The 99.9% Dispute: One Word, Three Meanings
Northwind Mutual, a 1,200-employee insurer, is launching a claims portal. The business sponsor signed an SLA promising '99.9% availability, 24/7.' Operations measures availability excluding a four-hour Sunday maintenance window; the vendor's contract lists that window as 'planned downtime.' In a steering meeting the sponsor cites two Sunday customer complaints and declares the SLA breached. The infrastructure lead says the portal was 'up' because the maintenance was scheduled and communicated. The service desk uses P1-P4; the vendor uses Sev-A-Sev-D with different thresholds. Vendor Sev-A requires a total outage, while a service-desk P1 includes any single claims adjuster unable to work. Two incidents last month were logged as P2 internally and Sev-B by the vendor. The CIO wants one glossary before the next board report.
1. Identify the distinct terminology conflicts in this situation and explain the operational or contractual risk each one creates. Then propose precise definitions the CIO could adopt.
Read the worked answer
Conflicts and risks
Availability vs. uptime vs. 'up.' The sponsor treats any user-visible failure during Sunday as an availability breach; operations treats the service as 'up' because the process was running and the outage was planned. Risk: the same event is reported two ways to the board, SLA credits are disputed, and customers get inconsistent explanations.
Planned downtime exclusion. The vendor's contract excludes the Sunday window, but the business SLA does not say so explicitly. Risk: the sponsor believes 24/7 means no exclusions; the vendor believes the window is free. Contract and SLA are misaligned.
Severity taxonomy. P1 includes a single adjuster unable to work; Sev-A requires a total outage. Risk: escalation is delayed because the vendor does not treat a P1 as Sev-A, and incident data cannot be compared across parties.
Definitions to adopt
- Availability: the percentage of agreed service hours in which the portal is usable by intended users for its intended functions, measured at an agreed monitoring point and validated by synthetic transactions. Do not use 'up' as a formal term.
- Service hours: 24/7 except maintenance windows listed in the SLA schedule, capped in duration, notified at least five business days ahead, and approved by the service owner.
- Severity: a single cross-party scale based on business impact, not on whether a total outage exists. Include impact dimensions such as number of users affected, workaround availability, financial or regulatory impact, and patient/customer harm where relevant. Publish a mapping table so P1 corresponds to Sev-A, P2 to Sev-B, and so on.
Reasoning: the root problem is not that one party is wrong; it is that the same words carry different decision rules. Definitions must be measurable, tied to monitoring, and embedded in the contract and tooling, or the dispute will recur.
2. Design a governance process for the glossary so future disputes are resolved without escalating to the CIO.
Read the worked answer
Governance design
Single owner and decision rights. Assign the glossary to a service-management office or equivalent, with a small cross-functional board: business service owner, operations lead, service desk lead, vendor manager, and a security/compliance representative where relevant. The board owns definition changes; the CIO owns only unresolved escalations.
Version control and change control. Each term has an owner, definition, measurement method, examples, anti-examples, and effective date. Changes require impact assessment on SLAs, contracts, reports, and tooling, then approval by the board. Publish a change log.
Contract alignment. Reference glossary terms in SLA schedules and vendor contracts. Where a vendor uses a different taxonomy, require a mapping table rather than forcing the vendor to rename everything.
Operational embedding. Load severity mappings into the service desk and monitoring tools. Include glossary terms in reporting templates and vendor scorecards. Train service desk and vendor staff on the impact-based severity rules.
Dispute resolution path. First, the term owner applies the definition; second, the board decides within an agreed SLA; third, the CIO or contract owner decides. Record the decision and update the glossary.
Reasoning: a glossary is only useful if it changes behavior in incidents, reporting, and contracts. Governance must connect definitions to decision rights and tooling, not just publish a document.