Scoring Methodology
Indie Pledge is dedicated to building a transparent recommendation ecosystem. We are publishing our core scoring logic and principles here.
Activity Board
Answers: "Is this product being actively maintained?"
Primarily based on effective update frequency in the last 90 days, combined with staleness penalties.
Treasure Board
Answers: "What is the real user experience like?"
Primarily based on blended rating quality and bug report rates.
Activity Board Details
Core Baseline: Effective Updates
The base score is entirely based on effective updates in the last 90 days. We apply a layer of filtering:
- Only 1 effective update per day is counted.
- Pure marketing announcements are excluded.
- Duplicate updates are normalized and removed.
Staleness Penalty
The longer the time since the last effective update, the higher the penalty. Long-term inactivity can result in negative scores.
Engagement Bonus
Claimed products that continuously maintain their roadmaps and publish social content receive a minor bonus.
Tie-Breaker Mechanism
When multiple products have the same activity score, we tie-break sequentially by:
- Effective updates in the last 90 days
- Days since the most recent effective update
Treasure Board Details
Rating Quality & Smoothing
To prevent a tiny number of reviews from pushing a product to the top, we use Bayesian smoothing and multi-source blending:
Bug Report Rate Penalties
Real user feedback on crashes and errors is a crucial indicator. Reputation = Rating Quality + Bug Health.
- High bug rates reduce the score, while active responses provide minor trust restorations.
- If a highly-rated product has a significant bug rate, its reputation score is capped.
Why Scores Change
Rolling 90-Day Window
A recent update does not prevent an older one from rolling out of the 90-day window, which may lower the base score.
New Activity Not Counted
Marketing-heavy, duplicate, or same-day repeated posts may not be counted as effective updates.
Bonus Signals Shifted
Roadmap maintenance, social content, and claim status provide bonuses. If these signals weaken or expire, the score will adjust.
We Want More Than Up or Down
We are continuously improving score explainability. Timeline descriptions aim to clarify whether a change was due to rolling windows, staleness penalties, or filtered updates.
How to Read the UI
Activity Example
This helps distinguish between recent product activity and whether the system treated it as an effective update.
Reputation Example
Values Behind the Boards
Activity Is Not Marketing
We reward verifiable iteration and continuous maintenance, not bursty marketing activity.
Reputation Is Not Praise Inflation
We use sample thresholds, smoothing, and bug signals to avoid inflated scores from a small number of positive reviews.