Engagement scores engagement-scores

An engagement score is a number that indicates the level of engagement for the members of a buying group. These scores are based on the buying group member activities, weighted actions, and weighted roles. The resulting scores are normalized within a tenant (instance) to enable consistent comparison and allow for actionable insights. Score calculation starts as soon as you create the buying group. The Journey Optimizer B2B Edition data hub system computes the scores daily and uploads them to the Multi-Level Marketing (MLM) MySQL system using the ingestion service.

There are two types of engagement scores:

  • Buying group engagement score - The buying group engagement score is a normalized score between 0 to 100 and is based on the engagement score calculated at the person level.

    The buying group engagement score is displayed in the Buying group details page. You can also view the most engaged buying groups in the Intelligent dashboard.

    Most engaged buying groups {width="700" modal="regular"}

  • Person engagement score - The person engagement score is based on the activities of an individual buying group member.

    The person engagement score for each buying group member is displayed in the buying group details page Members tab. These scores are also displayed in pages and dashboards that include top-engaged members and overlapping contacts information.

    Most engaged buying group members {width="550" modal="regular"}

recommendation-more-help

The person engagement score is an attribute that is available to use for filtering in roles templates and journey splt-path-by-people nodes.

Access the configured event definitions {width="550" modal="regular"}

Any engagement weighted activity performed by the members of the buying group in the last 30 days is used to calculate the scores. With the 30-day window, activity occurrences expire and scores can move downward (score decay). Displayed scores are rounded (for example, a score of 75.89999 is displayed as 76).

Activities used for engagement scoring

Buying group scoring is not triggered-based. It is a daily process that evaluates the activity across all the members of the buying group and recomputes the score. Activities use weights to inform buying group scoring according to the active weighting model, which determines how each activity is weighted.

There is a daily frequency cap of 20 for each activity. If a member of a buying group performs the same activity more than 20 times in a single day, the count for the activity is capped at 20.

Activity name
Description
Engagement type
Max daily frequency count
Default model activity weight
Attend Event
A member attended an event
Event
20
60
Email Clicked
A member clicks a link in an email
Email
20
30
Email Opened
A member opens an email
Email
20
30
Form Filled Out
A member fills and submits a form on a web page
Web
20
40
Interesting Moment
A member has an interesting moment
Curated
20
60
Link Clicks
A member clicks a link on a web page
Web
20
40
Page Views
A member views a web page
Web
20
40
Register for Event
A member registered for an event
Event
20
60
NOTE
Engagement score activities are recorded in the Marketo Engage activity log for a person. You can access this log in the connected Marketo Engage instance. For more information, see Locate the Activity Log for a Person in the Marketo Engage documentation.

Role template weighting engagement-score-weighting

Users can assign weighting to each role in the roles template to allocate different weights for a role.

Set weighting to each role in the roles template {width="700" modal="regular"}

Each weighting level translates to a value, which is used for calculating the engagement score:

  • Trivial = 20
  • Minor = 40
  • Normal = 60
  • Important = 80
  • Vital = 100

A roles template with three roles weighted as Vital, Important, and Normal convert to the following weighted percentages:

Role
Weighting
System value
Value calculation
Percentage
Decision Maker
Vital
100
100/240
41.67%
Influencer
Important
80
80/240
33.33%
Practitioner
Normal
60
60/240
25%
Total
240

Score calculation example

The following example illustrates the engagement score calculation. It uses the outlined role weight percentage, count of inbound activities for each buying group member, and a daily cap of 20 for each event occurrence.

Role
Member
Activity type
Yesterday’s count
Today’s count
Calculation
Total score
Decision Maker
Adam
Visited website
37
15
20 + 15
35
Clicked email
1
1
1 + 1
2
Mark
Visited website
5
3
5 + 3
8
Clicked email
1
1
1 + 1
2
Downloaded pub
3
2
3 + 2
5
Decision Makers total score
52
Influencer
John
Visited website
19
9
19 + 9
28
Influencers total score
28
Practitioner
Bob
Clicked email
1
1
1 + 1
2
Paul
Clicked email
1
1
1 + 1
2
Calvin
Clicked email
1
1
1 + 1
2
Visited website
1
7
1 + 7
8
Downloaded pub
1
2
1 + 2
3
Practitioners total score
17

The final engagement score is calculated by applying the weighting for each of the role scores:

Role
Role total score
Role weight %
Score X weight %
Decision Makers
52
41.67%
21.67
Influencers
28
33.33%
9.33
Practitioners
17
25%
4.25
Final engagement score
35.25

Scoring logic

In addition to the calculation logic outlined in the calculation example, there is a significantly complex normalization of scores that occurs in the system, across all people, buying groups, and accounts in your instance. A buying group engagement score has a dependency on the person engagement scores, according to the following ordered logic:

Person engagement score calculation logic

  1. Identify all engagement-weighted activity types that have an associated weight and daily quota, such as website visits, email clicks, and webinar attendance.

  2. Identify all person engagement-weighted actions performed within the activity look-back window, which is currently hard-coded to 30 days.

  3. Normalize the activity type weights across all engagement-weighted activity type weights identified in step 1, ignoring the ones that did not occur within the look-back window.

    This step leverages Min-Max Normalization and significantly reduces the artificial dilution of activity type weight for a tenant that does not leverage most of them.

  4. Apply the daily quota filtering per person and activity type.

    This step mitigates having very large outliers by avoiding lower value / high volume activities skewing the scores.

  5. Calculate the raw person engagement score by summing the daily activity per activity type, multiplying it by the associated weight, and then summing the results for all days of the look-back window.

  6. Use a Power Transformation (Square Root) transform to stabilize variance by reducing possible outliers.

    This transformations helps to reduce skewness and make patterns in the data more linear.

  7. Apply an additional Scaled Normalization transform to ensure that the scores leverage the entire range from 0 to 100.

Buying group engagement score calculation logic

  1. Apply a normalized weight to each buying group member by role, according to the weight configured in the roles template.

  2. Normalize the buying group role weight for each buying group.

    This normalization avoids unnecessary role weight dilution if a buying group does not use all roles.

  3. Aggregate all buying group member person engagement scores by multiplying the person engagement score by the person’s role normalized role weight, and add them together.

  4. Apply a Power Transformation (Square Root) transform to stabilize variance by reducing possible outliers, especially for very large buying groups.

  5. Apply an additional Scaled Normalization transform to ensure that the scores leverage the entire range from 0 to 100.

6ef00091-a233-4243-8773-0da8461f7ef0