This video walks through Contribution Analysis and its ability to use machine learning to quickly and easily explain the factors that contributed to a trend in your data.
Hi, this is Trevor Paulsen from the Adobe Analytics product team. Analysis workspace for Adobe Analytics has quickly become my favorite tool for doing analysis, and I’m excited to show you a powerful machine learning tool that we’ve brought into Analysis Workspace called Contribution Analysis. One of my favorite new aspects of Analysis Workspace is that it now automatically surfaces statistical anomalies in any trended report that you pull. In my case, I’ve got a revenue report here, and I’ve noticed that I had an anomaly on August 1st. You can see it here in the table, too. I wanna know why I had this anomaly. What happened? Who is responding to what campaign, or referral, perhaps something went viral, what are the specific factors that contributed to this anomaly, and most importantly, how can I use that and repeat that in the future, or, if it was a dip, how can I avoid that in the future? So, here’s how you run it. By simply hovering over this anomaly icon I can click analyze. Or, in the line chart, we’ve also added the ability to hit analyze. When I click that analyze button, it opens up a brand new contribution analysis panel. In the panel, I can exclude dimensions that I don’t want Contribution Analysis to consider, that are not particularly relevant to me. For example, persistent cookie support. That’s useful sometimes, but not particularly interesting for this analysis, so I’ll exclude it. When I am ready, I’m gonna hit Run Contribution Analysis, and when I do that, it kicks off an extremely optimized and powerful machine learning algorithm that combs through hundreds of millions of data points every second, trying to identify what are the statistical contributing factors to this anomaly.
When contribution analysis completes, I get a great readout of all of the contributing factors to that anomaly. You can see here in the Top Items table, I got the top five contributing factors based on their contribution score. A higher contribution score closer to one means it was more likely to be a contributing factor to that anomaly. You can see that my top item here was the product details exit page Sheeprock Gloves. In addition, the second most contributing item here was my Men entry page in my blog, so, very, very interesting information for me to dive in further down the road. I can also notice that I’ve got these contribution segments that the panel has created for me. When I look at the contribution segments, you notice they’re based off of those contributing factors. But the greatest thing about this in Analysis Workspace is the flexibility that it gives me to do deeper analysis. So, for example, I may be interested in knowing a break down of the browser type for each of these contribution segments. It’s as simple as dragging that item over and breaking it down. Contribution analysis now in Analysis Workspace gives me such additional flexibility to deep dive, answer more questions, and really get to the bottom of what is driving my business, and that’s Contribution Analysis in Analysis Workspace.
Contribution Analysis is available in Analysis Workspace for customers of Adobe Analytics Premium.
For more information on this feature, visit the documentation.