Cloud migration for small business: what actually matters

Cloud migration is the rare technology topic where both the evangelists and the sceptics are wrong. The evangelists promise transformation to businesses that mainly need their files to be reachable and their backups to be real. The sceptics point to failed migrations and conclude the old server was fine. Both miss the actual decision. After running and rescuing cloud projects for businesses from ten to a few hundred staff, our view is that the question is never "cloud or not cloud" — it is "which parts of this business genuinely benefit, and at what cost of disruption." Five questions answer that.

The five questions

1. Where do your people actually work? If the answer is "one site, one office, everyone at a desk", the cloud's headline benefit — access from anywhere — buys you little. If supervisors work from sites, staff from home some days, or the owner from anywhere the business calls them, the case starts building itself.
2. What would a day without your server cost you? The resilience argument is about more than backups: it is how long the business operates without its files, and whether the recovery actually works. An on-site server with an untested backup is one bad afternoon from a very bad month — and we test exactly this during every assessment.
3. Is the data clean enough to move? Migration is a forcing function: eleven years of files get inventoried, classified and pruned before the move, because moving everything is wasteful and moving the wrong things is risky. Businesses are often surprised that the migration's biggest payoff is the cleanup, not the technology.
4. What does the total cost look like over three years? Not the licence comparison the vendors lead with. The full model: migration labour, training, the productivity dip during cutover, ongoing subscription growth, and the avoided costs — hardware refresh, the backup regime, the person who quietly maintains it all.
5. Can the change be absorbed? The honest failure mode of cloud projects is not technical; it is that the team never adopts the new way, and the business runs two systems for a year. If there is no capacity for training and a stabilisation period, the timing is wrong — and saying so is advice, not defeat.

What the move should look like when it is done properly

A phased migration is the standard we hold: data classified and cleaned first (the labour-intensive part nobody markets); a pilot group cut over while the old system remains read-only; department-by-department moves, each rehearsed on a copy; a stabilisation month with the rollback path available; and decommissioning of the old system only after the new one has earned trust. Done this way, "we moved to the cloud and nothing happened" is the success story — production kept running, nobody lost a file, and the only difference anyone notices is that the files are reachable from wherever the work is.

Where our opinion annoys both camps

We routinely advise businesses not to migrate something. The specialist production system that works and has no cloud equivalent worth the disruption; the archive data that is rarely touched and can simply be archived properly on-site; the tool the team loves that is cheaper left alone. And we equally frustrate sceptics by insisting that certain things — anything where a tested backup does not exist, or where the business's continuity depends on one ageing machine — should move (or at minimum be made recoverable) regardless of enthusiasm. The cloud is not a religion. It is a delivery method, and parts of your business deserve it while others do not.

If you are weighing a move, the useful first step is an honest assessment: an inventory of what you run, the five questions answered with your real numbers, and a costed recommendation — including the parts we advise you to leave alone.

The benefits nobody prices into the business case

The business case for a cloud move is usually argued on licences versus hardware, and that comparison flatters neither side honestly. The benefits that actually decide satisfaction in hindsight are quieter ones that rarely appear in a quotation. Onboarding new staff becomes trivial — access is granted in minutes rather than an IT visit, and for businesses that hire seasonally this alone changes operations. The backup stops being a hope — tested, off-site recovery replaces the drive nobody has restored since installation, and we consider that benefit worth the move by itself. The key-person risk around "the server room" disappears — knowledge of the infrastructure is documented and shared instead of held by whoever set it up. And security posture generally improves, because reputable cloud providers patch, monitor and certify at a standard a small business cannot sustain alone. We put these in every business case we write — not because they are glamorous, but because they are the reasons clients tell us they would never go back.