It stays up, and you find out before your customers do

In accounts held in your company’s name, with your finance team paying the bill directly, on whichever provider your team can live with. Beyond that the honest answer is usually smaller than people expect: managed services that someone else patches, rather than machines you have to look after. The decision turns on where your customers are, what must stay inside India for compliance, and what your team can support at three in the morning. We will tell you when the answer is a modest server rather than a platform.

It is common, and it is fixable. Bills rise when nothing is labelled, when test systems are left running, when storage is never cleared, and when something was sized for a launch that never came. The first step is making every line traceable to a system and an owner, which usually removes a visible share of the bill in the first pass. After that the cost is reviewed monthly against what we said it would be, and a jump raises an alert rather than a surprise at the end of the year.

An alert reaches a named person, on a phone, with the thing that broke and what to do about it written beside it. That is the difference between a system that is watched and one that is merely hoped for. Which failures are worth waking someone for is agreed with you, because an alert that fires for everything is an alert nobody reads. During an engagement a founder is on that list; at handover your team is, with the written what-to-do in your own repository, and a monthly care plan is available if you would rather we stayed on it.

Because one is restored in front of you, on a timetable, and the time it took is written down. A backup nobody has ever restored is a belief, not a protection: the usual discoveries are that one database was never included, that the restore takes far longer than anyone assumed, or that nobody has the credential needed to start it. We agree two plain numbers with you, how much data you could afford to lose and how long you could afford to be down, then test against them rather than against a promise.

Usually yes, and the first piece of work is finding out what is actually there: which accounts exist, who holds the keys, what is running that nobody remembers starting. You get that inventory as a document whether or not you continue with us. Where the previous builder left things undocumented we write them down as we go, and where something is genuinely beyond saving we say so plainly, with what replacing it would involve, rather than charging to keep it upright indefinitely.

For nearly every business this size, renting is enough and owning is a distraction: your own machines mean someone in your building is responsible for power, cooling, spare parts and the patching nobody enjoys. Buying earns its place in narrow cases, such as a steady heavy load where the rental price never falls, or a rule that says data must sit on hardware you control. We work that out with real numbers at your volumes before anyone signs anything, and the comparison is written down.