Skip to content

How /last30days Scores Research Results by Real User Engagement Instead of SEO

Problem

Google ranks by SEO signals, backlinks, and editorial curation. For what’s happening right now — this month, this week, today — SEO-centric ranking misses the real signal: what real people are actually talking about and engaging with. A blog post nobody read can rank #1 because it has good SEO. A Reddit thread with 1,500 upvotes might not appear on the first page.

The Scoring System

/last30days scores every result by real user engagement signals. Each platform contributes its native signal:

PlatformScoring SignalWhat It Measures
RedditUpvote count on posts and top commentsCommunity consensus
XLikes, retweets, repliesReal-time engagement
YouTubeView count, like ratioWatch-time attention
TikTokView count, engagement rateViral cultural signal
Hacker NewsPoints, comment countTech community interest
PolymarketOdds percentage backed by real money volumeMoney-backed belief

How Scoring Works

The scoring pipeline has three stages:

1. Native Scoring. Each result is scored within its own platform. A Reddit post with 2,300 upvotes gets a high native score. A YouTube video with 3.6M views gets a high native score. These are raw, unnormalized signals.

2. Cross-Platform Fusion. The fusion engine (weighted_rrf — Reciprocal Rank Fusion) combines scores across platforms. This normalizes the different scales: a Reddit upvote is not the same as a YouTube view, so the fusion algorithm weights them appropriately. The fusion.py module implements this.

3. AI Re-Ranking. An optional AI judge re-ranks by relevance to the query. The final ranking favors three things in order: higher engagement, more recent, and more relevant to the topic.

Scoring pipeline
Raw results from each platform
┌─────────────────────┐
│ Native platform │ Upvotes, likes, views, odds
│ scoring │ per-platform normalization
└──────────┬──────────┘
┌─────────────────────┐
│ Cross-platform │ weighted Reciprocal Rank Fusion
│ fusion (fusion.py) │ combines different signal types
└──────────┬──────────┘
┌─────────────────────┐
│ AI re-ranking │ Optional judge re-ranks
│ (rerank.py) │ by relevance to query
└──────────┬──────────┘
Final ranked results

Why Engagement Beats SEO

A Reddit thread with 1,500 upvotes is a stronger signal than a blog post nobody read. A TikTok with 3.6M views tells you more about cultural relevance than a press release. Polymarket odds backed by $66K in volume beat a pundit’s guess.

The synthesis ranks by social relevancy, not SEO relevancy. This matters most for current topics — breaking news, product launches, controversies, trends — where editorial curation hasn’t had time to catch up.

Summary

In this post, I explained how /last30days’s scoring system surfaces what real people are actually engaging with. The key point is that social relevancy beats SEO relevancy for understanding what’s happening right now — engagement signals from real users are harder to game than backlinks.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!

Comments