Response time tracking means recording how long a customer waits between getting in touch and receiving the first meaningful reply from a person. To do it well, you need one written rule for when the clock starts and stops, one figure per channel, and a target you report against every week. For a service or engineering business, it also helps to track the next stage separately: how long until an engineer is booked and on site.
This guide shows you how to set those rules, which metrics are worth the effort, what targets are realistic for a UK business, and how to turn the numbers into fewer missed calls and fewer SLA breaches.
TL;DR
- Response time is the wait for the first human reply. Resolution time is the wait until the job is closed. Track them separately, or the real bottleneck stays hidden.
- Write down one rule for the clock: start at the customer's first message, stop at the first reply written for that customer, and ignore automatic acknowledgements.
- Report one figure per channel, and show the share answered inside target rather than a single average.
- Pick targets by channel and urgency, then review the trend weekly and look at where requests wait.
- The less re-keying between inbox, job and customer record, the easier all of this becomes.
Table of Contents
- What response time means, and what it does not
- Set the rules once: where the clock starts and stops
- Which metrics are worth tracking
- Setting targets that fit your business
- Tracking response time across inboxes and channels
- Using the numbers to find delays
- Our perspective
- See response, scheduling and SLAs in one place
- FAQ
- Sources
What response time means, and what it does not
Response time is the time from when a customer submits a request until your team acknowledges it and starts working on it. Resolution time is how long it takes to actually fix the problem and close the ticket. CustomerExperience.io describes the first as a communication commitment and the second as an outcome commitment, and says both should be tracked separately so the real bottleneck is not hidden.
The distinction matters in field service. A customer reporting a failed extraction fan or a tripping circuit wants two things: to know somebody has taken ownership, and to see the problem fixed. A team can be excellent at the first and poor at the second, or the reverse. If you only track the average of both, you cannot tell which one is slipping.
It also helps to be clear about what counts as a response. An automatic "we have received your message" email tells the customer the message arrived. Break.Beat recommends measuring the meaningful response instead, because an auto-reply should not be treated as the true response if the customer then waits for help. The same source suggests recording several timestamps: receipt, acknowledgement and first meaningful response, plus the next step agreed where relevant.

Three clocks for an engineering business
For a business that sends people to site, we find three clocks give a fuller picture than one:
- Response: from the enquiry arriving to a person replying.
- Booking: from the enquiry arriving to a visit being scheduled and confirmed with the customer.
- Resolution: from the enquiry arriving to the job being completed and closed, with certificates or reports issued where needed.
The first is a measure of your office. The second is a measure of your scheduling. The third is a measure of the whole operation. When complaints rise, comparing the three shows you where the delay sits.
Set the rules once: where the clock starts and stops
Inconsistent definitions are the most common reason response time figures cannot be trusted. Queue Bench advises starting the clock when the customer's first message arrives (ticket creation for email and forms, the first message for chat and social) and stopping it at the first reply written for that customer, excluding automatic acknowledgements. It also says to decide whether you use business hours or calendar hours, and to apply that choice consistently.
Turn that into a one-page rule for your team. It should answer these questions:
- Start: is it when the email lands in the shared inbox, when someone logs it as a job, or when the phone rings? Choose arrival, not logging, or your figures will flatter you.
- Stop: is it the first human reply, a call back, or a confirmed visit time? Name it.
- Hours: does the clock run overnight and at weekends, or only during opening hours? A Friday 5pm email answered at 9am on Monday is a very different story under each rule.
- Exclusions: spam, supplier emails and internal chatter should not be in the data.
- Ownership: who is responsible for a reply when a message sits in a shared inbox?
That last point is where most small teams fall down. A shared mailbox belongs to everyone, so nobody feels the clock. Giving each incoming message a named owner, even if it is just whoever is on the desk that day, is the cheapest improvement available.
Which metrics are worth tracking
You do not need a long list. A handful of figures, reported consistently, will tell you more than twenty dashboards.
| Metric | What it tells you | How to calculate it |
|---|---|---|
| First response time | How quickly customers hear from a person | Time from customer message to first human reply |
| Average response time | The typical wait over a period | Total response time divided by the number of responses in the same period |
| Median and 90th percentile | What most customers experience, and how bad the slow tail is | Sort response times, take the middle value and the value that 90% fall under |
| Share answered within target | Whether you meet your service level | Replies inside target divided by all replies |
| Waiting now | Messages with no reply yet | Live count of unanswered items |
| Response time by day or agent | Where peaks and gaps occur | Same figures, split by weekday, hour or owner |
The average calculation is simple. SuperOffice explains that you add up the time taken to respond to emails in a chosen period and divide by the number of responses sent in that same period.
