Attributing Referral Sales Inside Your Shopify Store

Illustration by Isabela Humphrey
Words by Nirman Dave


Continuing our Guide to Machine Learning and eCommerce, we break down how to predict attribution with machine learning inside your Shopify store.

Let’s get into it.

What is Channel Attribution?

First, you might be wondering what we mean by channel or referral sites. Both are ways to bring traffic to your Shopify store and are used in marketing strategies. For example, when a startup is featured in TechCrunch, the link they provide in the article brings traffic to the startup’s website which could increase signups, thereby increasing revenue. There are other community sites which could lead potential customers to your store or blog posts which mention your company. These are also referral sites.

Channel attribution in machine learning is predicting which channels will bring you the most profit. In terms of our Shopify app, you can predict how much your referral sites are going to make you in the future based on data. This is useful for knowing which referral sites to double down on and build a presence on. 

Using historical data inside your Shopify store, our app automatically predicts how much referral sites make you. All you need to do is click one button.


For this post, we will focus on the question “Which referral sites to double down on?”

Hit the check box and press “Predict” to get a report.

Inside Your Prediction Report

  1. You can instantly see the predicted revenue you will make from each referral site promoting your product. For mkrpd.io, you can see there’s a predicted $65.204 plus or minus $.021.


  1. Scrolling down, you can get a more detailed explanation of the prediction graph and see the top drivers that impact referral revenue. 




  1. Name your report something memorable with a date, we recommend stating what you predicted to easily scroll and compare month by month. Other great names would be “Referral Attribution - Q1 2020” or something like that.


  1. At the top of your report you can export your report to a CSV file to archive or share later. By hitting the “More” drop down menu, you can also refresh your report as new sales come in if you want to keep your predictions all in one updated report.
  2. At the bottom of your report, you can see your tech specs of the report. For more information on how we preprocess data and build machine learning models, read this post.

Explore Other Applications of Machine Learning in Shopify

The uses of machine learning for eCommerce doesn’t stop at sales forecasting.

Read how to predict sales forecasting and attribution.

If you’re interested in using non-eCommerce data to make predictions, join our Shopify waitlist.

Exclusive datasets, guides, and insights to your inbox.

Join 3,000 subscribers. GDPR and CCPA compliant.