Skip to main content
All CollectionsProfile
How is the Profile score calculated?
How is the Profile score calculated?

Breaks down the calculation of the Accoil Analytics score based on event occurrence and weights, for understanding engagement scoring.

Kate Caldecott avatar
Written by Kate Caldecott
Updated over a month ago

Summary

Breaks down the calculation of the Accoil Analytics score based on event occurrence and weights, for understanding engagement scoring.

How this helps

Provides insights into the engagement scoring process, allowing for refined event weighting and more accurate scoring.

Your Accoil Analytics score is based on two things:

  1. Events;

  2. Event weights (the weights you assign them)

Events represent the main actions users can take in your product.

How the Score is Calculated

Let's consider a hypothetical CRM application with the following events and weights:

Event

Weight

Create New Lead

9

Schedule Meeting

7

Log Call

5

Send Email

3

Update Contact Info

1

Suppose a user performed these events over a specified period:

Event

Count

Weight

Score
(Count x Weight)

Create New Lead

3

9

27

Schedule Meeting

5

7

35

Log Call

10

5

50

Send Email

20

3

60

Update Contact Info

15

1

15

Total Raw Score

187

The total of these event scores is the Raw Score, which in this example is 187.

In order to give you a more “usable” and easily digested, we normalize everyone’s scores to a number between 1-100.

Normalization of Scores

To normalize scores between 1-100, we use an exponential transformation that takes into account the overall distribution of activity. This process involves:

  1. Calculate all raw scores based on the score configuration

  2. Identify the 90th percentile of raw scores

  3. Apply an exponential transformation that normalizes scores relative to this threshold

The 90th percentile threshold helps create a meaningful scale by:

  • Acting as a reference point for what constitutes "highly engaged" behavior

  • Ensuring scores are normalized against actual usage patterns

  • Allowing the system to adapt to changing engagement levels over time

Example Normalization

Consider a set of raw scores:

[475, 89, 101, 7, 3, 21, 2, 149, 223, 1, 13, 9, 37]

The 90th percentile for this set of scores is 208. Using this as our threshold, the exponential transformation produces these normalized scores:

Raw Score

Normalized Score

475

90

223

66

149

51

101

38

89

35

37

16

21

10

13

6

9

4

7

3

3

1

2

1

1

0

Key features of this normalization:

  • Unlike linear scaling, it provides better differentiation between lower scores

  • Higher raw scores show continued improvement but with diminishing returns

  • The transformation naturally handles outliers without artificial caps

  • Scores remain proportional to actual engagement levels

Account Scoring

When scoring accounts, Accoil Analytics aggregates all activities within an account, regardless of the number of users, and applies the same normalization process against all other accounts. Therefore, an account with 20 engaged users will typically score higher than an account with 5 engaged users, and an account with 5 engaged users will likely score higher than an account with 1 engaged user and 19 less engaged users.

This approach ensures that Accoil Analytics provides a comprehensive and fair assessment of user and account engagement, enabling you to make informed decisions based on accurate data.


Understanding Relative Scores

It's important to note that scores are relative to the overall engagement across all accounts. This means that maintaining the same level of raw activity doesn't guarantee the same score over time. Here's an example:

Day 1:

  • Account A Raw Score: 100

  • 90th percentile threshold across all accounts: 200

  • Account A Normalized Score: 39.3

Day 30:

  • Account A Raw Score: 100 (unchanged)

  • 90th percentile threshold across all accounts: 400 (increased due to higher overall engagement)

  • Account A Normalized Score: 22.1

This decrease in score doesn't mean Account A is doing worse – they're maintaining the same level of activity. Instead, it indicates that other accounts have increased their engagement levels, raising the overall benchmark.

This relative scoring approach:

  • Reflects real-world engagement patterns where "good" engagement levels evolve over time

  • Encourages continuous improvement rather than maintaining static activity levels

  • Provides context for how an account's engagement compares to the current user base

  • Helps identify accounts that may need attention even if their raw activity hasn't decreased

When interpreting score changes over time, it's helpful to consider both the absolute raw activity and the relative normalized score to get a complete picture of account health.

Did this answer your question?