Response Time Reporting: What to Track and Why It Matters

Response Time Reporting: Which Metrics Matter and Which Are Vanity Numbers?

If you only do one thing with your response time reporting, track three things: first response time, time to resolution, and how both look at the slow end (your 90th percentile), not just the average. Split them by channel and review them weekly. Then add one easy companion: the share of conversations answered within your target. It turns the slow end into a promise your whole team can rally around. Use these customer support metrics to find bottlenecks and coach your team, not to chase a flattering number.

That's the short version. The rest of this post is the coffee-shop version: why those numbers matter, which ones to treat with caution, and how a small team can do useful response time reporting without a data analyst. 😊

Why response time reporting matters, and where it goes wrong

Picture a two-person support team. On Monday morning they check their dashboard and see a proud '2-minute average first response'. High fives all round. Then a customer emails to say they waited three days for an answer about a billing error, and another says they've given up chasing a refund.

Both things are true. The average really is two minutes, because most of the team's volume is quick live chats. But the number says nothing about the email that sat untouched over the weekend.

That's the trap. A good report answers one question: are customers feeling looked after? A bad one answers a different question: do we look busy? In this post I'll split metrics into two groups. Diagnostic metrics tend to line up with happy customers. Vanity metrics look good on a slide but tell you little. No metric is a perfect crystal ball, so I'll flag where each one can mislead.

This matters even more once live chat and email share one shared inbox. Customers expect very different speeds from each channel. Someone typing into a live chat widget is waiting right now. Someone sending an email is usually fine with a few hours. Blend those together and you get a number that describes nobody's actual experience.

The good news is you don't need a data team. You need a handful of metrics and a regular habit.

First response time vs. resolution time: which metrics matter?

Let's define the two big response time metrics properly.

• First response time is the time from a customer's message to the first human reply.
• Time to resolution is the time until the issue is solved, not just until someone said 'we're looking into it'.

Customers feel both, in different ways. First response is often where trust starts, because it tells them a real person has seen their message. Resolution is usually where satisfaction is won or lost, because it tells them the problem is gone. A fast 'we're on it' only helps if the fix follows. Otherwise you've just sped up the moment they start feeling ignored.

There's plenty of evidence that people care about speed up front. HubSpot's 2022 State of Customer Service report found that 90% of consumers surveyed said an immediate response was important or very important when they had a customer service question. Zendesk's 2023 CX Trends research reported that 72% of customers expect immediate service. (Exact wording and survey populations vary, so check the original reports before quoting them.) Treat these as context. They show that people care about speed. They don't tell you what the right target is for your team, and they don't prove that faster first replies make customers happier on their own.

Exclude automatic replies from first response time

Here's a common way first response numbers get quietly inflated: automatic acknowledgements. If your customer support software counts 'Thanks, we've received your message!' as a first response, your average will look wonderful and mean almost nothing. I'd recommend excluding auto-replies and bot greetings from your reporting. It's also worth checking how your tool treats pauses and reopened conversations, so the numbers match what customers actually experience.

First response time vs. resolution time at a glance

I've left specific targets out of the table on purpose. Set your own based on your baseline (more on that shortly).

Measure live chat in minutes and email in hours

Live chat is best measured in seconds and minutes. Email is naturally measured in hours. Blend them into one figure and chat will drag the number down and hide email delays, which is exactly what happened to our two-person team above.

Should you measure response time in business hours or calendar hours?

This one catches out a lot of UK teams. If you work standard hours, an email arriving at 6pm on a Friday will 'wait' 63 hours before you're back at 9am on Monday. Measured in calendar hours, that looks terrible. Measured in business hours, it looks fine. Neither is wrong, but pick one, apply it consistently, and be clear about it with customers. Many teams report business hours internally and promise them to customers too, such as 'within four working hours'.

Which response time metrics are worth tracking, and which are vanity numbers?

A useful test for any metric is whether you can tie it to a customer feeling. Here's how I'd sort them. The first three are your core set, and the rest play supporting roles.

Response time metrics worth tracking

• First response time, by channel. Did they hear back quickly? Keep chat and email separate, and look at the median and the slow end, not only the average.
• Time to resolution. Was it fixed, and how long did it take?
• The slow end (90th percentile) of both. How bad is it for the unluckiest one in ten? More on this in the next section.
• Share of conversations answered within your target. This is the easy companion to the slow end. 'We answered 92% of emails within four business hours' is a clear goal your team can rally around, whereas 'our average is 3.1 hours' is just a number. It doesn't replace the 90th percentile. It makes the same idea easier to talk about.
• Reopen rate. Was it really fixed? If customers keep coming back on the same issue, speed may be hiding poor resolution.
• Backlog age. Is anyone waiting in the shadows? Your oldest unanswered conversation tells you things an average can miss, including demand building up.
• Replies per conversation. Treat this as a prompt for curiosity, not a verdict. A lot of back-and-forth can signal confusion, but it can also reflect a complex issue or a customer who simply likes to chat. Read a few threads before drawing conclusions.

