Using Alarmduck for Product Spikes
Recently, I have been sharing cool ways that you can use Alarmduck to monitor and improve your web analytics capabilities. For those unfamiliar with Alarmduck, this post will explain what it is and how it can be used with Adobe Analytics to find data anomalies. Last time, I shared how I use Alarmduck to find “hot” onsite search terms. In this post, I am going to use a similar methodology to look for website products that have data anomalies.
Product Anomalies
Most (not all) website sell products or at least display products that can be sold offline. As a web analyst, it is common to capture a product name or ID in a dimension (in Adobe Analytics this is done in the Products variable) and then use that dimension with metrics. Common metrics used with products include Product Views, Cart Additions, Orders, Revenue, etc. A typical products report may look like this:
However, as was the case with internal search terms, you probably don’t have time to monitor this report every day to see if there are certain products that have spiked or dipped when it comes to product views, orders or revenue. Instead, you’d probably just like to be notified when a data anomaly for any product occurs when it pertains to product-related metrics.
This is where Alarmduck really shines. Using the steps similar to those in this internal search term post, you can create a data anomaly report in Alarmduck to send you a message in Slack if a data anomaly occurs. For example, let’s imagine that you work for Super Company and sell products for online purchase. One day, you come into work and as you open Slack, you see an alert in the Alarmduck channel like this:
As you can see, Alarmduck has reviewed your Adobe Analytics data and found that product #847284937 had 153 Orders yesterday, which is far more than have taken place in the recent past. From here, you can link to the report in Adobe Analytics and then begin the process of discovering why this anomaly took place. Perhaps there was a specific external campaign that led to the increase, or a new onsite promotion that drove more visitors to this particular product.
While the preceding example uses Orders, the Alarmduck anomaly report can be setup to look for multiple product-related metrics. Perhaps you want to see if there is a spike in product reviews or product comparisons. Regardless of your need, Alarmduck has the machine learning power to mine your data on your behalf and highlight the most important changes that you need to be aware of on a daily basis.
