Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Metric

Time to First Reply

The time from a customer's first message to the first reply from a person; read beside repeat contacts, as a quick reply that solves nothing still scores well.
AnalyticsmetricUpdated

Formula

Time of the first reply from a person − Time of the customer's first message

Unit: time · lower is better

Definition

Time to first reply is the elapsed time between the moment a customer first gets in touch about a problem and the moment a person on the business's side first answers. The guide to customer service metrics defines it that way in its table of core measures and treats it as one measure among several, not the verdict on a service.

Time of the first reply from a person − Time of the customer's first message(time)

  • Time of the customer's first message is when the earliest message about the problem arrived. If the customer sent a second message or tried another channel, and the contacts were merged, the earlier one sets the start.
  • Time of the first reply from a person is when someone on the team first wrote back. A message sent automatically to confirm receipt is not a reply from a person and does not stop the clock.
  • The result is a duration per contact, summarised across contacts for a channel and a period.

A team reports the summary, but the summary is a choice too. An average lets one long wait vanish among many short ones, so a median, the longest waits and the proportion of contacts that met the business's own standard say more than a mean alone.

How to interpret it: replies, clocks, channels and local hours

Four rules make the figure mean the same thing from one week to the next.

What counts as a reply from a person. The definition excludes automatic messages; write that into the report itself. A person's reply can still say nothing useful; a holding line that promises a proper answer later stops the clock while the customer still waits for help. That gap is why the figure should not be read on its own.

The clock. Elapsed time and business hours give different numbers for the same contact: a message sent just after closing waits overnight on one clock and barely at all on the other. Customers live through elapsed time. Choose the clock per channel, name it under the chart, and keep it. A contact moved into a waiting status before anyone has replied is still waiting from the customer's side, so time in such statuses is reported apart rather than quietly removed.

By channel. Replying in chat and replying by email are different promises, and each channel is read against its own standard. Blended, a shift in where customers write in moves the average with no change in service.

In local hours. Across markets, the same queue may be answered while some customers are asleep. The guide to supporting customers across languages and time zones reports time to first reply per market in local hours, so a gap that the head office's clock hides becomes visible.

Teams outside support face the same choices. The tutorial that sets service levels for marketing requests writes down when its clock starts, pauses and stops, for the same reason.

What moves it

  • Staffing by hour and channel. A reply can come no earlier than the first person on the channel. Cover that starts after customers' working day has begun builds a queue before anyone logs on.
  • Queue order. Picking easy contacts first lowers the average while difficult ones sit.
  • Clock settings. Changing from elapsed time to business hours, letting automatic messages count, or pausing for waiting statuses moves the figure without changing a single customer's wait.
  • Who is judged on it. When the people answering are targeted on the number, the cheapest ways to move it get used. Goodhart's law describes the pattern: a measure under that pressure stops tracking what it was chosen for.

What good looks like

There is no reliable benchmark for time to first reply that the site cites, and none is given here. The metrics guide says the same of response and resolution times in general, adding that another business's figure tells you little about your own. Set a standard per channel from what customers in that channel expect and what the team can keep in a normal week, then judge the figure in two ways. Against that standard, as the proportion of contacts that met it. And against the business's own history, on unchanged clock rules, before and after a change to hours, routing or staffing. A falling figure is good news only if the measures read beside it hold steady too.

Where an outside team answers customers, the guide to working with an outsourced support team leaves the definitions, and the reports built on them, with the business and its own records; the partner's speed figures help it run its team but do not decide whether the work is good.

Illustrative example — not a real result

A phone repair chain splits its single average

Jestuar Phone Repairs runs 5 shops and also takes mail-in repairs from a neighbouring country 1 hour ahead. Its report showed one figure: an average time to first reply of 1 hour 50 minutes, counted in the team's business hours of 10:00 to 18:00. Split by market and measured as elapsed time in each customer's local hours, the 930 home messages in the quarter averaged 1 hour 40 minutes and the 310 mail-in messages 3 hours 25 minutes. Most of that gap sat in the morning: the neighbouring market's working day starts at 8:00 local time, 3 hours before the team logged on, and the business-hours clock had not been counting those hours at all. The chain moved one agent to a 7:00 start on the mail-in inbox and now reports the two markets separately, each in its own local hours.

Repeat Contact is the partner the metrics guide reads beside it, by reason, because a quick reply that resolves nothing still scores well here. That pairing is one case of a Guardrail Metric: a measure that has to hold while the headline figure is pushed down. Read by Contact Reason, the figure separates a reason that is slow because it needs another department from a team that is slow across the board. When an outside partner answers customers, a SLA between the business and that partner writes the reply standards down, set from the business's own records. Contact Rate measures how often customers have to get in touch in the first place. Time to resolution tracks how long the problem lasts rather than the wait to be heard.

Common mistakes

  • Counting the automatic acknowledgement as the reply. It confirms receipt; the customer is still waiting for a person.
  • One average for every channel and market. The mix moves the number, and customers in other time zones disappear inside it.
  • Changing the clock without saying so. A switch from elapsed time to business hours improves the trend overnight.
  • Hiding waiting statuses. Time a contact spends parked is time the customer waits.
  • Per-agent league tables. Speed targets for individuals reward holding replies and easy contacts first.
  • Borrowing another business's figure. Standards come from the business's own customers and capacity.
  • Confusing it with lead response time. How fast sales answers a new enquiry is a separate measure, as is the first statement in a crisis.

More terms to look up

Where this term is used

Related