Data 101: Avoid drowning in the analytics data lake

Now more than ever, businesses are focusing on data collection and application in order to stay ahead in this competitive market. Maybe your company is looking to run an analysis on the costs of maintaining your physical data center. Maybe there’s an ongoing project over in marketing and they want to see the trends of their latest strategy. Or maybe someone on the C-suite just wants IT to start collecting data for the sake of collecting data.

And to better analyze that data, the C-suite expects IT to solve their analytics woes. What they want to analyze is called the “data lake,” a storage point where all the data collected from company websites, blogs, social media, email newsletters and event attendance gets dumped.

Now, if the data in the lake is properly stored and maintained, it can be a wealth of information, useful in making critical company decisions. That’s the magical ‘thrust a sword out to the man who will be king’ sort of lake execs are looking for.

But if it’s stored improperly or made without a management plan in place first, the lake can be daunting, overwhelming and, worst of all, completely useless.

What often ends up happening is, due to some sense of urgency, companies are tempted to pull the trigger early on many of their data projects and start collection right away. But that’s a risky move that might end up doing more harm than good.

Manage your data streams

Collecting data without any sort of plan in place isn’t the best of ideas. It’s like trying to build a dam to create an actual lake without first considering where all the pent back water will go and where it might flood elsewhere. It has disaster movie written all over it.

What usually ends up happening when companies ask IT to start collecting data on everything, is it turns into a digital disaster. It’s the equivalent of turning fire hoses on your face and putting them on full blast. Instead of full force blasts, you want a steady and easily maintained trickle.

So what can you do to avoid getting drenched?

Work smarter, not harder

For starters, always keep in mind that data alone does not equal intelligence. Data is merely a means to make more informed decisions.

Data also is so much more than just IT’s game. When the data is going to be used to make financial decisions that impact the company, or is collected from specific departments, people from those departments are going to be involved in some way. Those people need to be identified and their contact information kept on hand to feed them updates on the data process, if periodic meetings are out of the question.

IT needs to understand the broader implications of trends within the numerical data, and that requires creating a system to balance out those torrents into a trickling stream.

Getting the most out of your data

Before data collection even begins, IT needs to let those pushing for analytics know that a baseline must be established and that none of these changes are going to happen overnight. After all, how do you know what’s out of the ordinary or whether decisions have had a positive impact if you don’t have a starting point to look back from?

To establish the best baseline possible it’s going to take at least 90 days, with three data point “captures” of your network and performance systems – the first at day 30, the second at day 60 and the third at day 90. These will serve as snapshots of a typical performance that can be measured later.

For this step, you don’t need to buy any products. In fact, you shouldn’t be shopping for analytics programs until you know what you’ll need. Definitely don’t trust any product that promises to solve all of your analytics problems or bills itself as “the one program to rule them all.” That one is probably destined for a different sort of lake – a molten one.

In the meantime, you can use Excel to track numbers and pick out any trends. Make sure it’s tidy and presentable, taking advantage of Excel’s table feature where each observation set is its own table with columns and rows of individual data points.

Hammer out the details

While you’re establishing a baseline foundation for your data, IT should be getting together with the C-suite and planning what data will be examined and why. You’ll want a clear end-goal in mind for this data project, otherwise you’re just casting sail in a sea of never-ending data. Make sure to hit the following points:

  • work out the details before you take the day 90 network snapshot
  • cover who is going to receive the data, who will analyze it, and who will receive summary reports on the analyzed data, and
  • consider how the present the results from the collected data, keeping in mind the target audience – you may need to repackage the same analysis two different ways for two different crowds.

Make Smarter Tech Decisions

Get the latest IT news, trends, and insights - delivered weekly.

Privacy Policy