Nuage Solutions All articles
Cloud Strategy

Counting the Minutes You Can Prove: A Service-Credit Readiness Routine for UK Cloud Buyers

Nuage Solutions
Counting the Minutes You Can Prove: A Service-Credit Readiness Routine for UK Cloud Buyers

Photo: CutG, CC BY-SA 4.0, via Wikimedia Commons

Most organisations read their service-level agreement in the week after an outage, when the status page has turned green and the operational noise has drowned the commercial question. By then, the claim window may already be running. A service credit is not compensation for inconvenience; it is a defined remedy that must be requested, evidenced and timestamped. The work therefore starts before the incident, not after it.

What the SLA actually commits to

An availability figure is only as meaningful as its measurement rules. A promise of "99.9% monthly uptime" can be calculated over a calendar month, a billing month or a rolling thirty-day period, and each produces a different answer for the same incident. The measurement point matters just as much: availability observed at the edge of the provider's network is not the same as availability experienced at your firewall, your identity provider or your staff's homes.

Definitions of "unavailable" are usually narrower than customers assume. Partial degradation, elevated latency, intermittent errors below a stated error-rate threshold, and loss of a single feature may all fall outside the definition. Credits typically scale in bands, expressed as a percentage of the monthly fee for the affected service rather than the whole contract value. Read the band table alongside the service description, because the fee base is frequently the subject of disagreement.

The exclusions that quietly close the door

The reasons a claim fails are rarely dramatic. They are administrative, and they are written into the schedule. Common exclusions include scheduled maintenance notified in advance, emergency maintenance, customer misconfiguration, use beyond documented limits, failures in third-party networks outside the provider's reasonable control, preview or beta features, and suspension for non-payment.

Two procedural traps account for a large share of unclaimed credits. The first is the notice window: many contracts require a claim within thirty days of the incident, in a specified format, sent to a specified address. A request made to an account manager in a weekly call may not constitute valid notice. The second is the "sole remedy" clause, which limits the provider's liability for availability to the credit itself. That does not make the credit optional, but it does mean the claim has to be made properly rather than negotiated informally.

Three functions, one evidence file

A defensible claim is a joint product of three teams, and each holds a piece that the others cannot supply.

When these three sets are kept separately, the claim becomes a reconstruction exercise under time pressure. When they are drawn together at the point of incident, the claim becomes a matter of assembly.

A pre-outage checklist

Build the following once, then use it each time an incident occurs.

  1. Create a single availability evidence folder with a named owner in IT and a named deputy.
  2. Obtain written confirmation from the provider of the contractual notice address, the claim template and the claim window.
  3. Configure monitoring to produce a timestamped availability report per service, retained for the contract term plus six years.
  4. Adopt a one-page incident log: detected at, resolved at, services affected, users affected, business process affected, workaround used, estimated business cost.
  5. Identify a finance contact who can confirm the monthly fee for the affected service within two working days.
  6. Set a calendar reminder for day twenty of every thirty-day claim window, listing any incident still open.

The quarterly review

Set aside one session each quarter to test the machinery rather than the provider. Review incidents from the period and ask, for each, whether a credit was owed and whether one was claimed. Check that monitoring coverage still matches the measurement points named in the contract, particularly after any architecture change. Confirm whether amendments have altered the exclusions or the fee base. Reconcile the claim register against finance's invoice records, so that credits granted are visibly applied rather than quietly absorbed.

Finish by walking one historical incident end to end as a tabletop exercise: could you produce the timeline, the availability export, the fee calculation and the correct contract reference within the window? If any of those four is missing, the gap is in your records, not in the provider's service.

A claim is a records exercise

Cloud spend is under closer board scrutiny than it once was, and unclaimed credits are a measurable leakage in that spend. The organisations that recover what their contracts allow are rarely the ones with the most aggressive procurement posture; they are the ones whose IT, finance and procurement teams share a single, pre-agreed evidence file. Treat that file as part of the service you are buying. The status page will turn red again, and when it does, the question will not be whether you were inconvenienced, but whether you can prove it.

All Articles

Related Articles

The Billing Trail Is Not an Asset Register: Owning Every Cloud Subscription in a UK Estate

The Billing Trail Is Not an Asset Register: Owning Every Cloud Subscription in a UK Estate

Before the Request Arrives: Rehearsing Subject Access Retrieval Across a UK Cloud Estate

Before the Request Arrives: Rehearsing Subject Access Retrieval Across a UK Cloud Estate

A Map You Can Audit: Evidencing Data Residency Across a UK Cloud Estate

A Map You Can Audit: Evidencing Data Residency Across a UK Cloud Estate