Trello Implementation: 30/60/90 Rollout That Sticks

Write three outcomes Trello must change — status visibility, owner accountability, or time-to-update — and who owns each. 1. Name the job: lightweight kanban boards. Explicitly exclude a portfolio PMO or an engineering issue tracker.

1. Pick three measurable rituals (example: Friday status lives in Trello, not Slack).
2. Name the admin owner and the weekly users.
3. List nice-to-haves you will refuse until day 61. Worked example: Harbor Marketing (a small campaign crew) freeze “replace the status spreadsheet” as the only 90-day outcome. Portfolio dashboards wait.

Configure one board, list names that match how you actually work, and members who will move cards. Success looks like: create one board, move three cards across lists, and prove comments replace a side chat for that work. 1. One live container of work with real owners.

1. No marketplace apps except the connectors the loop needs. Research names Zapier on the Trello side — confirm the connectors your work loop depends on.
2. A manager can see status without asking for a screenshot.
3. Log every “can we also…” request for day 61. Worked example: Harbor Marketing (a small campaign crew) delays optional AI and extra hubs until the core loop has a week of real use. Week-one marketplace apps are treated as a failure mode.

Train the people who must update Trello every week — not a one-time all-hands. Our snapshot flags a trial on Standard, Premium, and Enterprise without a published length — confirm the window on the Trello pricing page. 1. Role-based sessions: contributors vs managers vs guests.

1. Include one sceptic so adoption risk shows up before speeches.
2. Practice a contributor moves a card and a lead sees it without a Trello admin live — no slideware-only training.
3. Capture every question as a training-cost line, not a feature request. Worked example: Harbor Marketing (a small campaign crew) includes one sceptic user in training so “I still ping the owner in Slack” shows up while you can still fix the ritual.

Check whether the core loop is actually used. Only then add automations, extra hubs, or AI. 1. Count weekly active updaters vs invited seats.

1. Confirm managers stopped asking for spreadsheet status.
2. Review watch-outs. Research flags: Shallow planning depth; Outgrown by complex work; Reporting weak.
3. Unlock one extra: automations or a dashboard — not both in the same week. Worked example: Harbor Marketing (a small campaign crew) reviews whether create one board, move three cards across lists, and prove comments replace a side chat for that work happened for four consecutive weeks before expanding scope. Empty dashboards mean the rollout is not done.