Give Your Traffic Sources the Credit they Deserve with Simple Attribution Enhancements
SwellPath is co-hosting a couple of local events (Web Analytics Wednesday and the new Portland Google Analytics User Group) coming up that I needed to prepare some material for. As I started researching potential topics for both, I kept coming back to the subject of attribution and attribution modeling. My first thought was that surely this was a played out topic, as analysts have been talking about moving beyond last-click attribution for years now. But the more I looked at it, the more I realized it is one of the fundamental concepts in web analytics that most marketers still do not know how to tackle. With this post, I’m going to try to give the average reader some quick ideas they can take away, that will hopefully help them understand their traffic sources a bit better, and give credit where credit is due.
First, a quick primer on attribution. When we talk generally about attribution in web analytics, we’re talking about how you attribute the conversions (revenue, leads, etc.) on your site to various traffic sources. You obviously want to know as much detail as you can about this, because you can use that information to decide how to allocate your marketing resources. For example: if you know that your email campaigns are involved in the decision making process for 50% of the visits that convert on your site, you probably want to allocate a very significant amount of your resources towards your email program.
How we handle crediting traffic sources, is referred to as attribution modeling, or an attribution model. The standard old-school way of doing this is referred to as last-click attribution. This means that the last traffic source that a visitor clicks through to our site from, gets 100% of the credit for a conversion. Let’s say I first visit your site by clicking through a bit.ly link in a Tweet of yours. Then I come back the next day through an organic search result. Later that afternoon, I click on one of your paid search ads, visit your site for the third time, and convert. Your paid search ad gets all the credit for the conversion, despite that fact that your Twitter efforts really provided the awareness, and the organic search helped me research your products. Shouldn’t they get some credit? With last-click attribution they don’t.
So why are we using last-click attribution? Because it used to be the only readily available method, and it was so much better than anything pre-web (i.e. traditional media attribution modeling) that marketers were fine settling with it. But times have changed, and analytics platforms now offer other attribution models. If you are not taking advantage of this (and most of you are not), it’s time to start.
To keep this post simple, but effective, I’m going to focus on simply using some of the Multi-Channel Funnels reporting in Google Analytics. These are baby steps towards a sophisticated attribution model, but they are giant leaps if you’re still just doing last-click attribution. In particular, we’re going to look at the Assisted Conversions report. To get to this report, you’re going to want to navigate to Conversions > Multi-Channel Funnels > Assisted Conversions in the Standard Reporting section of Google Analytics. This report is going to show us all of conversions that our traffic sources assisted, meaning they were a touch point for the visitor, but not the last touch point. It can also show us all the first interaction conversions (the traffic source was the first touch point for the visitor), and of course the last interaction also.
In the screenshot, I’m looking at a week’s worth of data for a ecommerce client. As you can see, looking at the email channel for last interaction, a total of $568 is attributed. But if we look at the assisted conversions, there is $2,746. All in, the conversions that can be attributed to email go from $568, up to $3,314, that is an increase of almost 500%. This supports the notion that although customers may not click through on an email they receive, and purchase during that visit, the visit keeps them engaged with the brand, and many of them eventually come back and purchase through other channels as a result.
This is just touching the tip of the iceberg with the Assisted Conversions report. You can also look at your First Interaction data (meaning conversions where the traffic source was the first touch point), and you can break down the source into more granular segments. For example, you can see the how your organic search visits, through non-branded keyword results, assist conversions. What all of this allows for, is a greater understanding of how your customers move through the purchase decision-making process, and what sources play fundamental roles at the different phases of that funnel. We can back up our assumption that email isn’t the first touch point (obviously), and isn’t often the final touch point before a conversion, but it is often “in the middle”. Meaning, it is key to the research phase of the customers’ decision-making process. Knowing this, we can adjust the messaging and content in our email campaigns, to better nurture customers at this point.
Hopefully this has been helpful, and at least can get you started with moving beyond last-click attribution models, into something that gives more appropriate credit to your traffic sources. Utilize Google Analytics’ help section to navigate the various features and details of this report, and as you explore the other Multi-Channel Funnel reporting.