Call center KPIs explained: service level, handle time, first-contact resolution and CSAT

The short answer

Most outsourcing contracts measure six things:

• Service level: how quickly calls are answered
• Abandonment rate: how many callers hang up before reaching an agent
• Average handle time (AHT): how long each contact takes
• First-contact resolution (FCR): how often the issue is solved the first time
• Customer satisfaction (CSAT): how customers rate the contact
• Quality score: how well agents follow your standards, from monitored calls

Pick the few that match what the program is for, define exactly how each is measured, and review them together. A single number in isolation is easy to game.

Service level

Service level is the share of calls answered within a target time. It is written as two numbers: a percentage and a number of seconds, for example a stated share of calls answered within a stated number of seconds.

How to calculate it: calls answered within the target time, divided by calls offered, multiplied by 100.

What it tells you: whether enough agents are on shift when customers call. It is the main input for staffing. Agree in writing whether calls abandoned within the first few seconds are excluded, and over what period the figure is averaged (per half hour, per day or per month). A good monthly average can hide busy hours when customers waited a long time.

Email, chat and social messages use the same idea with a response-time target, such as the share of emails answered within a set number of hours.

Abandonment rate

How to calculate it: calls where the caller hung up before reaching an agent, divided by calls offered, multiplied by 100.

What it tells you: how many customers gave up. It usually moves with service level, so the two are best read together. A rising abandonment rate is often the first sign that staffing no longer matches volume.

Average handle time (AHT)

How to calculate it: (total talk time + total hold time + total after-call work) divided by the number of calls handled.

What it tells you: how much agent time each contact needs, which drives how many agents you need. Sudden changes point to something worth investigating: a new product issue, a confusing policy, a system that is slow, or a training gap.

Handle time is a planning measure more than a quality measure. If it is pushed down too hard, agents rush, problems go unsolved, and customers call back. That is why it should sit alongside first-contact resolution.

First-contact resolution (FCR)

How to calculate it: contacts resolved on the first attempt, divided by total contacts, multiplied by 100.

There is no single standard method, so agree one. Common approaches are:

• treating a contact as resolved if the same customer does not contact you again about the same issue within a set number of days
• asking the customer in a post-contact survey whether the issue was resolved
• having the agent mark the outcome, checked by quality reviews

What it tells you: whether customers are actually getting their problem solved. It also shows whether agents have the knowledge and the authority they need. Low FCR is often fixed by changing what agents are allowed to decide, not by more training.

Customer satisfaction (CSAT)

How to calculate it: the number of positive survey responses (for example, 4 or 5 on a 5-point scale) divided by the total number of responses, multiplied by 100.

What it tells you: how customers felt about the contact. Watch the response rate and who answers: unhappy and very happy customers respond more often, so small samples swing. Separate scores about the agent from scores about your product or policy, so your partner is judged on what it controls.

Some companies also track Net Promoter Score (NPS), which asks how likely a customer is to recommend you. NPS reflects the whole relationship with your brand, so it is better as a company measure than as a contract target for a call center.

Quality score

A quality score comes from reviewing a sample of calls, emails or chats against a scorecard. Typical items are greeting and identity checks, accuracy of information, following the process, tone, and correct notes in your systems.

What it tells you: whether agents handle contacts the way you want, including things customers never see, like data entry and compliance steps. Agree the scorecard together, and hold regular calibration sessions where both teams score the same calls, so a score means the same thing to everyone.

Sales programs: conversion and revenue

For sales and lead qualification, add outcome measures: conversion rate (sales divided by contacts or qualified leads), average order value and revenue per agent hour. Pair them with a quality score, so results are not won with promises your business cannot keep. See outsourced sales and our guide to commission-based outsourced sales.

How to set targets in a contract

1. Start from your own baseline. Use your current figures where you have them. Industry averages vary widely by sector, channel and call type, so your own history is a better guide.
2. Choose a few measures that pull in different directions. Service level with handle time, and handle time with FCR and CSAT, keep each other honest.
3. Define every measure in writing. State the formula, the data source, what is excluded, and the reporting period.
4. Allow a ramp-up period. New agents get faster and more accurate over the first weeks. Agree interim targets for launch, then the steady-state targets.
5. Agree how volume forecasts work. Service level depends on staffing, and staffing depends on your forecast. Decide what happens when actual volume is well above the forecast.
6. Review targets regularly. As your products, policies and customers change, the right targets change too.

The KPIs you choose also depend on how you pay. Per-seat contracts usually focus on service level and quality, while per-minute contracts focus on answer speed and quality per call. See per-seat vs per-minute pricing.

How KCG BPO reports on KPIs

At KCG BPO, service levels are agreed with each client before launch. Clients receive weekly performance reports and a monthly review meeting against those agreed service levels, and quality is monitored through call monitoring and quality assurance. Agents work in your own CRM and ticketing systems, including Salesforce, HubSpot and Zendesk, so the data comes from the tools you already use. Every program runs 24/7, and most go live in 2–4 weeks.

If you are comparing providers, see how to choose a call center outsourcing partner, and how to brief an outsourced call center for what to prepare.

Last reviewed: October 2026