The average alone can mislead. Queue Bench advises reporting the median plus the 90th percentile, or the share of tickets answered inside the target, because a few tickets can skew a mean. One enquiry that sat untouched for a week will drag a month's average up far more than it reflects normal service. It also recommends watching trends over weeks, since a slow week after a launch or an outage says more than a monthly average.
A useful supervisor-style report combines several of these. Cirrus describes an email analysis report covering messages received, waiting and answered, the average time agents take to reply, and the percentage answered within a defined service-level time. That combination, volume, backlog, speed and share on target, is a good model for any service desk.
Do not blend your channels
Phone, email, web forms and chat behave differently, and a single combined average hides the differences. Queue Bench recommends one figure per channel. GOV.UK's Service Manual makes a related point: response times depend on the kind of enquiries you get and when, including variation by day, and on which channels staff are assigned to and whether they can work across several at once.
If your office manager is answering the phone, dispatching engineers and replying to emails, your email response time will reflect that. It is a staffing signal, not necessarily a performance problem.
Setting targets that fit your business
A target is only useful if it is specific, tied to a channel and measured as a percentage. GOV.UK suggests defining a service level as the percentage of enquiries you aim to handle for each channel in a given period, for example answering 80% of email enquiries within 24 hours. It also advises using historical data and average handling time by channel to estimate how many enquiries you will receive and how quickly you can deal with them.
For UK small businesses, CPD Online suggests these workable response targets:
- Phone calls answered within 60 to 120 seconds.
- Live chat first response within 60 seconds during opening hours.
- Email and web form acknowledgement within 4 working hours for priority issues, or 1 working day for standard enquiries.
- Social media first response within 1 to 2 hours during staffed hours.
For resolution, the same source suggests 1 to 3 working days for standard issues with clear fixes, 5 to 10 working days where another team or supplier is involved, and 10 to 20 working days for complex investigations, with regular holding updates.
These are starting points for general customer service, not rules. In engineering and facilities work, your real targets usually come from contracts. If a maintenance agreement promises a four-hour response to a priority fault, that is your target for that customer, and your tracking needs to be able to show it. Break.Beat's advice is to vary standards by intent and channel, and to define the hours the standard applies, a target for acknowledgement and one for meaningful response, a fallback route and a weekly measure.

A simple tiering example
Here is a pattern many service teams use. Treat it as an illustration to adapt, not a benchmark:
| Priority | Example | Response target | Booking target |
|---|---|---|---|
| Emergency | No heating at a care home, gas leak report | Same hour | Same day |
| Urgent | Refrigeration fault at a commercial kitchen | Within 4 working hours | Next working day |
| Routine | Planned maintenance query, quote request | Within 1 working day | Within an agreed window |
Write your own version, link each tier to the contracts that define it, and measure the share of requests that met each target. If you want to see how priority rules can feed scheduling, our piece on priority-based scheduling for SLA breaches goes through it.
Tracking response time across inboxes and channels
The practical difficulty is that enquiries arrive in different places: a shared mailbox, a web form, a phone, a message to an engineer's mobile. Each has its own timestamp, and none of them talk to each other.
There are three ways teams cope. The first is a manual log, which works for a handful of enquiries a day and fails the moment the office gets busy. The second is exporting data from each tool into a spreadsheet, which is slow and tends to happen only when someone complains. The third is working from one system in which an enquiry becomes a job or a customer record the moment it arrives, so the timestamps are captured as part of the work rather than as extra admin.
The principle is the same whatever the channel. If you want a deeper look at how response targets are set for inbound leads and the cost of slow replies, this piece on lead response time benchmarks covers the sales side of the same discipline.
Two cautions apply to any set-up. Automatic acknowledgements must be excluded from the first-response figure, otherwise your numbers will look excellent while customers still wait. And channel clocks should be reported separately. KPI Depot's definition page makes the same point: measurement choices about the statistic, the timing definition and clock consistency change what an average response time represents, so definitions need to be standardised before periods are compared.
Using the numbers to find delays
A response time report is only worth having if it leads to a decision. Here is a routine that takes 20 minutes a week.
- Check the share on target. Did you meet your service level this week, by channel and by priority?
- Look at the slow tail. Open the ten slowest responses. Was it a staffing gap, a message nobody owned, or a request that needed information you did not have?
- Split by day and hour. Peaks usually show up quickly. Monday mornings and the hour after lunch are common candidates.
