Understanding Lytics / Product Documentation / Descriptive & Predictive Modeling

Behavioral Analysis

Lytics constantly tracks events from many data sources: web, email, salesforce, and custom sources as well. Looking at this data directly is not useful; it needs to be summarized. Lytics allows for simple summaries by means of user fields and for use in audiences, but there is more to be learned from all this behavioral data.

Behavioral analysis asks many questions and computes the answer. Questions such as:

  1. When did the user last visit?
  2. When did they visit before that?
  3. Is it unusual that they visited right now?
  4. During their visit, what were they up to?
  5. Did they visit a lot of pages?
  6. Did they stick around for a long time?
  7. How does that compare to other users?
  8. Are they going to come back?
  9. Is this a good time to encourage more interaction?

After finding the answers to these questions, for every user, Lytics uses these answers to inform six low-level scores that get assigned to every user who has shown enough behavior for a confident assessment.

These scores are computed automatically, they are kept up to date continuously, and they are assigned relative to other users in the account.