What the Cloud Actually Means for a Small Business

The cloud is somebody else's computers, in a building you will never visit, rented by the hour. The company that owns them has bought a very large number, keeps them running, replaces them when they fail, and rents you a slice for as long as you want it.

In practice it reaches a small business in two forms. One is software you sign into instead of installing: email, accounts, payroll, the tool your sales team uses. The other is the raw material, where you rent a server or a database and put your own system on it. Most businesses use both without ever drawing a line between them.

Nothing about it is magic and nothing about it is new in kind. You already rent your premises and your broadband, and nobody calls that a strategy. This is the same arrangement applied to the equipment in the cupboard by the printer.

Which means the sensible questions are ordinary ones. What does it cost, what happens when it breaks, who has the keys, and how hard is it to leave.

The server in the cupboard, and its replacement in four years. That is the visible one, and it is usually the smallest part. What goes with it is the surrounding equipment that never appears in the conversation. The battery unit that keeps it alive during a power cut. The extra cooling in a room that was never designed for it. The spare disks somebody was sensible enough to buy, and the software licences that come with running your own hardware.

You also stop buying capacity for a peak. When you own hardware, you buy for the busiest hour of the busiest month of the busiest year you can imagine, and then run at a fraction of it for the rest of the time. Renting lets you pay closer to what you use, at least in principle. Collecting that benefit takes some deliberate work, and plenty of businesses never do it.

And you stop buying a certain kind of risk. The disk in the corner failing on a Friday. The air conditioning unit dying in May. The break in that takes the machine along with the laptops. None of those become impossible, but they stop being yours to solve at short notice.

What you do not stop buying is people who understand your systems. Renting the hardware does not rent the knowledge of how your business is wired together, and that is the part nobody warns you about.

Compute, which is machines by the hour. Storage, which is space by the amount you hold and by how often you read it. Network, which is mostly free coming in and charged going out. And subscriptions for the software your staff use, which are charged per person per month and go up when you hire.

The important change is not the total. It is the shape. A server was one large payment every few years, approved once, then forgotten about until the next one. Cloud is a smaller payment every month that nobody approves individually and that quietly moves.

Your accountant will care about that more than your IT supplier does. Capital spending and running costs are treated differently, they are budgeted differently, and they get scrutinised by different people. Have that conversation before the first invoice rather than after the fourth.

One practical point that catches people out. Almost everything in this world is priced per user per month, which makes headcount the biggest single driver of what you pay. Growing from a dozen staff to two dozen roughly doubles a large part of the bill, and no one ever budgets for that at the time of hiring.

Your data is still your responsibility. The platform keeps its own systems running. Whether somebody in your company deleted the wrong folder, or shared a file with the whole internet, or left with a copy of the customer list, is entirely yours. Most subscription platforms keep deleted items for a short window and no longer, and finding that out in month seven is an unpleasant morning.

Backups are still your responsibility for the same reason, and this is the most common gap we find. People assume that because the email is in the cloud, it is backed up. What exists is usually a recycle bin with a short life, which handles an accident noticed on Tuesday and does nothing at all for a mistake noticed in April.

Who has access still matters, and it matters more, not less. When everything lived on a machine in the office, an attacker had to get into the building or onto the network. Now a password and a login page are enough. Two step sign in on every account is the single change that does the most, and it costs nothing but a fortnight of grumbling.

And your staff still need to know how to use the thing. Every platform move that went badly in our experience went badly on the training, not on the technology.

This is the part nobody mentions in the sales meeting. With a server, the cost was settled at purchase and then went quiet for years. With rented services, the cost is a live number that responds to everything you do, and it only ever moves in one direction unless somebody is watching it.

It moves when you hire. It moves when a team turns on a feature that is priced separately. It moves when a system gets busier, when somebody spins up a copy for testing, when a storage bucket fills, when a supplier raises their subscription price at renewal.

Give the number an owner. In a small business that is often the finance manager working with whoever handles the technology, and the review takes half an hour a month. What you are looking for is the difference against last month and the reason for it. That habit is worth more than any tool you could buy.

It is worth being honest that this is a genuine cost of the model. Renting is more flexible and easier to start, and it asks for attention every month for as long as you use it.

When the systems were in the cupboard, a broadband failure was an inconvenience. Email stopped, the website was somebody else's problem, and the order system carried on. Once everything is rented, a broadband failure stops the business.

So the line becomes infrastructure rather than an office service. That usually means a second connection from a different provider, taking a different physical route into the building, with equipment that switches over automatically. For a small site, a mobile connection as the fallback is often enough and costs very little.

Test the failover, and test it during working hours with people trying to work. A backup line that has never carried real traffic has an unknown state, and the day you find out is the day you need it. Check that the important systems are actually usable on it rather than just reachable.

Ask about the speed in both directions too. Plenty of business connections are fast coming down and slow going up, which is fine for browsing and painful once your staff are uploading files and sitting in video calls all day.

People can work from anywhere without a clumsy remote connection into the office, which stopped being a nice extra some years ago and is now simply expected. Sites and home working become the same problem rather than two.

Recovery gets better, if you set it up. Restoring a business that runs on rented services is a very different exercise from restoring one that ran on a machine which is now under water or in the back of somebody's van. That improvement is real, and it is not automatic. It still needs backups you have tested.

Weekend upgrades mostly stop. Patching the underlying platform becomes somebody else's job, and while that does not remove all of it, it removes the part that used to eat Sundays.

And capacity arrives in an afternoon. When a new contract needs another system by the end of the month, you are no longer waiting six weeks for hardware to be delivered. For a growing business that speed is often worth more than any saving on the bill.

Move in an order that puts the risk at the end. Email and files first, because the tools are mature, the migration is well understood, and the benefit is immediate. Then the general purpose systems. The one that takes your orders or runs your production goes last, once your team has done this twice and knows what surprises look like.

Keep the old system until the new one has survived a month end. Month end is when the odd reports run, when the accountant needs something unusual, and when the parts nobody tested get exercised. Switching off the old system the week after go live saves a little money and removes your only way back.

Insist on owning the accounts. The subscriptions, the domain name, the cloud account itself should all be in your company's name with your people holding the top level access, whoever set them up for you. This is the single most common piece of damage we clear up for new clients, and it is entirely avoidable at the start.

And write down what you now depend on. Which supplier holds what data, what you pay, when the contract renews, and how you would get your information out. That page takes an hour to write and answers most of the difficult questions you will be asked in the next three years.