Most cloud projects do not fail because the technology is wrong. They fail because the business was never ready for the change in habits — and nobody planned for the week when everything feels worse.
Decide what you are actually solving
Cloud is a means, not a goal. Write down the problem first, because it determines the whole plan:
| If the problem is… | You probably want… |
|---|---|
| Files on a server only reachable in the office | File storage with sync |
| No way to work together on a document | Real-time collaboration |
| Nothing works when the internet is down | A hybrid setup with local cache |
| We cannot hire outside Metro Manila | Remote-first tooling |
| Compliance asks where the data lives | A region-specific plan, documented |
Start with identity, not storage
This is the step people skip and later regret. Get your accounts, groups and sign-in policy right before moving data. Migration becomes dramatically simpler when access is already organised, because permissions follow the person rather than the folder.
- One account per person. No shared logins — you cannot audit a shared password.
- Multi-factor authentication on everyone, no exceptions for executives.
- Groups that match how the business actually works, not an org chart from 2019.
- A documented offboarding checklist.
Migrate in three waves
Do not try to move everything in one weekend.
- Wave 1 — the willing. A small team already comfortable with change. They surface the real problems while the audience is small.
- Wave 2 — the majority. Apply what wave 1 taught you. This is where training materials earn their keep.
- Wave 3 — the holdouts. Give them a named person and a date, and keep the old system read-only for a defined period.
Plan the bad-internet day
If you are in the Philippines, this is not theoretical. Brownouts and ISP outages happen. Decide in advance what still works:
- Which files are cached locally on each laptop.
- What can be done on a phone hotspot.
- Who keeps a printed list of emergency contacts and account recovery numbers — not stored only in the system that is down.
The one thing that makes or breaks it: training that respects people’s time
A single sixty-minute session is not training. People need to hear it three times, in three formats: a live walkthrough, a one-page written guide, and short videos they can replay at the moment they are stuck.
Nobody reads the manual. They search for the answer at the moment they are blocked. Put the answer there.
Set a definition of done
Write down what success looks like, with dates, so the project can actually end:
- Everyone signs in with their own account and MFA.
- A file shared with an outsider works, and is revocable.
- The team has completed one full week with no use of the old system.
- You have restored a deleted file as a test, and know how long it took.
- Offboarding a leaver is a checklist, not a scramble.
Planning a move and not sure where the traps are? Talk to a specialist — we plan around how your team actually works.