Hosting decisions get made on monthly price because the alternative costs are invisible. They are calculable, and once calculated they usually change the decision.
What an hour offline costs you
For a business where the site takes orders, it is direct: average revenue per hour during business hours, times the hours lost. Add the orders that do not come back later.
For a business where the site generates enquiries rather than sales, work it through the funnel: enquiries per hour, conversion rate, average value. Less precise, still far better than treating it as zero.
Then add what does not appear in revenue — staff unable to work, support time absorbing complaints, and the customers who quietly went to a competitor. That last one is real and uncountable; acknowledge it rather than pretending it is zero.
Write the number down. Every subsequent hosting decision compares against it.
Uptime percentages, in minutes
Availability figures are designed to look reassuring. Converted to time they are more concrete — the difference between two nines and three nines is hours per month, not a rounding error.
Two things to check in any commitment:
- What is excluded. Scheduled maintenance is typically outside the calculation, so advertised availability is availability excluding planned downtime.
- What the remedy is. Usually a credit proportional to the outage — which, against the cost of the outage itself, is close to symbolic. Treat a commitment as a statement of intent rather than as compensation.
Managed against unmanaged, honestly
The monthly difference is visible. What it buys is not, until something breaks.
Estimate annual hours your team would spend on: OS and software updates, responding to security advisories, monitoring and responding to alerts, verifying backups, and handling incidents including out of hours. Multiply by a loaded hourly cost.
For most small teams this exceeds the managed premium, and that is before valuing the incidents that do not happen because someone was paying attention. The calculation flips only where you already employ someone who does this work and has capacity.
Speed is a revenue metric
Page load time affects conversion measurably. If you are choosing between tiers, measure your own site on each rather than comparing specifications — a faster host that improves load time meaningfully can pay for itself through conversion alone.
Measure at peak, not at a quiet hour. Shared hosting in particular performs differently under neighbour load.
Backups: value the restore, not the backup
The question is not whether backups exist. It is how long a restore takes and how much data is lost.
Multiply your hourly downtime cost by realistic restore time. A host that restores in an hour versus one that takes a day is a large difference in exposure, and it is worth asking about explicitly — including whether restore is a service they perform or something you do yourself.
The annual review
- Actual downtime last year, in minutes, and its cost at your figure
- Hours your team spent on hosting administration
- Whether a restore was tested, and how long it took
- Whether current resource usage justifies the current tier — in either direction
The fourth question saves money as often as it spends it. Workloads shrink, and nobody downsizes without being prompted to look.
About the author. Kislay Chandan has worked with schools since 2014. Rather than gathering requirements from the outside, he worked inside a school to see what actually goes wrong day to day, and built what he found into the software. He still works with Shemrock Glitters in Patna — most changes to the product start with something that came up there.
Gallery
