Amey-Thakur Support SLA management
An SLA is a promise with consequences. The common failure is promising what looks competitive rather than what capacity supports, then missing it, which is worse than a modest promise kept.
Method
- Derive targets from measured capacity. Current performance at the percentile you intend to hit, not the average and not an aspiration.
- Commit to response before resolution. Response time is within your control; resolution depends on the problem, and promising it invites breach on the hard cases.
- Define the clock precisely. When it starts, whether it pauses awaiting customer reply, and which hours count. Ambiguity here is where disputes come from.
- Tier by severity with different targets. One target across all ticket types either over-serves the trivial or under-serves the critical (see support-ticket-triage).
- Alert before breach, not after. A ticket approaching its target needs intervention while it can still be met.
- Report honestly, including misses. Selective reporting is found eventually and costs more than the misses would have.
- Staff for the peak you accept. An SLA met only at average volume is not met, and queues arrive in bursts (see capacity-planning).
Boundaries
Contractual SLAs carry financial and legal consequences and need review before commitment (see agent-legal-desk). Meeting an SLA is not satisfying a customer; a fast unhelpful reply meets the target and fails the person. Targets shape behaviour and can crowd out quality if measured alone.