A review of what you pay for
A list of subscriptions, accounts and licences with the unused ones marked. Often that alone brings the bill down, with no migration at all.
Moving a service to the cloud does not change what it does, only what you pay for and who answers when it stops working. Instead of a one off purchase there is a bill every month, instead of your own server room somebody elses, and instead of your outage somebody elses outage that you cannot influence. Sometimes that is a fine deal and sometimes it is not, and that is exactly what we work out before a migration rather than after.
We are not a party to this argument and we earn no commission on what you choose. We sell the work of rolling it out and looking after it, and that work looks much the same on either side. So a quote shows you both routes side by side, together with the cost of your own server, and the choice is yours.
Migrating everything at once is the most expensive scenario there is. It almost always pays to split the company into three groups and treat each differently.
The easiest decision, because a mail server of your own stopped paying for itself long ago once you count the work of filtering spam. Microsoft 365 or Google Workspace close the subject, and the difference between them usually comes down to what your people already know how to use.
Here the maker has to be asked first, because many ERP and trade specific programs still need a local database or run in the cloud only in the makers own version, on the makers terms. It is the maker, not us, who sets the limit of what may be moved.
Design files, video material and scanned documents can stay put, because working on them across the internet is often slower than off a disk next to the desk. The sensible compromise is working locally with a copy in the cloud, not moving everything for the sake of it.
Most often it ends with part staying in the company and part going to the cloud. That is not indecision, it is the normal state of affairs, and in this trade it is called a hybrid environment. Keeping one costs slightly more than either side on its own, and we say so plainly rather than promising that everything can be closed into one place. One profession has an extra criterion here: at a law firm a demand to produce data may go straight to the supplier, who will not invoke professional privilege. We set it out under IT for law firms. Construction has its own criterion here: documentation has to open from site on a weak signal, which tips the choice towards the cloud more firmly than in an office. We return to it under IT for construction companies.
The cloud rarely grows dearer through price rises. Most often it grows because you pay for things nobody watches, since they do not appear on the invoice separately.
A server in the cloud is paid for by the hours it is switched on, not the hours anybody uses it. With a steady, predictable load around the clock your own machine can simply be cheaper, because there the spare capacity does not cost anything each month.
Licences for people who have left can run for years, and packages bought for one need stay for good. This is the simplest saving in all of IT and usually the first thing we find when reviewing subscriptions.
Putting data into the cloud is usually free, taking it out is not. That asymmetry decides the cost of recovering after an incident and the cost of any future move, so we ask about it before anything is migrated.
Honestly: for a small company of a dozen or so people the cloud nearly always wins, because your own server, its licences and the care around it spread across too few people. The point where that reverses depends on how steady the load is, not on the number of staff. If the bill is already growing and you want to understand it first, we start with a review of IT costs, not with a migration. Compute workloads are a calculation of their own, where differences between providers can be several-fold, which we set out under IT for a software house.
The gravest risk in the cloud is not an outage, it is the moment when leaving is no longer possible. European law now governs this and it is worth using.
From 12 January 2027 a cloud provider may not charge you anything for moving you to somebody else, including a charge for taking your data out. Until that day the most it may charge is what the move actually costs it.
Notice to start the move is at most two months, the move itself to another provider or onto your own server at most thirty days, and you then have at least another thirty days to retrieve the data. If that is technically unfeasible the provider has fourteen working days to explain and to name another date, no later than seven months.
A penalty for ending the contract early is still allowed and it is often the real cost of leaving. Dropping the charges covers the move itself; it does not release you from paying for the unused period you agreed to.
What to do with this in practice: before signing, check those three deadlines and the size of the early termination penalty, and for services already running ask the provider to confirm in what format it hands the data back. If the answer is in its own format, it will have to be rewritten anyway, and the law will not do that for you. The basis: Regulation 2023/2854 of the European Parliament and of the Council, known as the Data Act, whose switching provisions have applied since 12 September 2025.
Six areas that a migration has to close. We usually take several at once, but we start with the first.
A list of subscriptions, accounts and licences with the unused ones marked. Often that alone brings the bill down, with no migration at all.
Moving mailboxes, calendars and files with their history and permissions intact, and then putting order into who has access to what.
Part of the services in the cloud, part on site, with one sign in and consistent permissions on both sides. We also work out whether your own machine comes out cheaper.
The provider looks after its own infrastructure, but your data is your responsibility. A separate copy of mail and files guards against deletion and against losing access to the account.
In the cloud the password is the only boundary, because the service can be reached from anywhere in the world. A second sign in factor and access rules are compulsory here, not optional.
Matching plans to what people actually use and watching renewal dates, so that nothing renews itself automatically once it has stopped being needed.
What we do not do: move a system to the cloud when its maker never intended it, because that ends in a program running with a delay and a loss of support. Nor do we promise that migrating will lower the bill until we have seen what you pay today. Sometimes the sums say that staying where you are is cheaper, and then that is what we say.
The most important part happens before the move and the most visible one just after. So the first week after the switch is planned separately.
We go through subscriptions, accounts and systems, and ask the makers of your trade software what may be moved. Without that answer there is no point planning the rest.
We set the cost of the present setup against the target one over the same period, with licences, labour and the cost of recovering after an incident. You get a table, not a recommendation on trust.
One department first as a trial, then the rest. Switchovers run in the evenings and at weekends, and the old environment stays usable until the new one has worked several days without a hitch.
That is when most requests come in, and small ones: a missing shortcut, a different way of sharing a file, a password to fix. We set time aside for it in advance, because that week decides whether people call the change a success.
A law firm asks where the data sits, a factory asks what happens when the line goes down. The answers differ, which is why the conversation starts with the industry.
The pages that come back when the cloud is discussed.
For a small company usually yes, because the cost of a server, its licences and its care spreads across too few people. With a steady load running around the clock it can be the other way round, because in the cloud you pay for the hours it is on, not for actual use. So we cost both routes over three years and show a table rather than answering in general terms.
With the largest providers the European Union can now be named as the place of storage and processing. Microsoft completed such an arrangement for its services in February 2025, covering data from technical support cases as well, although some side services and the traffic needed to run them still leave that area. For services under data protection law or trade requirements we check this for the specific plan, not at the level of a slogan about a European cloud.
The company stops working, and that is the real price of moving to the cloud. So during a migration we discuss a backup link, usually on a SIM card, which switches over by itself and carries mail and basic work. With a single link and a system critical to production we advise against moving everything.
It looks after its own infrastructure, so the data survives a failure of its server, but it does not survive deletion by a member of staff, a dispute over payment or a hijacked account. The bin inside the service holds files for a limited time and that is not a backup. A separate copy of cloud mail and files is cheap insurance and we recommend it with every migration.
Yes, and it is better protected now than it used to be. From 12 January 2027 a provider may not charge for moving you elsewhere or for taking your data out, and the contract has to allow the move within thirty days. A penalty for ending the contract early still applies, so the real cost of leaving depends mostly on the term you tied yourself to.
Both cover mail, documents and meetings to a comparable standard, so the choice is rarely settled on features. What decides is what your people already work in, what your customers use, and whether you run programs that need Office installed on the computer. On an even result we look at the price of the plan that actually applies to you.
In a company of up to a few dozen people usually two to eight weeks, and most of that is copying data in the background rather than downtime. The mailbox switchover itself we do in an evening or at a weekend, and in the morning people find their mail in place. It takes longer when permissions to files have to be tidied up along the way, which with old network drives is the rule.
The first meeting takes 30 minutes and commits you to nothing. We will tell you straight whether we can help and what we would propose.
2019 - 2024 Interactive Workspace Ltd. - All rights reserved.