Where to Start When Developing XLAs

Beyond the SLA

The Service Level Agreement has been the dominant measure of IT service quality for three decades. Response times, resolution times, availability percentages — these metrics are familiar, measurable and widely reported. They are also, in isolation, a poor proxy for the quality of experience that users actually have with IT.

An IT organisation can meet every SLA and still be seen as poor by its users. Resolution within four hours is within target — but if the issue required three contacts to resolve and each interaction was frustrating, the user experience was poor regardless of what the metrics say.

Experience Level Agreements represent an attempt to close that gap — measuring the quality of experience directly rather than inferring it from technical metrics.

The Common Starting Point Problem

Most organisations that attempt to develop XLAs struggle with the starting point. The concept is appealing but the implementation is unclear. What should you measure? How do you measure it? How do you set meaningful targets? And how do you manage performance against a measure that is inherently more subjective than a resolution time?

A Practical Starting Point

The most effective starting point we have found is to work backwards from the moments that matter most to users — what CX practitioners call 'moments of truth'. These are the interactions with IT that users remember and that shape their overall perception of IT quality.

For most organisations, these moments cluster around a small number of scenarios: the experience of reporting an incident, the experience of requesting something from IT, and the experience of onboarding as a new employee. Starting with these three scenarios — understanding what good experience looks like in each, and how far current reality falls short — creates a manageable scope for early XLA development.

Measurement Without Surveys

One of the practical challenges of XLAs is measurement. Transaction surveys (the feedback request at the end of every ticket) are widely used but have significant limitations — low response rates, selection bias, and survey fatigue that makes the data increasingly unreliable over time.

More effective approaches combine targeted pulse surveys with behavioural indicators — self-service adoption rates, re-open rates, escalation patterns — that reveal experience quality without requiring users to fill in forms. Building a measurement approach that triangulates across these sources gives a richer and more reliable picture than any single metric.