Understanding Score Discrepancies Between Account and Developer Activity Score

You may occasionally notice situations where an account has a high Activity Score (e.g., 72), while the linked developers have relatively low scores (e.g., 2 and 46).

This may seem contradictory at first, but it’s expected behavior based on how scores are calculated differently and relatively within Reo.Dev.

Here’s what’s happening under the hood:

Key Difference: Scoring Models Are Relative & Contextual

  • Developer Activity Score (DAS) is individual-level and ranked relative to all de-anonymized developers across the platform. A developer with a score of 46 is more active than only 46% of developers.

  • Account Activity Score (AAS) is an aggregate metric computed by summing up engagement across all linked developers, both anonymous and de-anonymized, and is then ranked relative to all other accounts in your tenant.

So while developer scores apply only to known individuals, account scores factor in all signal sources — including activity from anonymous developers or unidentified interactions (e.g., doc views or telemetry with no identity match).


Real-World Example: Relative Scoring in Action

Imagine the following:

  • You have 7,000+ accounts in your workspace.

  • Over 5,700 of them have an Activity Score below 20 (Low).

  • If your account has even minimal meaningful engagement from 2 known developers, it may still rank in the top 30% of accounts, giving it a “High” score (60+) even if the linked developers' scores are modest.


Why This Matters

  • An account score is not the sum of its developers’ scores. It is a percentile rank of total aggregated activity (across all sources), which may include:

    • Anonymous traffic

    • Partial GitHub engagement

    • Docs usage

    • Telemetry pings

    • Page visits or signups not tied to specific identities

  • Developer scores only consider activity from de-anonymized individuals. So, even if developers appear inactive, background activity at the account level can still push up the overall AAS.


TL;DR:

The system uses relative percentile scoring for both developers and accounts — but each within their own population.

A high account score with low developer scores doesn't mean an error; it reflects broader engagement signals across your product and ecosystem.


Last updated

Was this helpful?