MadKudu trains behavioral models which learns from historical patterns to uncover specific behaviors that separate leads who were on the path to conversion from others.
For the model training to be done smoothly, there is a key element in the MadKudu workflow called Data Management. This document aims to detail how Data Management works in MadKudu's platform.
What is Data Management in the MadKudu world?
You have events from multiple data sources such as Salesforce, Segment, Amplitude and more. Today, it is a challenge for you to gather all these events in your database since they live in different systems and have different ways in which they are tracked.
What we do is to:
- Link the events a contact did to one identifier
- Name these events in a human-friendly way
- Aggregate similar events from different system to one meta event that makes sense from a business logic stand-point
How did we come up with the standard events?
After analyzing 30 B2B SaaS companies, we've noted typical patterns of web, email, sales, app and marketing activities to come up with a list of standard events that are critical in identifying in order to train a behavioral model.
Here's the list:
You can also add custom events if you have more events outside of this list, and we use them in the behavioral model.
What happens next?
- We would then aggregate these events or use them as-is to identify the impact of them on your conversions.
- Based on that analysis, they will then be applied accordingly in the model.
- Once the model has been configured, these events are also used as part of the Signals that we push into your CRM platform.