bigdata

Feeling buried under a mountain of data? In an age where every byte can be tracked and measured, the challenge isn’t about having the data or accessing it, but making sense of it all. We’ve got countless tools to store, filter, manage, review, sort, highlight, and analyze it but it’s still easy to get lost in the details.

Here are five big data analysis mistakes to avoid:

  1. Seeking What You Expect

A good scientist (and any BI consultant or marketer is a scientist, after all) tests a hypothesis but refuses to ignore data patterns that don’t support it. Be honest about your expectations and the lens of opinion that you see through to analyze your data. Don’t hesitate to get someone to play ‘devil’s advocate’ for you to get an opposing perspective. For example, if you’re pleased with the latest conversion rate on a new landing page, how did the traffic metrics affect your outcomes? Were the leads qualified?

  1. Getting Distracted

Okay, so we live in the Age of Distraction. But when it comes to data, getting distracted by events that aren’t directly related to your analysis goals can lead you far off stream. With a seemingly endless supply of data, be sure to set parameters of your search: such time and resources. Then only analyze the metric that will answer the question you are trying to answer.

  1. The Chicken or the Egg?

Just because two events occur concurrently or in proximity of each other, it doesn’t necessarily follow that one caused the other. For example, you might find a strong correlation between the performance of website traffic and revenue trends, but that doesn’t mean that website traffic is a direct cause of  revenue activity. There could easily be another factor that they both have in common, or another variable involved that increases the likelihood for revenue to improve website traffic is good.

  1. Half Baked Results

If you don’t have data sets that are large enough to extrapolate results or suggest a trend then you are wasting your time. Likewise, if you are comparing campaign performance, the difference in the results should be large enough to be significant for the comparison to be of value to you.

  1. Misinterpreting Intention

Actions speak louder than words, but intent is what drives them. It’s seductive to base conclusions on the actions recorded in your data rather than on the intention behind them. For example, if you’ve revised your messaging on a home page, and notice a spike in bounce rates, you could easily conclude that there was something wrong with the new message. But perhaps you did a better job of filtering out unqualified visitors and the qualified ones ended up converting. Be sure that your calls to action clearly speak to the intent of your audience.

Share this on