Over recent years, the average size and complexity of web analytics implementations has absolutely exploded. It is no longer uncommon to see organizations tracking over 100 different success events. The funny thing about events (everything from page views to orders to Facebook shares) is that they tend to fluctuate. Change any report in Adobe Analytics to a long date range and a small granularity and you’ll see what I mean. Maybe it’s the day of week, maybe it’s seasonal, maybe it’s just typical noise, maybe some segment of users has suddenly shifted, or maybe something is broken?
With so many potential variables, it can be very hard to look at a chart and just know when one of those normal spikes or valleys is something that should require further attention. This is not to say a lot of people don’t *think* they can accurately eyeball data by quickly looking at it. But think about how many false alarms you have seen raised due to swings that turned out to be just another day. Or perhaps swings that looked ordinary and actually turned out to be indicative of much bigger issues.
So when Adobe rolled out Anomaly Detection it was a huge deal. And I certainly don’t know how often it was used across the board, but from my experience I’ve seen it used way less than I would have expected. It seemed to become more of a tool to be used once there were already suspicions about how certain metrics were trending. For most users, I’m guessing that it didn’t become as much a part of the daily workflow as it should have. The problem was that it was a separate report, in a place you had to intentionally find. Useful when you needed it, but otherwise just another report to monitor.
But those days are behind us. With the October release fresh out of the gates, Adobe has managed to seamlessly integrate anomaly detection into all trended charts in Analysis Workspace. Whenever you’re looking at an event, anomaly detection will be enabled by default and allow you to quickly see inflection points in the data you might not have seen otherwise.
No matter what other analysis you might be working on, Adobe will constantly be showing you what fluctuations were statistically significant. Said another way, without even changing your workflow, you now have that superpower to eyeball any trend line with scientific precision. You can spend more time finding insights instead of chasing shadows in the noise.
[author] Jon Narong [author_image timthumb=’on’]http://i2.wp.com/33sticks.com/wp-content/uploads/2016/05/630630c74efb4c3e5c2d6df74cd0322c_400x400.png?w=192 alt=”Jon Narong”[/author_image] [author_info]Jon is a Principal Analyst at 33 Sticks where he leverages his more than 10 years of experience in the digital analytics space in both technical and business focused analytical roles. Jon joined 33 Sticks from lynda.com, recently acquired by LinkedIn, where he ran the digital analytics and optimization practice. Prior to joining lynda.com, Jon lead the analytics platform migration for the POPSUGAR media network. As an industry veteran, Jon’s tenure includes managing e-commerce analytics and optimization for Apple, Disney, and Beachbody.[/author_info] [/author]