When I planned Clienzo's networks, I estimated the work as code: an OAuth flow, an upload endpoint, error handling. That part was usually quick. What actually decided when a network could launch was the platform's review.
Reviews gate the features people care about
On most platforms an app can do almost anything in development mode — for its own developers' accounts. Doing the same for customers needs permissions the platform grants after a review. And the gated permissions are exactly the valuable ones: posting to someone else's page, reading comments, reading post numbers.
What each one asked for
- Meta (Facebook, Instagram, Threads): a written use case and a screencast per permission, in two separate apps, in two rounds — publishing first, comments and insights later. Plus one permission the dashboard needed at a level its own API did not show (part 6).
- Google / YouTube: OAuth verification with a demo video, a reviewer login and a privacy policy that matches the code — and then a separate API audit before uploads can be public (part 7).
- Pinterest: trial access that creates pins only the author can see, until Standard access is approved (part 8).
- Tumblr: no review to publish, but a shared rate limit for the whole app; raising it means telling them who you are and what traffic to expect.
- Vimeo: connecting works; uploads wait for a manual approval that is still pending.
How I plan around it now
- Submit as soon as the feature is demonstrable, not when it is polished. The review clock is the long pole.
- Record the screencast while building, with a clean test account ready for reviewers — one that works without a paid plan.
- Make the privacy policy match the code, line by line, before submitting.
- Show "In review" honestly in the product and on the website, instead of a tick. A network is called live only after a real post has gone through it.
- Keep a list of what each platform declared where. I will not remember which permissions live in a dashboard and which in code.
Next in this series: when the blocker is not code or review, but paperwork.
Originally published on jonasnguyen.com, where I write about building Clienzo and the next startups.
Top comments (0)