(Time-)Normalize Performance over time in Adobe Analytics’s Analysis Workspace
In Digital Analytics, one of the most common requests from business stakeholders is to compare the performance of two or more items on our websites, like marketing campaigns or content pages. While it is immediately obvious why this comparison is important to the business, it quite often leads to graphs like this, where the analyst tries to visualize performance over time:
This solution is technically correct but makes it hard to really compare how both pages perform in direct comparison with each other. They went public on different dates and while Page A is rather stable in regards to traffic, Page B got a boost at around the middle of its time online. So, how do we make this simpler?
When enjoying my free time between jobs, I caught up on some older videos from the Superweek Analytics Summit’s Youtube Channel. In 2019, Tim Wilson demonstrated how to align dates and display cumulative values (I highly recommend you watch the video) using R. I really like how the cognitive load can be reduced by a lot by aligning each line to the left of the chart and using a cumulative sum. Luckily, it is really easy to do this in Analysis Workspace as well:
That looks much clearer! Now we can see how Page B outperforms Page A right from the start, and even increases around the middle of the displayed date range. Awesome, let’s build this view together!
Align dates using custom Date Ranges
The first step to this aligned view is to familiarize ourselves with the data. To do this, we start with a simple Freeform Table containing the trended performance for our two pages, identifying when they went online:
Here we can clearly see when the two pages became relevant to the public. For Page A, that date is Nov 16, while it is Dec 1 for Page B. Now we can create a custom Date Range to cover the following three weeks (or however long you want) by clicking the plus-icon on the component rail:
In the Date Range Builder, we can then select the range of dates we want to use for each column. For Page A, I simply select three weeks from Nov 16 like this:
Once we have created both dare ranges, we can just drag them over our Freeform Table’s columns as filters like this:
Here we can see Analysis Workspace doing its magic. Now both columns have their very own date range and start at the same row, showing the respective date in each cell of the table. At the same time, the date column is basically useless now, so we should only look at the row number instead of the date. Next, we drag a Line Chart into our project to visualize both pages. To reduce the confusion about the date, we can just deactivate the X axis in the chart settings:
Now our chart looks quite nice already:
That is the first of the two charts done in no time! This aligned view can help our business users tremendously when they want to compare different campaigns, pages, or products. And since it was really easy and fast to do, I don’t have any doubts we could teach them how to do this themselves. Now let’s go one step further!
Cumulate performance over time
The other part that Tim mentioned in his talk is how cumulative views help users understand the total impact of a piece of content for the site in total. Luckily, Adobe Analytics can do this very easily with a simple Calculated Metric. To start, let’s create a new Metric, again by clicking the plus icon:
In the Metric Builder, we search for the Cumulative function and drag it onto the canvas:
This gives us two slots to fill. In the “number” field, we just put a zero, so Analytics sums up all previous rows. In the “metric” slot, we can drop our Page Views Metric:
Go ahead and save this metric. Now we can just add this new Metric to our table, using the same base filters as with the Page Views metric:
As the last step, we add another line chart with the same settings as before. If we lock the selected data range for both charts, we can have both charts based on the same table. I think it looks quite nice as a dashboard:
Conclusion
That was quick and easy! There are quite a few applications I can think of for this kind of analysis:
- Compare the performance of pages that have gone public on different days
- Align different marketing campaigns to a common start date retroactively
- Analyze how two videos contribute to the cumulative performance along their lifecycle
I hope you enjoy this kind of quick tips on Analysis Workspace, even if there are no super-complicated calculated metrics involved. As always, let me know what you think about this tip! Have a nice day.
German Analyst and Data Scientist working in and writing about (Web) Analytics and Online Marketing Tech.