XLA metrics explained: CSAT, FCR, sentiment, and customer effort.
An Experience Level Agreement is only as good as the metrics inside it. Here is what each component measures, why it carries the weight it does, and how the six combine into a single experience score.
- The XLA composite Simetrix uses combines six metrics: CSAT, FCR, sentiment, NPS, resolution quality, and customer effort.
- Each metric is measured directly and weighted by how closely it maps to retention.
- No single metric is enough. Each has a blind spot that the others are there to cover.
- Sentiment is weighted heavily because it reads the customers a survey never reaches.
- Simetrix scores all six across 100% of interactions, not a sample.
The XLA composite at a glance
An XLA is a weighted composite. Each component below is measured directly, and the weighting reflects how closely that signal maps to whether a customer stays. The full background on the framework lives on the what is an XLA page. This page goes one level deeper, into the metrics themselves.
| Metric | What it measures | Weight |
|---|---|---|
| CSAT | Satisfaction with a specific interaction | 25% |
| FCR | Issues resolved in a single contact | 20% |
| Sentiment | Emotional tone across the conversation | 20% |
| NPS | Loyalty and willingness to recommend | 15% |
| Resolution Quality | Whether the fix actually held | 10% |
| CES | Effort the customer had to spend | 10% |
The six metrics, one by one
Customer Satisfaction CSAT · 25%
CSAT measures how satisfied a customer was with a specific interaction, usually captured as a rating immediately after the contact.
CSAT is the most direct read of how a customer felt, which is why it carries the most weight. Its limitation is coverage. Only a fraction of customers answer a survey, and the ones who do tend to sit at the extremes. That is why CSAT is paired with sentiment in the composite: one captures the customers who responded, the other captures the ones who did not.
First Contact Resolution FCR · 20%
FCR measures the share of issues resolved in a single interaction, without the customer having to come back.
FCR is a quiet predictor of both cost and loyalty. A customer who has to make contact a second or third time for the same issue is a customer whose effort is climbing and whose patience is falling. Simetrix measures FCR through recontact tracking inside a defined window, so a problem that returns within days is counted as what it is: a resolution that did not hold.
Sentiment 20%
Sentiment measures the emotional tone of an interaction, read across the full conversation rather than asked in a survey.
Sentiment is weighted heavily because it sees what surveys miss. It is read from the interaction itself, across every contact, which means it captures the experience of the silent majority who never fill out a survey. A drop in sentiment is often the earliest signal that a customer is on their way out, weeks before it shows up in CSAT or churn.
Net Promoter Score NPS · 15%
NPS measures how likely a customer is to recommend the brand, tied to the relationship rather than a single contact.
Where CSAT is about the moment, NPS is about the relationship. It moves more slowly and reflects how the customer feels about the brand as a whole. It is weighted below the interaction-level metrics because it lags, but it matters as a check that good individual interactions are adding up to a relationship worth keeping.
Resolution Quality 10%
Resolution Quality measures whether an issue was actually solved, not just closed.
A closed ticket is not the same as a solved problem. Resolution quality guards against the most common form of metric gaming, closing contacts to hit a target while the underlying issue lives on. It is assessed by whether the resolution held, which ties it closely to FCR and recontact. A fix that the customer comes back about was not a fix.
Customer Effort Score CES · 10%
CES measures how much work the customer had to do to get their issue resolved.
Effort is one of the strongest predictors of disloyalty. Customers rarely leave because a single interaction was imperfect. They leave because the whole thing was hard: repeated contacts, transfers, re-explaining the same problem, switching channels. CES captures that friction. A low-effort experience is one customers quietly stay for.
Why a composite beats any single metric
Every one of these metrics has a blind spot. CSAT only hears from responders. NPS lags. FCR can be gamed by reclassifying contacts. Sentiment needs full coverage to be trustworthy. Taken alone, any single number can be made to look good while the experience underneath it gets worse.
The composite removes that hiding place. A high CSAT cannot mask a rising effort score. A strong close rate cannot mask deteriorating sentiment. Because the six are read together, leadership sees the whole shape of the experience rather than the one number that happened to look best this month.
How Simetrix scores the composite
Two things make the composite trustworthy at Simetrix. The first is coverage. Every metric is scored across 100% of interactions, not a three to five percent sample, so the score reflects the whole experience rather than an estimate of it. We make the full case for this on the 100% interaction analysis page.
The second is timing. The composite is scored in real time, not at the end of the month. That is what turns the metrics from a report into a steering signal: a team lead can act on a contact while sentiment is still turning, and leadership can see a drift in effort the week it starts. The weighting itself is configured to each operation, so the composite reflects what actually drives retention in that business.
XLA metrics, answered.
See your operation scored on all six metrics, across every interaction.
Start with a 30-minute CX Review. No call data required to begin.
Tell us about your industry.
30 minutes with our operations team. Industry-specific scoping, written gap summary in 48 hours, honest next-step recommendation. No call data required to start. We respond within one business day with a written gap summary and honest next-step recommendation.