Tip Sheets & eBooks

Catch Data Errors Before They Happen

Issue link: https://resources.observepoint.com/i/1267814

Contents of this Issue

Navigation

Page 1 of 6

1 Would you make business decisions based on the advice of a psychic? Hopefully not. A reasonable professional would recognize that basing business decisions on Madam Perdita's predictions won't get you far. But here's the clincher: You, the reputable, data-driven professional, are as vulnerable to misguided advice as a psychic's patron if you don't have mechanisms and process- es in place to validate the accuracy of the data guiding your business decisions. Whether you're an analyst, marketer, deci- sion-maker, or all of the above, the accu- racy of your implementation, and the data collected from that implementation, has a significant influence on your ability to make effective decisions—and therefore your credibility. Don't lose that credibility, and don't miss out on business opportunities because of faulty data. Here's how you can catch analytics errors before they happen and become your company's data-driven champion. The Reactive Approach For Catching Data Errors Let me tell you a story that might sound familiar. Following what appeared to be a successful website release, a marketing team met to re- view their results. The analyst pulled up the re- ports to show his boss what kind of performance increase they were seeing but, instead, saw that their top metrics had dropped noticeably. How could this be? The analyst opens a key metric report only to find it's completely empty. For some reason, the implementation hasn't been collecting this data point since the launch. With considerable effort, the analyst manually combs through the analyt- ics implementation to find the problem: conflict- ing JavaScript in the new release compromised the analytics implementation. The team has lost critical data for accurate analysis of the new site's performance, and the analyst lost credibility with his supervisor. On top of that, if he hadn't noticed the error (not all data errors are obvious), the decision-maker could have made a bad call on bad information. Many companies are still testing their analytics implementations after a release (or not at all) and unsurprisingly face similar scenarios—dis- covering they've been collecting inaccurate data that is unusable. Then they lose even more time, data, and confidence trying to figure out what happened. The reactive approach is much less than ideal—it's crippling and costly.

Articles in this issue

Archives of this issue

view archives of Tip Sheets & eBooks - Catch Data Errors Before They Happen