whatisrpm.com

Revenue wing / head-to-head

Mediavine Journey vs Ezoic

Reviewed10·2026

For years this was the small-site fork: Mediavine's on-ramp or Ezoic's open platform. In 2026 the two moved in opposite directions. Journey dropped to 1,000 tier-1 sessions; Ezoic raised new sites to 250,000 monthly active users. The entry terms are published; the uplift numbers, as usual, are not.

By Oliver Wakefield-Smith. Oliver Wakefield-Smith runs a portfolio of ad-monetized publisher sites and reads these dashboards weekly.

Entry terms, side by side

Journey vs Ezoic published terms, checked 2026-10-21
DimensionJourneyEzoic
Entry (new sites)1,000 tier-1 sessions in 30 days Mediavine2026-10250,000+ monthly active users Ezoic Support2026-10
Below the barn/a at this scaleIncubator Program, 20 publishers a month Ezoic Support2026-10
Existing sitesn/aOn Ezoic before 19 Feb 2026: grandfathered Ezoic Support2026-10
Published revenue shareNot stated on the Mediavine pages checked not publishedNo flat % printed not published
Reporting basisSession RPMEPMV Ezoic Support2026-10
Growth pathGraduate to full Mediavine at $5,000 annual ad revenueService tiers from 250K and 1M+ monthly users

Who actually has a choice

A new site under 250,000 monthly active users can join Journey on 1,000 tier-1 sessions Mediavine2026-10, but can only reach Ezoic through the Incubator, which takes 20 publishers a month and does not guarantee acceptance Ezoic Support2026-10. So for most small sites the realistic fork is now AdSense or Journey, with Raptive joining at 25,000 pageviews.

The real Journey-vs-Ezoic decision belongs to sites already on Ezoic. They are grandfathered, but removing the Ezoic integration for more than 7 days voids that status Ezoic Support2026-10. A trial of Journey that runs longer than a week is therefore a one-way door for a grandfathered site under 250,000 users: weigh it as a switch, not a test.

The unpublished part, named

Neither side prints a revenue-share percentage for these programs not published. Head-to-head RPM claims between them are therefore untestable from public data. What to do instead: keep a dated earnings-per-session record on the one you run, and model the alternative with the what-if dial before switching. The published entry rules at least make the eligibility half of the decision clean; the network table keeps all eight in view.