Skip to content

feat(ops): capture top referrers per repo in weekly snapshot#538

Merged
amavashev merged 1 commit intomainfrom
ops/snapshot-referrers
May 3, 2026
Merged

feat(ops): capture top referrers per repo in weekly snapshot#538
amavashev merged 1 commit intomainfrom
ops/snapshot-referrers

Conversation

@amavashev
Copy link
Copy Markdown
Contributor

Summary

  • Adds .outreach/adoption-snapshots-referrers.csv (long format: date, repo, rank, referrer, views, uniques) seeded with today's data across all 16 public repos.
  • Extends .github/workflows/adoption-snapshot.yml to fetch traffic/popular/referrers per repo and append rows alongside the existing totals + by-repo writes.
  • Adds the new CSV to the auto-PR's git add list.

Why

Pulling referrer data today made the picture much sharper than clone counts alone could. In a 14-day window across 16 repos:

  • ~95% of page views came via github.com (internal navigation, scrapers, indexers)
  • 60 from runcycles.io — 4 unique humans
  • 7 from LinkedIn — 1 unique
  • 4 from HN — 2 uniques
  • 1 from Reddit — 1 unique

Translation: ~8 unique humans arrived from any external content. Clone counts are dominated by automation. Referrer data is the cleanest signal for whether distribution efforts (HN, LinkedIn, podcasts, blog) are actually moving the needle. Weekly trend lets us attribute traffic spikes to specific posts.

Test plan

  • After merge, gh workflow run adoption-snapshot.yml and verify a PR is opened with rows appended to all three CSVs
  • Confirm referrer CSV has one row per (repo, rank) and that repos with no referrers (rare) just contribute zero rows

Add a third CSV (.outreach/adoption-snapshots-referrers.csv, long format)
recording top-10 referrers per repo per snapshot. Today's seed shows
the why: across 16 repos in 14 days, only ~8 unique humans arrived
from any external content (HN: 2, LinkedIn: 1, Reddit: 1, runcycles.io: 4).

This is the single best signal for "is HN/blog/social actually working"
because clones are dominated by automation. Weekly trend lets us
attribute spikes to specific posts/threads instead of guessing.
@amavashev amavashev merged commit 06e06e4 into main May 3, 2026
5 checks passed
@amavashev amavashev deleted the ops/snapshot-referrers branch May 3, 2026 01:19
@amavashev amavashev restored the ops/snapshot-referrers branch May 3, 2026 20:16
@amavashev amavashev deleted the ops/snapshot-referrers branch May 3, 2026 20:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant