Using Cross-tab Analysis to Explore Basic Marketing Attribution in Analysis Workspace

There are many ways you can take your attribution methodology to the next level with Adobe Analytics. In this video, we highlight how you can derive deeper insights from the Marketing Channels report using cross-tab analysis in Workspace.

Hi this is Jen Lasser with Adobe Analytics Product Management. In this video, I’m going to - show you how you can take your marketing attribution step further with marketing channels reporting, in cross-tab analysis - and analysis workspace. Now, for the demo, I’m going to assume you have a marketing - channel’s report enabled, but if you don’t, you can read more about why it is our most effective - external traffic report at the link in the description - of the video below. The help documentation link - will outline the process for quickly enabling the reports. It’s something that can - be done within minutes, you just need to put in a little bit of forethought - and preparation, about your tracking strategy and your tracking code structure.
Okay, back to our demo. So the marketing channel’s - report, once set up, will automatically create a first and last touch channel view for you. Well, most of the time we - defer to a last touch view. There is some helpful - information that can be gleaned from the first touch report as well. A few questions that - come to mind for me are, you know, which channels are assisting and driving conversions? How often are my customers interacting with multiple marketing touch points? Or how often is a channel - that found a customer the same one to convert - that customer in the end. So while we can’t do full - blown attribution modeling just yet in analysis workspace, we can use what we know about - first and last touch channels of our visitors to kind - of take our analysis a step further past this last touch view. So I’m in workspace here - and I’ve pulled together a first touch channel report, and a last touch channel report. Now, the first touch channel report is telling us what channels - found our customers. What are those original channels that brought those - customers to our website? The last touch report is telling us what channels were were most close to when the customer converted. What are those last touch points - that they interacted with, before they converted on our property? So while both of these views are certainly extremely helpful, it doesn’t help us in seeing the overlap between our first touch and last touch marketing touchpoints. We really need to bring together - both of these dimensions to get that kind of - information on the relationship between first and last.
So, this is where cross-tab - analysis comes in handy. cross-tab analysis is the - ability to bring in dimensions, both as rows and columns of a report.
So I’m going to start by bringing - in the last touch channel like we had in the original table and I’m also going to bring in my key success metric of orders. Now I’m going to use orders - because success metrics are only credited to the - last channel that was present for that customer before they converted. If we used a metric - like visitors or visits, it could happen where a - visitor would be credited to many last touch channels, because they could have - come in through email and then returned to the - website a couple of days later through paid search. So to get a very clean - view on your attribution, you want to stick to - your main success metric.
So the next thing we need to do, is pull in our first touch channel and add our cross dimension here.
So I’m going to pull over the variable and drop it right under orders. And by default, it’ll bring - in just the top five channels. But you can see very quickly - that I’ve sliced my order data by new dimensions. This is the beauty of cross-tab analysis, rather than having to create segments for each one of these - first touch channels, I was simply able to - drag over a dimension, and slice my data by it. So a huge time saver.
Since only the top five - channels are brought over, what I need to do now, is - open up first touch channel, and bring over the other two - channels that are missing here. So Social Networks and SEO will round out my first - touch channel report. Okay. So the next thing I want to do, is make sure my columns are - in the same order as my rows. And it looks like I’m pretty close here, I just need to move social - networks ahead of paid search.
Okay, great. So the last thing that I like to do, is actually simplify - the table a little bit. I like to just look at the - percentages as a starting point. So to do that, you can change - your column at the top here, and turn off the number. That will just leave the - percentages in the column, which gives you a much clearer view here. So you can go column by column - and turn off the numbers. I went ahead and actually - did this ahead of the demo just to save us a little bit of time. So you can see here, we - now have the same table as before for last touch channels and the percentage of orders for each. So this resulting view - is quite a powerful one. So let’s break it down a little bit here.
Along the diagonal, we’re able to see the percentage of orders that were credited to the same - first and last touch channel. For example, for email, 72% of orders that were credited to email as a finder channel, we’re - also credited to email as a closer channel.
Higher percentages along the diagonal, can possibly indicate a lower overlap in your marketing efforts. Meaning that, if a channel finds an order, it’s likely to also close that order. Your marking touchpoints - aren’t interacting with one another very much. Lower percentages along - the diagonal can indicate that you have channels that - are assisting quite a bit in driving conversions - for other closer channels.
The next thing you can do, is - focus on individual columns. Columns will sum to a hundred percent. So this percentage is of the total orders that email was a finder for, you know, what percentage of those orders were ultimately credited to each of these last touch channels.
So you can see here, of all the customers that email found, 72% of their orders were ultimately credited - to email in the end, indicating a lower likelihood of assisting channels in between.
The remaining 28% of orders - were ultimately stolen away by other marketing touch points, and credited elsewhere in the - end, in a last touch model. And this indicates that email assisted in driving those orders at some point.
So, in this case example - here you can see that, referring domains benefited the most from an email assist here. So, of all of the original - customers email found, referring domain stole away - about 10% of the orders that they had found - ultimately, did on the website. Okay? So, I prefer to look at the percentages 'cause I really just want to - understand the relationship between first and last touch. But it is also helpful to - look at the numbers as well, and you always want to ground the information you’re looking at in terms of volume of data as well. So, I created the exact same table. I just toggled the percentages - off and the numbers on. I also added in all visits segment here, just so I could get the - total count of orders attributed to each last touch channel.
In a stable, you can also see - the total number of orders along the top, credited to - each first touch channel. So that’s kind of helpful to see as well. So for our email example, there was 8,857 orders that email, you know, found customers for, and ultimately email got credit in the end for about 9,300 orders. So, pretty interesting to see.
You can highlight each row at a time here, to understand, you know, of - all the last touch orders, how many were assisted, - how many were unassisted. So for email 9,300 last touch orders of those 6,300 were likely unassisted because the first touch - channel was also emailed, the remaining 3000 thousand - orders did have an assist though from at least one other channel. Because you can see, you - know, referring domains originally found this thousand orders that ultimately went to email.
So, you know, while this - approach doesn’t give you a full multi-touch view, it is a great way to start diving into your marketing attribution, past the traditional - level of last touch only. And, you know, once you have the marketing - channels report set up, you can get to this data immediately, and it’s automatically created, for you by way of capturing - first and last touch. So we at Adobe are continuing to improve the attribution capabilities - that we’re offering within analysis workspace, but in the meantime, know that as an Adobe Analytics customer, you do have options like - this to dive further into your marketing analysis, than you might have - originally thought you had. So I hope you found this video on bringing together marketing channels with cross-tabs analysis helpful, and we’ll see you next time. -

UPDATE: There have been several improvements made to Workspace since this video was published. We are leaving the video live because it has some great tips that you should know anyway.

First, be sure to check out the Attribution IQ features that will help take your marketing attribution analysis even further. This includes the following videos and the videos around them:

Second, if you are following the steps in this video, be aware that you can use the Freeform Table Builder to setup your table.

  • 3:20 - Cross-tab dimensions are automatically dynamic when dropped, meaning their values can change over time. For this analysis, it’s best to bring over static columns instead. Learn more about Dynamic Columns at 6:33 of the video: Row and Column Settings in Freeform Tables
  • 4:30 - Column settings can be updated in bulk now, instead of 1-by-1. Learn more at 8:45 of the video: Row and Column Settings in Freeform Tables

For more information on this topic, please visit the documentation.