Measuring Contact-Centre Quality beyond Handling Times

Why handling times are not enough#

Contact centres have long been managed on productivity: speed of answer, abandoned call rate, average handling time (AHT). These indicators measure the centre's capacity to absorb volume. They do not say whether the customer got what they were looking for.

Supervisors know the trap: an AHT target that is too tight pushes the agent to shorten the call, transfer the difficult case or promise a callback. The customer calls back, volume rises, average AHT goes up again and the centre concludes that it is understaffed. To break out of this cycle, you need indicators that measure outcomes. It is also the precondition for selling higher-value services, as our guide on repositioning a contact centre explains.

First contact resolution#

First contact resolution (FCR) is the share of requests settled without the customer having to come back. It is measured in two ways, which should be combined:

• declared: the customer is asked, after the conversation, whether their request has been settled;
• behavioural: you check that they do not contact you again for the same reason within a window set in advance (seven days, for example), across all channels.

The gap between the two is revealing. A resolution declared by the agent or the customer, then contradicted by a repeat contact, signals a problem of diagnosis, procedure or information.

Repeat contacts, transfers and complaints#

• Repeat contact rate: the share of customers who come back for the same reason. It is the flip side of FCR and often a centre's first hidden cost.
• Transfer rate: the share of conversations passed to another agent or department. A high rate reveals unsuitable routing, decision rights that are too limited or missing skills.
• Complaint rate: complaints relative to the number of contacts. It measures the failures the customer considers serious enough to formalise. Our guide on complaint management explains how to trace these cases back to their causes.

These three indicators are behavioural signals: they reflect what customers do, not just what they say. The guide on behavioural signals places them within a broader listening system.

Customer effort and surveys#

CES measures the perceived effort needed to settle a request; CSAT, satisfaction with the conversation. How they work is explained in our guide on NPS, CSAT and CES. Two precautions for a contact centre:

• Satisfaction scores are almost always concentrated at the top of the scale (Peterson and Wilson, 1992), and surveys administered by telephone tend to produce more positive answers than written questionnaires. The guide on courtesy bias details these effects.
• In the study by De Haan, Verhoef and Wiesel (2015), conducted with a Dutch panel covering 93 companies, no declared-feedback indicator strongly predicted retention; CES had little or no predictive power. A survey complements behavioural measures; it does not replace them.

Quality assessment of conversations#

Quality assessment consists of listening to or rereading a sample of conversations and scoring them against a scorecard. A useful scorecard:

• distinguishes critical criteria (wrong information, personal data disclosed, unauthorised commitment) from skill criteria (diagnosis, clarity, rephrasing);
• assesses the outcome for the customer (has the request been settled?) as much as adherence to the script;
• is the subject of calibration sessions, in which several assessors score the same conversation to align their judgements;
• feeds individual coaching, not just a ranking.

Recording and listening to calls are processing of personal data: in Morocco, Law 09-08 requires individuals to be informed and the corresponding CNDP formality to be completed. See our CNDP compliance method.

Speech analytics and AI: what the AI Act allows#

Automated analysis tools transcribe conversations, classify reasons, spot keywords and can suggest a quality score. For a centre serving European clients, the AI Act draws precise limits.

The detail is in our guide on emotion analysis in call centres. For complaints, see also the safeguards specific to AI applied to complaints.

A balanced dashboard#

Read these indicators together: a fall in AHT accompanied by a rise in repeat contacts is not progress.

Putting it in place in six steps#

1. Code contact reasons consistently across all channels.
2. Choose the repeat contact window and measure baseline behavioural FCR.
3. Revise the quality scorecard and hold a first calibration session.
4. Remove AHT from individual targets and reserve it for planning.
5. Check every automated analysis tool against the AI Act and Law 09-08.
6. Present the dashboard every month, with the three main causes of repeat contact and the actions under way.