Vanity metrics to treat with caution

• Total conversations handled. It's context, not performance. A big number might mean you're busy, or that your product has a confusing sign-up flow.
• Raw average first response with no context. No channel, no median, no sense of the slow end.
• 'Messages sent' per agent. High volume isn't the same as high productivity. An agent sending twelve messages to solve something that should have taken two might be stuck, or the customer might be. Again, read the thread before judging.

Zendesk's CX Trends research adds a related point: around 70% of customers surveyed expect anyone they speak to to have the full context of their situation. A quick reply that makes the customer repeat everything isn't a win, so repeat contacts are worth watching.

Check response time metrics against customer satisfaction

Where you can, pair these numbers with satisfaction ratings, even a simple thumbs up or down. That lets you check which of your metrics line up with happy customers in your business. Speed alone can reward rushed replies and premature closures, so a speed metric should always sit next to an outcome metric like satisfaction or reopen rate.

Why average response time can be misleading

Let's work through a quick example, because this is the part that surprises people.

Imagine ten customers. Nine chat with you and get a reply in one minute. The tenth sends an email and waits 30 hours (1,800 minutes). Add it all up and divide by ten, and your average is roughly three hours.

Three hours! That number describes nobody. Nine people were delighted and one waited well over a day. The average smooths everything into something that sounds mildly acceptable, and the unhappy customer vanishes into it.

Median and 90th percentile response time, in plain English

Two simple ideas rescue you here:

• The median is the middle result: half your customers waited less, half waited more. In our example it's one minute, which tells you what a typical customer experiences.
• The 90th percentile is the wait that nine in ten customers beat. It shows you the slow end. If your 90th percentile is eleven hours, roughly one in ten customers waited longer than that.

A word of caution, though. With only ten conversations, the 90th percentile is shaky. Depending on how your tool calculates it, our example could land anywhere from one minute to the full 30 hours, because a single conversation decides the answer. That's why I haven't drawn a 90th percentile line on the chart above. Percentiles become useful once you have a few weeks of data and a reasonable volume of conversations. It's also worth checking how your reporting tool calculates them, so you compare like with like over time.

When you do have enough data, look at all three together: average, median and 90th percentile. If the average improves but the 90th percentile gets worse, you may have a growing group of people waiting too long while quick chats flatter the overall picture.

What can skew your response time reports?

A few usual suspects pull averages around:

• Outliers. One forgotten email can wreck a week's average.
• Weekends and overnight gaps. These stretch calendar-hour numbers unless you use business hours.
• Blended channels. Fast chats hide slow emails.
• Unanswered messages. A SuperOffice benchmark of around 1,000 companies found that 62% didn't respond to a customer service email at all, and the reported average response time (12 hours 10 minutes) only covered those that did. It's an older study, so don't treat it as a current benchmark. The lesson still sticks, though: leave out the non-replies and your numbers look better than reality.

Look at the response time distribution, not just one number

The most practical question you can ask is: what percentage of customers waited longer than we promised? That single figure is hard to hide behind. Most help desk software lets you filter by channel, tag and time period, so you can see beyond one blended number without building a spreadsheet.

How to set realistic response time benchmarks for your team

It's tempting to Google 'good email response time', copy the first number and call it a target. Please don't. Published figures vary widely, depend on who was surveyed and when, and are best treated as rough guides. (If you do quote any, double-check the source and date first.)

Here's a calmer way to do it.

1. Pull two to four weeks of your own data as a baseline. Before borrowing anyone else's targets, find out what you actually do today. You may be better, or worse, than you think.
2. Set separate targets for live chat and email. Customers expect minutes in one and hours in the other. A single target will be too slack for chat or impossible for email.
3. Factor in team size, opening hours and peak times. A team of four can't match a 24/7 enterprise operation and doesn't need to. If Monday mornings are chaos, build that into your expectations.
4. Frame your target as a promise. For example: 'We reply to 90% of emails within four business hours.' It's specific, it's measurable, and it tells everyone what 'good' means.
5. Review quarterly, not monthly. Raising the bar every few weeks is a fast route to burnout. Let a target settle, then adjust it when your team, volume or opening hours change.

One more thing: tell your customers. Put your target on your contact page or in your auto-reply, such as 'We usually reply within four working hours'. It sets expectations, reduces chasing emails, and gives your team a public standard to live up to. Just remember that the auto-reply itself shouldn't count as your first response in the reports!

How to use response time reports to spot bottlenecks

Once you have a baseline, the reports become a detective tool. Here's where I'd look first.

