Bug / Security Issue – Inactive Podcasts Ranking in “Top — Past Week”

[lowenthalrobin-dotcom]

Summary
The “Top — Past Week” sort is surfacing podcasts that have been inactive for years, ranking them above or alongside actively publishing shows. This suggests the sort is based solely on pageview counts, without any penalty for inactivity. It misleads users and is trivially manipulable.

---

Steps to Reproduce

1. Open Podverse (web or mobile).

2. Perform a search (examples below).

3. Set sort order to “Top — this week.”

4. Observe results.

Example 1 – Keyword: “Frum”

Inactive shows (last episodes: 2014, 2020, 2021) appear above or alongside:

Active show (last episode Sep 9th 2025).

Example 2 – Keyword: “News”

Inactive shows (last episodes: 2016, 2021, 2022) appear in the top results alongside:

Active shows (last episodes: Sep 16th 2025, Nov 25th 2024).

---

Why This Matters

User trust: “Top this week” implies currently active/popular shows. Showing feeds inactive for years undermines trust.

Easy to game: Pageview counts are trivial to inflate via bots, especially in politicized categories like News & Politics.

Discovery harm: Active creators are buried under inactive or artificially boosted feeds.

---

Probable Cause

Popularity counters (e.g. pastWeekTotalUniquePageviews) are computed in
podverse-api/src/services/stats.ts and used in
src/controllers/podcast.ts and the episodes query endpoint for sorting.

These counters are based on pageviews only, with no activity filter or anti-bot measures.

---

Suggested Fixes

1. Activity filter: Exclude or down-rank podcasts with no new episodes in the past 6–12 months from “Top” sorts.

2. Hybrid ranking: Blend pageviews with recency of publish.

3. Bot detection: Throttle abnormal pageview spikes; de-duplicate by IP/device fingerprint.

4. Transparency: Document how “Top” is calculated. If it’s pageviews only, rename it to “Most Viewed in Podverse (past week)” to avoid misleading users.

---

Impact
Currently, Podverse’s “Top” sort can be manipulated and is producing misleading discovery results. This undermines the credibility of the app and its mission to support free podcast discovery.