed to see/edit/move data based on your own policies and roles. Open access sounds nice, but can lead to costly mistakes — especially with sensitive data.

4. What are my security requirements?

Depending on industry, you may be required to meet certain levels of security. Even if you don’t, data is a critical asset that is both monetizable and an open door into your business strategy. Did you consider how you’ll keep prying eyes away?

5. What budget do I have for data management?

Data management can be one of the most expensive parts of the whole game. From storage, to maintenance, to administration don’t over look this portion of the bill. You’ll notice we asked this about data collection as well. Money matters.

Data management is a huge topic in and of itself, and many things are missing here — such as considering if high-speed, frequent access is needed or if access policies should change depending age of the data. These 5 questions, and the others in the series, will hopefully lead you down the road of asking critical questions early and often.