I typed thirty Plausible API keys before I admitted there was a problem.
Each one took twenty seconds. Find the key in Plausible, switch tab, paste, save. Twenty seconds × thirty sites = ten minutes. Not terrible. Until I did it again two days later, because I’d added a typo somewhere and four clients had been silently showing zero traffic for a week. Then it’s twenty minutes. Then I migrated to a self-hosted Plausible box and had to type a different key into twenty of them. Then I was three months in and had lost an afternoon I couldn’t remember authorising.
The shape was wrong. The API key wasn’t really a property of the site — it was a property of me. One credential, granting access to all my agency’s Plausible data. SendTidings was making me write it down on every site like I was filling in a registration form at a different doctor’s surgery, every visit, forever.
What “connect once” actually means
So we moved every integration to the org level. Plausible. Matomo. Google Analytics. Search Console. UptimeRobot. Every one of them, you connect once. Add a new client site three months later and it inherits the connection. You don’t see the credential again. You don’t paste anything. You don’t even think about it.
Connect once, send forever. The tagline came out of the migration, not the other way around.
It sounds obvious in retrospect. Most useful things do. But per-site configuration is the SaaS default — every form on every dashboard asks you the same question on every record. The default is wrong. The default treats each record as if it might need a different answer, when in practice 99% of the time the answer is the same.
The product’s job is to know what’s the same and ask you only once.
Where the wrinkle lives
There’s an honest caveat worth flagging for early users. When you reconnect a Google account at the org level, the new refresh token cascades down to every site that was using the old token — but only if the old token matches exactly. If a site was on a slightly-different copy of the credential (which happened more than I’d like in the early days), it needs reconnecting individually.
We’re refactoring the schema so it can’t get into that state. Until then, if a reconnect doesn’t quite fix everything, click into the affected site and re-link. Tell me about it — it shouldn’t be happening, and every report of it makes the case for finishing the refactor.
Why this matters more than it sounds
Quietness is a feature. The product is delightfully boring on purpose. Reports get sent on the 1st. Credentials get refreshed. Reports land in clients’ inboxes. Nobody needs to type anything.
The agencies I respect most don’t talk about their tools — they just talk about their work. SendTidings’s job is to stay invisible enough to enable that. Per-site configuration would make the product visible every time you added a client, which is exactly the wrong moment to be reminded that your tool exists.
So if you’re new to SendTidings and you’ve found the integrations page — congratulations, you’ve found the most important screen in the product. Set it once. Then never visit it again unless something breaks.
That’s the promise.