• Time of day and day of week. Patterns appear quickly. A mid-afternoon slump might mean lunch cover is thin, or that everyone's heads-down on a project. Either way, it's a staffing conversation, not a blame conversation.
• Tags. Filter by topic to see which kinds of conversation take longest to resolve. Billing, bugs and delivery queries are common culprits.
• Bouncing and unassigned conversations. In a shared inbox, things can slip between people. Look for conversations that get passed around or sit with no owner, because 'someone else is probably on it' is where customers get lost.
• Handoffs. Internal notes and collaboration data can show where work stalls between people, such as waiting for a developer to confirm a bug.
• Quietly neglected channels. It's easy for email to wait while everyone watches the live chat window. If email's slow end is creeping up while chat looks great, that's your answer.

Slow topics usually need a better system

When the same topic shows up as slow again and again, 'work faster' is rarely the right fix. More often there's a missing help article, a canned reply that would save ten minutes each time, or a simple tag-and-route rule that gets the right person involved sooner.

How to turn response time reporting into coaching conversations

This is the bit I care about most. Reports are a conversation starter, not a scoreboard. If your team feels watched, they'll game the numbers by closing conversations early, sending quick 'looking into it' replies, or avoiding tricky tickets. Nobody wins.

Try opening with curiosity. Something like: 'I noticed billing questions take longer than most. What's getting in the way?' You might hear that the refund process needs three approvals, or that nobody's sure about the VAT rules. Suddenly you've found a process fix instead of a performance problem.

Reports can also point to training needs. One agent with lots of reopens on a specific topic might need a quick walkthrough of it. A knowledge gap across the whole team probably means a missing doc, not a missing skill. Both are useful, and neither calls for blame.

A few habits that help:

• Celebrate team-level wins first. 'We answered 94% of emails inside target this month' lands better than singling anyone out.
• Don't rank agents on speed alone. It encourages rushed, low-quality replies, and customers can tell.
• Pair the numbers with real conversations. Spend ten minutes reading a few actual threads together. The data tells you where to look, and the conversation tells you what happened.

How often should a team review response time reports?

Little and often beats big and rare. Here's a rhythm that works for small teams:

• Daily (two minutes): glance at your backlog and your oldest unanswered conversation.
• Weekly (15 to 30 minutes): review first response, resolution, the slow end and the share answered within target, plus any outliers, as a team.
• Monthly: look at trends, tag analysis and whether your targets still fit.
• Quarterly: reset benchmarks and decide on any changes to process, staffing or tools.

Keep it light. One shared dashboard beats five spreadsheets, because the best report is the one people actually open.

Make response time reporting simpler with customer support software

Here's the unglamorous truth: accurate reporting is nearly impossible when your tools are scattered. If chat lives in one app and email in another, you're stitching numbers together by hand. Different tools also measure things differently, so you end up comparing apples with slightly different apples.

A shared inbox with live chat and email in one place gives you comparable data across channels: the same clock, the same definitions, one view. That's what good customer support software should do for you.

One example is Sonny, which brings live chat, email-to-ticket forwarding, internal notes, tagging and filtering together, with reporting on response times and team performance. It costs a flat $19.99 a month in US dollars (so the sterling amount will depend on the exchange rate) with unlimited agents, and there's a free trial. Check sonny.com for current plans before you decide. Whatever tool you choose, make sure it lets you filter by channel and shows more than a single blended average.

Your next step: pick three metrics (first response time by channel, time to resolution, and the slow end of both), grab a baseline this week, and add the share answered within target as your team-friendly goal. Then review it together on Friday. Don't aim to fix anything yet. Just look. 💛

Frequently asked questions about response time reporting

What response time metrics actually matter?

Start with first response time (excluding auto-replies), time to resolution, and the slow end of both (the 90th percentile), all split by channel. The share of conversations answered within your target is a handy companion that makes the slow end easier to talk about. Add reopen rate and backlog age once those feel comfortable.

How often should a team review performance reports?

Do a quick daily check on your oldest unanswered conversations and a short weekly review of response, resolution and the slow end. Then take a deeper monthly look at patterns by topic and time. Reset your benchmarks every quarter, or whenever your team or volume changes significantly.

Can response time reporting help identify training needs?

Yes. If certain topics take longer, get reopened more often or involve lots of back-and-forth, that can point to a knowledge gap or a missing resource. Treat it as a prompt for a supportive conversation and a chance to improve docs or canned replies, not as a way to blame individuals.

Is average response time a bad metric?

It isn't bad, just incomplete. Averages hide the customers who waited far too long. Pair them with the median, the 90th percentile and the share of replies that met your target to get a truer picture.

Should live chat and email have the same response time target?

No. Customers expect live chat replies in minutes and are generally comfortable with email replies in hours. Set separate targets for each channel and report on them separately.