- Compare response with booking and resolution. If replies are fast but visits are slow, the problem is scheduling, not the inbox.
- Change one thing. Add cover at the peak, assign a daily inbox owner, or tighten a handover, then check next week.
The Cirrus help page walks through exactly this loop: reviewing waiting items and response times above target, then adding staff at peak times and providing extra training, with improved response times and a higher service-level percentage shown the following week. GOV.UK recommends collecting data on contact volumes plus response and handling time, per channel, team and individual, to improve your estimates and plan improvements.
There is also a useful diagnostic in Freshservice's performance analysis guidance: if average first response time is poor but average response time is good, tickets may not be getting assigned quickly enough. In a service business, the equivalent is an enquiry that sits in the inbox before anyone allocates it to a job or an engineer.
Visibility without extra admin
The best response time data is a by-product of normal work. JBC Industrial Services, a UK field service firm with 40-plus engineers, reported spending around 8 man-days a week on engineer paperwork and manual data entry before integrating its systems, and said that work was eliminated afterwards. That is a case study from one company using a different product, not a promise about any system, but it shows the principle: when information flows through connected tools instead of being retyped, measurement stops being a separate job.
For the wider reporting picture, our guides to field service reporting metrics and field service dashboards show how response figures sit alongside utilisation, first-time fix and compliance measures. If your contracts carry compliance obligations, five compliance KPIs is a good companion.
Our perspective
Curcle was born inside a real UK service and engineering business, so we have lived the problem. Enquiries arrive by email, phone and text. Somebody writes it on a pad, somebody else chases it, and a month later nobody can say how long the customer actually waited.
We think response time tracking should not be a separate project. It should be a consequence of how the work is recorded. When the enquiry, the customer, the site, the asset, the job and the engineer sit in one connected system, the timestamps already exist. Nobody has to remember to log them.
Our approach is that every click should earn its place, and operational software should help you see what needs attention rather than only recording what has already happened. For response time, that means a view of what is waiting now, what is close to breaching its SLA and who owns it, not a report that tells you last month was bad.
We also believe AI should assist rather than replace human judgement. A system can flag the request that is about to miss its target. A person decides whether to move an engineer, call the customer or escalate.
Compliance-led businesses have a further reason to care. When a contract says four hours, you need to be able to show it, quickly, with an audit trail. That is why we build compliance into the backbone of the platform, with audit-ready exports shaped around UK regulatory standards, instead of bolting it on afterwards.
Luke Herridge
See response, scheduling and SLAs in one place
If your response time data currently lives in an inbox, a spreadsheet and somebody's memory, Curcle brings jobs, customers, engineers, assets, compliance, stock, quoting and invoicing into one connected system, with SLA management and role-based dashboards so office and field teams work from the same information.
Plans start from £99 per month for Starter, from £249 per month for Professional and from £499 per month for Business, with Enterprise available on request. You can compare them on our pricing page, see how the platform handles multi-site SLAs for facilities teams, or explore the full feature list.
FAQ
What is a good response time for a service business?
It depends on channel and urgency. For UK small businesses, CPD Online suggests answering calls within 60 to 120 seconds and acknowledging emails within 4 working hours for priority issues or 1 working day for standard enquiries. If you hold maintenance contracts, the response times written into those agreements take priority over any general benchmark.
What is the difference between response time and resolution time?
Response time is how long a customer waits for the first meaningful reply from a person. Resolution time is how long it takes to fix the problem and close the request. Track them separately, because a team can be quick to reply and slow to fix, and each needs a different remedy.
How do I measure response time in a shared inbox?
Start the clock when the customer's message arrives and stop it at the first reply written for that customer, ignoring automatic acknowledgements. Decide once whether to count business hours or calendar hours, and give every message a named owner. Report the share answered within target, not just the average.
Should I use average response time or median?
Use both, and add the share answered within target. A few very slow replies can inflate an average, while the median shows what most customers experience. Reporting the 90th percentile shows how bad the slow end is.
Sources
- Response Time vs Resolution Time as SLA Metrics Explained, CustomerExperience.io
- How quickly should you respond to a new enquiry?, Break.Beat
- First response time in customer service, Queue Bench
- Set up and manage user support, GOV.UK Service Manual
- 5 Ways to Reduce Customer Service Response Times, SuperOffice
- Email Service Analysis, Cirrus
- Customer service standards, CPD Online
- Service Desk Performance Analysis report, Freshservice
- JBC Industrial Services case study, APH
- Lead Response Time Benchmarks, SDR.ai
- Average Response Time KPI definition, KPI Depot

