Chain-Based Attribution Model
Are you using the last touch attribution model? first touch? even-weight? Are you finding out that these models don't accurately weigh touch points in the customer journey based on their influence?
Attribution is not easy, and using a simple attribution model rarely delivers the value that you're looking for.
What you need is a model that gives fair credit to different channels in the customer journey based on actual influence.
But imagine if instead of using an attribution model that unfairly gives credit to different channels in a customer’s journey you could use one that assigns credit based on actual influence?
Introducing the CaliberMind Chain-Based Attribution Model: Using data science to analyzing probabilities of linked events in the customer journey, we're now able to predict sales opportunity conversion with a much higher level of accuracy than previous marketing attribution models -- ultimately leading to more revenue and better decisions for your B2B Enterprise.
What is a Chain-Based model?
Coined by Andrey Markov, a Markov Chain model allows you to take chain of events (for example clicking on an email link and then visiting your company's pricing page), and predict whether another event in the chain will happen.
In attribution, Markov chains are extremely powerful and generally considered superior to other attribution models.
For example, in the last touch attribution model, the last interaction a customer has in their customer journey is given all the credit for the conversion, when in fact, every interaction in the customer journey likely influenced the customer to convert in some way - the visiting your website, reading your e-newsletter, chatting on the phone, etc.
How is a Markov chain model used in attribution?
Using a Markov chain model, each touch point is fairly given credit for their influence based on the outcome you want to achieve.
Want to close more leads? Increase e-newsletter subscriptions? Determine the event that generates the most revenue? A Markov chain model can help you do that.
Say you ran a webinar campaign. You could use a Markov chain model to see how organic search traffic contributed as compared to paid ads traffic, determine which event had the highest probability of generating webinar traffic, and then predict how to best allocate your marketing budget for the next webinar. You can also find out which events would cause the largest drop in traffic if they were to stop by looking at the Removal Effect report.
Like all attribution models, a Markov chain model has its flaws.
If you don't have enough data, you could get incorrect estimations. The general rule of thumb is to take the number of events that you want to track, and multiply it by 10. So if the outcome you're tracking has 10 touch points, you'll want to have at least 100 transitions between them.
The quality of your data affects the quality of the estimations. The lower the quality of your data, the lower the quality of the estimations.
Despite these flaws, the Markov chain model is a far superior model to traditional attribution models.
To find out how you can use Markov chain attribution models in your CaliberMind attribution reports, contact your Customer Success Manager.