Attribution Modelling is a complex topic and I could write an entire book just about this topic… For now, let’s focus just on the basics. If you would require some help finding the right fit for your campaigns, feel free to reach out.

Before we dive into the various attribution models and how you can fix their flaws, let’s quickly revisit what a conversion is.  

When I’m talking about “conversions”, we’re not only talking about sales. A conversion is any desirable action that a user performs. This could be signing up for your newsletter (form submission), making a donation, initiating or completing a purchase, submitting a request form, downloading a white paper, and even something as trivial as visiting a particular page on your website that’s important to you.

Conversions are the goals of each of your campaigns. This could be event registrations, downloads, subscriptions, demo registrations, purchases, and many more. Ultimately, why you are measuring your campaign’s successes is just as important as knowing what to measure.

Another important implication to understand is that oftentimes your emails, social media posts, and online ads are driving traffic to your website and it’s the website’s job to “convert” these visitors. Before blaming your website for the failure of a campaign, however, it’s important to ensure that you have done a good job in qualifying your leads and sending these visitors to the right place.

Another important factor to take into account is that marketing doesn’t happen in a vacuum. Don’t assume that because there are no conversions attributed to a certain channel (e.g. social media), your efforts are fruitless.

By default, Google Analytics uses the “last click attribution” model. To use an analogy, this would be like giving the cashier at a store 100% of the credit for a sale.  

Now let’s imagine that based on this last-click attribution, the company decides to lay off the entire marketing and sales department, because no credit was attributed to their efforts… This is why you shouldn’t base important business decisions on only one attribution model. 

When to use the Last-Click attribution model

If your sales cycle does not include a research and consideration phase, the last-click attribution model would be useful.

In most situations, however, a conversion is often the outcome of team effort not of an individual’s success. A lot of different departments play their parts in making a purchase happen. The same is true for marketing channels as well. So, let’s explore some other attribution models and when to use them.

As opposed to the last-click attribution model, the first-click attribution model gives 100% of the credit to the first channel the user interacted with.

When to use the First-Click attribution model

If you would like to know which campaigns scored the best in creating initial awareness for your company or product, this attribution model would be helpful.

This attribution model gives each channel that was involved in a conversion the same amount of credit.

When to use the Linear attribution model

This attribution model is very helpful in understanding the involvement of various marketing channels. If your conversion path involves multiple touch points that aim at reconnecting with a particular user until a conversion has happened, you might find this attribution model helpful.

This attribution model is a combination of the first and last-click attribution models. You could assign 40% of the credit to both the first and the last interactions and 20% of the credit to all interactions that happened in between.

When to use the Position-Based attribution model

This attribution model is for you if you place the highest value on the first interaction (awareness) and last interaction (purchase) and consider all interactions in between as less important. I personally hardly ever use this attribution model, however, it can sometimes help understand the customer journey.

This attribution model gives the most credit to touch points that occurred closest to the conversion and then works backward with attributing credit. In this attribution model, interactions that occurred within a 7-day period receive 50% of the credit a touch point receives that occurred on the day of the conversion. An interaction that occurred within a 14-day window prior to the conversion would receive ¼ of the credit and so forth.

When to use the Time Decay attribution model

If the research and consideration phase in your sales cycle is short or if you run “flash sale” promotions, you might find this attribution model helpful, because it gives more credit to more recent interactions.

This attribution model is similar to the last-click attribution model mentioned above. However, instead of attributing 100% of the credit to the “last click” it attributes 100% of the credit to the last click an AdWords ad received.

When to use the Last AdWords Click attribution model

This attribution model is helpful when you run Google AdWords ads and you would like to know which of your campaigns perform best in driving particular conversions.

This attribution model ignores direct traffic and instead assigns 100% of the credit to the last marketing channel the user interacted with.

When to use the Last Non-Direct Click attribution model

Filtering out direct traffic can often be helpful to monitor the performance of your marketing campaigns.

This attribution model is only available in the Google Analytics 360 Suite and a threshold of 400 conversions per conversion type is needed over a time period of 28 days.

This attribution model uses your account’s data to map and understand user interactions. It calculates conversion probabilities for the various interactions and thus generates a custom attribution model. It provides an actionable view of your best performers to enable you to make informed, data-driven marketing decisions.

Often, smaller marketing channels, such as email or advertising don’t get the credit they deserve. In Google Analytics, the bulk of credit goes to … drumroll please … direct traffic.

Yet, it is very unlikely that direct traffic is the superhero of marketing. Otherwise, your business would be just as successful (if not more successful due to cost reduction), if you cut out all other marketing channels. Anyone who has ever fallen into the trap of shutting off important marketing channels because of analytical misperception, knows this. 

While we’re on the topic of direct traffic, I’ve written a whole article on what “direct” traffic actually is.

Various attribution models allow you to understand the conversion behaviour of your website’s visitors. Whether you’re interested in finding out what happened just before a conversion occurred or you’re trying to identify your most effective marketing channels, attribution models are your friend, if used correctly.

Before you head over to Google Analytics to test them out, be warned: Attribution models only show you what you ask them to show. One important point to understand is that Google Analytics’ can make mistakes, when your conversion tracking and goals are set up incorrectly, when you’re segmenting your data the wrong way, or you make mistakes with sample sizes, among others. Consider these implications before you jump to conclusions.