Learn more about the Alteryx Maveryx Universe recently announced at Inspire 2023!

Alteryx Designer Desktop Discussions

Find answers, ask questions, and share expertise about Alteryx Designer Desktop and Intelligence Suite.

Customer Lifetime Value (CLV) and Digital Attribution

Inactive User
Not applicable

Hello Everyone,


We are relatively new to Alteryx, but have already had several successes using the download tool, blending tools, and forecasting tools (ARIMA).


Now that we have gotten our feet wet, I want to start researching how to really take advantage of everything Alteryx has to offer, in particular by designing and building two solutions that I feel Alteryx would be ideal for.


Specifically, I'm talking about:


  • Customer Lifetime Value (CLV) - A prediction of the value (could be net or gross) a customer will have over it's lifetime based on predictors.  The goal is to aquire high value customers based on the identified predictors, but also to move customers into higher-value segments by marketing to them in a certain way that historically has driven up value.
  • Digital Attribution - Identifying events / actions that contribute to an outcome, and then assigning value to each events.  This is used to understand how best to create purchase conversions.

I think the Spline model tool would be a good place to start w/ CLV, but not sure where to start w/ Attribution.


Both projects should take advantage of good portions of the Alteryx toolset, so we're excited, but figuring out where to start is a little daunting.  


Any ideas / starting points is very much appreciated!



Matt Simcox

Director of Business Intelligence



Alteryx Alumni (Retired)

Hi Matt,


This definitely sounds like you'll have to do quite a bit of model development, and I'd be happy to point you in the direction of a few resources. For ~starters~, take a look at these starter kits (pun intended) in the Predictive District on our Analytics Gallery.


  • Predictive Analytics Starter Kit Volume 1
    • Description: The Predictive Analytics Starter Kit enables you to learn the fundamentals of key predictive models with an interactive guided experience. This Starter Kit will help you understand how predictive analytics can be used to answer business questions. Contained in this kit are samples that will help you understand how: A/B testing can be used to answer "How does a price change impact my bottom line?", linear regression can be used to answer "How can I predict how much a customer will spend?", and logistic regression can be used to answer "How can I predict whether a customer will purchase the product I put on sale?". The Predictive Analytics Starter Kit demonstrates the steps necessary to develop the dataset need for analysis, and then how to actually build these predictive models yourself.
  • Importance Weights
    • Description: The Importance Weight tool provides methods for selecting a set of variables to use in a predictive model based on how strongly related each possible predictor is to the target variable.
    • This may be a great way to help determine which of your predictors are strongly correlated to succesful customer segments.
  • Variance Inflation Factors
    • Description: The Variance Inflation Factors produces a coefficient summary report that includes either the variance inflation factor (or VIF) or a generalized version of the VIF (GVIF) for all variables except the model intercept.
    • This may be useful as well. It's often the case that certain predictors are colinear, i.e. an increase in one is strongly correlated with an increase in another predictor. For best model-building, you would want to limit this as much as possible. This tool can help you spot these extraneous predictors and exclude them from any regression-based model building.
    • Check out this sample for more on this tool!

I think these two may be particularly useful for narrowing down predictors to use for your CLV, but then the next step would be to create a model. You can use historical success to help validate the model, and create scores for future success. There are a variety of ways to do this, but I think for your case you may want to consider using a Decision Tree. This would give you a good idea of how the predictors split up your customers into segments based on success.


  • This Virtual Training recording, Predictive Analytics, Part 2: Regression Modeling, that is a particularly good resource for regression methods, as well as the Decision Tree (40:00 min mark). This recording can be found on our Virtual Training site under "Watch a Past Session", and is one video in a series of four (all are great, I highly suggest watching them if you would like to get an idea for the model building process).

Hope these help you get the mental gears turning. If you run into questions or issues, let us know over at Customer Support. We won't be able to build models out for you, but we can certainly discuss options or provide resources for you to look to as you work through these cases.




Alex Koszycki
Program Manager, Community Platform