Finding Insights In a Data Haystack

Close-up of a SpreadsheetSo you’re actively engaged in this whole social media thing, and you’ve even figured out how you’re going to keep track of it. You’ve chosen a few tools that came highly recommended, you plugged in all the right information and now those tools are collecting data for you. All the time.

Now what?

With seemingly unlimited amounts of data coming at you in real-time, how do you make sense of it all? I’ve often been advised to look at it from the CMO’s perspective. What are the big bullet points that they would need to know? What are the insights?

Forgetting for a minute that very few of us actually have any first-hand experience knowing what a CMO wants, I wanted to walk you through my process for gleaning insights. After working in the media analysis business for two years now, I’ve found that often learning what to do with all that data can be just as tricky as finding the right data in the first place.

How to find insights in a data haystack

  • Organize the data. If you’ve just run a Twitter campaign for example, it would make the most sense to analyze tweets first and separately from other social media. If you’re looking at trends over time, it makes much more sense to arrange data by date.
  • Remember your goals. Before diving into the nitty gritty, reread your goals. Was your goal to increase positive chatter around your brand? Then focus on the sentiment of chatter more than the topic and define what you consider positive.
  • Read carefully. I know it may seem tedious, but read as much of the data as you can. Pay attention to common phrases that may be used to describe your brand or other subjects that often pop up in tandem. Watch for any pattern, even if it seems silly or meaningless.
  • Track patterns and examples. As you read, write down trends, patterns or keywords that you notice most often, and be sure to make note of specific, representative examples to refer to later. Write down anything that seems out of place or just plain interesting. This list will be whittled down.
  • Ask “So what?”. Go back through your list of patterns and trends and apply the “So what?” test. A lot of consumers used the word “blue” when discussing your brand, so what? If your logo or product is blue, this could be meaningful. Don’t just ask once, ask at least twice (Katie Paine recommends three times). Any trend worth mentioning can pass the “So what?” test twice.

What are your tips and tricks for finding insights in data haystacks? What methods work the best for you and which have failed?

[reus id=”6″][recent posts]