In the last post we introduced attribution and talked about a couple of the standard attribution models used within Google Analytics.

In this follow up post, we’ll talk about determining the level of effort we should put into looking at attribution models as well as an easy way to understand how marketing channels are impacting the final conversion.

Should We Obsess About Attribution?

First, unless all (100%) of our conversions happen within the first session, it’s critical to understand attribution. However, what is the line? When should we just use simple techniques and when do we need to really dig in and spend a lot of resources to figure it out?

Go to the conversions section in Google Analytics and click on multi-channel funnel → path length. Are the majority of conversions happening within the first session? 80%? 90%? If so, don’t obsess over attribution. Note: If you have multiple conversions (goals) set up, be sure to check each separately.

But, even if the overwhelming majority of our conversions are happening during the first session, does that mean we should ignore attribution? Absolutely not.

If we still need to understand the impact of our marketing channels, but either don’t need to or don’t have the resources to really obsess over attribution, we can use assisted conversions.

Assisted Conversions Explained

The label ‘assisted conversion’ makes perfect sense. What are those channels that assist the final interaction as opposed to being the final interaction before conversion?

Looking at assists is a good way to quickly see those channels that may not be the final interaction before the conversion, but contributed to conversion nonetheless. Let’s look at our holiday shopping example from the previous attribution post (Part One) and see which of those touchpoints would be seen as ‘assists’.

Here are the touchpoints along Jack’s journey to conversion (purchase):

• Jack clicks on a banner ad and goes to the website, but doesn’t convert → first touch: banner ad
• Jack searches Google for the company name, clicks on a paid search listing, goes to the website, but doesn’t convert → second touch: paid search
• Jack types in the URL directly, goes to the website and purchases a product → third (and last) touch: direct

In the above example, the direct channel gets credit for the final conversion while the banner ad and paid search will each get an assisted conversion.

“But,” you may say, “I thought the default attribution model for Google Analytics was last non-direct click. In this scenario, wouldn’t Google Analytics give credit to paid search for the final conversion?”

Well, yes and no. The Multi-Channel Funnel reports in Google Analytics don’t use last non-direct click as the default model. Only the Non-Multi-Channel reports do. That’s weird and confusing, I know. But, we move on…

By looking at the assisted conversion report, we can quickly see which channels are contributing as the final conversion and which are contributing by assisting the final conversion.

To quickly see which channels are more ‘assist’ channels, look at the last column – assisted/direct conversions. In this column, if the number is less than one, it’s more of a final conversion channel. On the other hand, if the number is more than 1, the channel assists in conversions more. If the number is right around one, the channel equally assists and drives the conversion. Google has a great explanation on their support site.

Lookback Windows

Within most of the multi-channel funnel and attribution reports, you’ll notice ‘lookback window’ at the top of the report. It defaults to 30 days. That means that when we’re looking at the report, we’re seeing touchpoints that happened in the 30 days prior to conversion. So, if a user converted after four sessions to the website and the sessions happened – 35 days ago, 20 days ago, 10 days ago, and 1 day ago (when the user converted), if we look at a 30 day lookback window, that first touchpoint (that happened 35 days ago) would not count when we look at attribution.

Why is this important?

If your buying cycle is quite long, a longer lookback window may make sense. Otherwise sessions that happened 30 days prior to conversion will not count when looking at attribution.

Some Caveats to be Aware of – But Not to Get Hung Up On

As we talked about in Part One, even if our entire buying cycle is long, we may not be able to measure that entire cycle – to the ultimate conversion (application, enrollment, donation, etc.) – on our website. So, for what we can measure, our ‘buying cycle’ might actually be short.

There are also offline touchpoints that happen (phone calls, SMS, campus visits, etc.). These are hopefully captured in a CRM (but probably not entirely tied to the what is happening online).

Further, you have repeat users that are surely being tracked as new. Why? Because people use different devices and browsers. Unless you have a login on your website and have Google Analytics setup to track User IDs, you’re definitely tracking repeat users as new users.

But, and this is important – don’t get stuck on these limitations. It’s simply something to be aware of. It’s not perfect at all, but it never will be. Even the most sophisticated attribution modeling and technical setup has limitations. But, there is so much we can get from the data and setup we have today.

Assisted conversions are a simple way to understand how all channels contribute to conversions (eve