Google Analytics Attribution Models and UTM Parameters
How to make accurate Google Analytics reports
The Statsbot team continues the series of advanced Google Analytics articles. In our previous post, we were talking about a quality Google Analytics audit and must have tools. Now, it’s time to choose the right attribution model and UTM parameters for your exact needs.
Today’s analytics tools are incredibly powerful, and there’s no shortage of metrics you can track. But that doesn’t mean you get meaningful data to make a major decision.
Are you confident spending thousands of dollars based on what Google Analytics is telling you about where your conversions are coming from? Are you sure?
If you are similar to most people, it’s likely you are using the default attribution model in Google Analytics, i.e. Last Interaction.
Just because it is the default option, that does not mean it’s the right one for you. In this article, we will tell you how to set up a Google Analytics attribution model and UTM parameters based on your product and the stage of your company.
Google Analytics attribution models
Let’s look at an example. Kate wants to buy a chair for her home office.
Interaction 1 (Organic Search)
Kate Googles “comfortable office chair,” looks at some pages on yourofficechairs.com, and leaves.
Interaction 2 (Social Network)
3 days later, Kate sees an article on Facebook that mentions how comfortable your chairs are. She clicks the link to yourofficechairs.com and looks at the recommended chairs.
Interaction 3 (Paid Search)
7 days later, Kate Googles “chairs for home office” and sees your ad. She clicks it and completes the purchase.
According to this case, you could have the following credits for a conversion for different GA attribution models:
1. First Interaction/Click
All the credit for a conversion goes to the first interaction.
Organic Search: 100%Social Network: 0%
Paid Search: 0%
2. Last Interaction/Click
All the credit goes to the last interaction.
Organic Search: 0%
Social Network: 0%
Paid Search: 100%
3. Linear
Conversion is attributed equally to all the interactions.
Organic Search: 33%
Social Network: 33%
Paid Search: 33%
4. Time decay
Recent interactions are given more credit than older ones.
Organic Search: 10%
Social Network: 30%
Paid Search: 60%
5. Position
Credits are given based on the specific position of the interactions.
Organic Search: 40%
Social Network: 20%
Paid Search: 40%
How do you know which attribution model is right for you?
You need to consider the combination of several factors before deciding on the attribution model for your specific needs. Here are a few factors that might influence your decision.
1. Product Price
For low price items, the Last Interaction model works well. This is especially true for one-time purchases, say, a t-shirt. For higher priced items, say, a couch, people tend to compare and take longer. For such cases, First Interaction, Linear or Time Decay are better options depending on other factors like Consideration Period and Brand recognition.
2. Consideration period
If people tend to buy your item within a few days of thinking about it, then Last Interaction or Time Decay models are a good choice. Time Decay is especially useful when you are running promotion campaigns that expire in a short period of time.
However, if your customers take a long time to decide, then look at other factors. This is true even for low priced items. For example, if you sell a subscription for $5/month (low price), but your customers take a few weeks to decide, then the Last Interaction model is not the right choice.
3. Time investment
If your product requires a time commitment from your customers to use it, say, productivity software or project management software, they’ll tend to do a lot of research and comparison shop. A low price tag may not overcome this need.
This is true for most software sold to businesses. In this case, Last Interaction is not a good option. Look at Brand recognition to decide between the First Interaction and Linear models. One of these might work for you.
4. Brand recognition/awareness
If you are just starting out and not many people know about you, First Interaction is the right choice. This helps you identify the marketing channels where you can focus to generate even more awareness about your company or product.
If you are a company with a well-established name, say, Hubspot or Salesforce, then you don’t need to worry about First Interaction. What you care about is how to convert visitors into customers. You want to know what interactions convince people to buy.
In this case, the Linear model might work well for you. For example, you should try to find out which of your email nurturing campaigns, ebook downloads, etc. are helping people convert the most.
5. One-time or subscription pricing
Subscription products always require more interactions than one-time purchases before people buy. Last Interaction can work for one-time purchases, but may not be a good choice for subscription products.
For subscription-based products, look at other factors. In this case, look at Brand recognition to decide between the First Interaction and Linear models.
6. Identifying ROI of Marketing Channels
Sometimes you want to know if your specific channels are generating enough ROI to continue spending money on them. Paid ads are one such channel that can get expensive pretty quickly. You need to know if it is working or not as soon as possible. In this case, the Last Interaction model is a good option.
Having said that, this is not an exhaustive list and there are probably other factors unique to your business that will affect your decision. But hopefully this gives you an idea of how to think about these attribution models.
How can you improve your attribution reports?
No matter what attribution model you choose, it still needs to detect the correct traffic source. It’s your responsibility to make sure that most of your website visits have this information.
The two most important things you can do to make your analytics reports more accurate:
Add UTM parameters to all the links you share in your campaigns
Your website can get traffic from a variety of devices, apps, websites, etc. But some channels, e.g. mobile apps and email clients, will not set any referrer information. UTM parameters are an excellent way to make sure that you can can fill that gap.
- Define your UTM conventions
This step may seem daunting at first, but even having some basic guidelines can have enormous benefits. Think about the reports and values you would like to see in Google Analytics and it becomes easier to come up with your own UTM conventions.
Quick Tip:
Always use utm_medium=email for your email campaigns. Google Analytics will automatically attribute these visits to the Email channel.
Always use utm_medium=social for your social campaigns. Google Analytics will put these visits under the Social Network channel.
For more exhaustive guidelines, look at UTM Parameters Best Practices course.
These conventions are key in making sure that your conversions are attributed correctly and your chosen attribution model gives you the right data.
2. Follow it consistently for all URLs
Make sure to add UTM parameters to all the URLs that you share. This includes email, social, ads, etc. The more consistent you are in tagging your URLs, the better your reports will be in attributing your conversions correctly.
3. Use a UTM builder and URL management system
- Google URL Builder helps you build UTM URLs, but it doesn’t track them.
- A spreadsheet is better since it also keeps track of your past URLs in one place. However, it can get messy pretty quickly.
- You can also use a more advanced UTM builder, like Terminus, that enforces your UTM conventions.
Enable HTTPS on your website (if you haven’t already)
HTTPS (or SSL) is not just about security and privacy for your users. It’s also important for your analytics.
Let’s say you publish a new blog post, http://www.yoursite.com/new-blog-post, on your website (no HTTPS). A blogger, John, mentions your new post on their website that has HTTPS, say https://www.johnblog.com. A visitor on John’s blog clicks on the URL and comes to your website.
All modern browsers will remove any Referrer information before they send a visitor to your website (non-HTTPS) from an HTTPS-enabled website. This visit will now show up as Direct in your analytics reports.
Since more and more websites are moving to HTTPS (including all the big ones like Google, Facebook, etc.), you are at a disadvantage if you have a non-HTTPS website. Having HTTPS on your website will make your attribution and other reports much more accurate and minimize Direct visits.
Final Thoughts
Attribution models are no longer reserved only for advanced users. In fact, the default Last Interaction model might even mislead you. Choose the right one for your business and change it as your needs evolve. UTM parameters can then be used to feed the right data to your analytics tools to get the most accurate reports possible.