Create explainable fit and activity lead scoring models
Combine explicit profile rules with recent completed Hub activity and review the score calculation before applying it.
Use these steps in your selected Pixie CRM company workspace. The module, dependent tools and actions available to you follow its configuration and your permissions.
Hub → Intelligence → Scoring models → Create
PermissionManage CRM resources; native source access; independent Approve and Publish before applying scores
Before you begin
Define what a useful score means for your team. Use measurable saved fields and completed Hub activities rather than assumed browsing or email engagement.
Follow these steps
- Create a scoring model for Contacts, Companies or Leads. Add property rules and/or activity rules.
- Name each rule and assign its points. For activities, choose the supported activity type, lookback period, half-life and maximum counted events.
- Set fit and activity caps and the qualification threshold. Save the draft.
- Assess a representative record in Intelligence desk and inspect the fit/activity breakdown. Adjust the draft where the result is misleading.
- Obtain independent approval and publish before applying the model to a live profile or referencing it in a workflow.
Fields and options
- Rules
- 1–20 rules; label up to 100 characters; points −100…100. Property rules use supported lifecycle, score, health and native record fields.
- Activity
- Completed Call, Meeting, Note, Task or Draft activities; lookback and half-life 1–365 days; maximum count 1–20.
- Bounds
- Fit and activity caps are each 1–100. Threshold is 1–200 and cannot exceed the combined caps.
Rules and limits
- Activity points decay with age using the configured half-life; the calculation examines the latest 100 accessible Hub activities.
- Fit and activity subtotals are bounded by their caps and rounded before combination.
- This is an explainable rule model, not external AI, a conversion probability or evidence of website/email tracking.
Check the result
- Compare the assessment breakdown with the source values and completed activities.
- Check the threshold and caps give sensible outcomes for both strong and weak examples.
If something does not look right
- If an expected event is absent, confirm it is an accessible completed Hub activity within the lookback period.
- If a model cannot be used by a workflow, check source compatibility and current published approval.
Common questions
Can I assess a draft model?
Yes. Assessment is a review step. Applying a score requires the current published model.
Does the model read email opens or website visits?
No. These rules use the supported saved profile fields and completed Hub activities.