Data Smog: The Too-Much-Data Problem

I was in a meeting the other day with one of the Web analysts on my team. He threw out the term “data smog” to describe the massive amount of data a new client was regularly reviewing. They were struggling to pull insight from the data because all their time was devoted to just putting all their different data types together.

He later provided a definition he found online: “To be overwhelmed with too much information. Example: A Google search that turns up six million references is data smog.”

In no other industry is data smog a greater issue than in Web analytics and the related site data that goes with it. Our clients commonly look at the following data types to understand site performance:

  • Web analytics data (Omniture, WebSideStory, WebTrends, etc.)

  • Attitudinal data from surveys, user studies, and the like
  • Customer satisfaction data (ForeSee, etc.)
  • Competitive insight (comScore, Hitwise, etc.)
  • Customer segment information
  • Industry research reports (JupiterResearch, Forrester, etc.)
  • Customer data
  • Transaction data
  • Campaign data
  • Affiliate information
  • Call center data

These are only a sampling of the data types you may get from your site. Usually, these data types aren’t combined to tell a full story. There could be 11 different owners for the 11 items who rarely speak or work with one another.

Each data source can easily create data overload; combining them can be flat out overwhelming. But combining this information properly can result in a ton of insight and help you understand the overall visitor experience.

How do you avoid Web site data smog? A few tips to keep you in the clear:

  • Focus only on the most important behaviors or metrics. Determine what behaviors drive your business and dig deep into those, rather than taking a superficial, high-level view at everything. Too often people try to look at every piece of data they have and get bogged down. That leaves no time, patience, or resources to dig into the things that really matter. Focus on the top business and site goals to prioritize where you’ll spend your time.

  • Start small. To begin, dig into one issue to determine the best ways to use the data and solve the problem. Once you’ve identified an opportunity and forecasted the results, move on to A/B or multivariate testing (e.g., Offermatica).
  • Focus on actionable data. There may be a bunch of “good to know” things you could determine by analyzing the data. But you must focus your time on analyzing the information you can ultimately act on.
  • Tell a story. Make sure you tell a story with the data you’re looking at, don’t just regurgitate all the numbers.

As analytics tools evolve and more people try to combine behavioral data with attitudinal and competitive data, data smog will only become a bigger problem. Don’t get caught up in spending all your time pulling and combining data without conducting the analysis that provides the insight that really drives your business.

Related reading

site search hp