Skip to content

Resource

The real cost of network downtime

Most downtime calculations produce a number that is far too low, because they count idle staff and stop. The costs that matter arrive after the systems come back, and in some sectors they accrue on somebody else's meter while you wait.

What to include

Six components. The first is the one everyone counts. The rest are usually where the real number lives.

Components of downtime cost
ComponentHow to estimate
Idle staffNo, this is the one everyone countsAffected headcount multiplied by loaded hourly cost
Lost revenueSometimesTransactions or billable hours not completed
Recovery labourFrequentlyHours spent restoring and verifying
Reconciliation tailAlmost alwaysManual work re-entered after systems return
Contractual penaltiesAlmost alwaysMissed windows, service credits, demurrage
Reputation and churnNever quantified, occasionally the largestCustomers who quietly do not return

Where the tail comes from

The outage stops when systems return. The cost does not.

  • Orders taken on paper have to be entered, checked and reconciled
  • Claims and invoices delayed by days push revenue back by weeks
  • Appointments cancelled during the outage displace future capacity
  • Partners who received nothing over automated exchanges start calling
  • Staff work overtime clearing the backlog, which is a real cost nobody logs

Using the number properly

The point of the calculation is not to frighten anyone. It is to make a redundancy decision rational.

Once a business knows what four hours costs, questions like whether to buy a second internet circuit, a standby server or a shorter recovery objective stop being technical preferences and become straightforward arithmetic.

It also stops the opposite error. Where the number is genuinely small, paying for a one hour recovery objective on everything is waste, and the calculation gives you permission to accept a longer window on the systems that can tolerate it.

Common questions

Answered before you ask.

Is there a standard cost per hour figure we can use?

Published averages exist and they are close to useless for an individual business, because the range across sectors and sizes is enormous. A figure built from your own headcount, revenue per hour, contractual exposure and reconciliation burden is both more accurate and more persuasive internally than any industry average.

How do we estimate reputation cost?

You generally cannot with precision, and pretending otherwise weakens the rest of the analysis. The workable approach is to note it qualitatively alongside the quantified figures, and to be specific about which customers or contracts would be affected rather than assigning a speculative number.

Does this apply if we are mostly cloud based?

Yes, because the dependency moves rather than disappearing. A cloud business is exposed to its internet connection, its identity provider and its application vendors instead of to its own server. The calculation is the same. What changes is which single point of failure you are protecting against.

Start with the assessment, not the contract.

We document what you have, test whether your backups restore, and give you the findings in writing. Yours to keep either way.

CallFree assessment