The real cost of a slow reply

Every business with an inbox has a number it has never calculated: what the queue costs while it waits.

It is not a mysterious number. You can work it out in about five minutes from figures you already have, and the result is usually large enough to change a hiring decision.

The arithmetic

You need four things, all of which you already know or can look up:

• Conversations per month. Count every inbound message thread, across every channel.
• The share that are buying conversations. Not support — people asking about price, availability, or whether you do the thing they need.
• Your close rate on those, when you answer quickly.
• Average order or deal value.

Now the part that is usually missing: how many of those buying conversations get a first reply after the person has stopped waiting. Not “after an hour” in the abstract — after the point where they have gone and asked someone else.

That threshold is different per business. For a restaurant it might be minutes. For a B2B service it might be a day. Ask your best salesperson; they will know it precisely.

Multiply: late buying conversations × close rate × average value. That is the monthly number.

Most teams find it is larger than the salary of the person they were debating hiring. That is the actual finding, and it does not require anyone else’s statistics.

Three places response time leaks

The average is not where the damage is. The damage is in the tail, and the tail has three usual causes.

Out of hours

The honest version of most response-time reports excludes nights and weekends, which is exactly when a large share of consumer messages arrive. A nine-to-five team with a twenty-four-hour inbox has a response time of several hours for anything that lands at 8pm, no matter how fast the team is during the day.

This is the leak that automation actually closes, because it is not a staffing problem you can solve by working faster.

The channel nobody has open

If messages arrive in four apps, the slowest channel sets your reputation. Customers do not average their experience across your channels; they remember the one where they waited.

The second reply

First response time is the metric everyone tracks, which means it is the metric everyone games — an instant “thanks, we’ll get back to you” resets the clock without helping anyone. The number that correlates with outcomes is time to a useful answer. Track that one instead, even though it is less flattering.

What to do about it, in order of cost

1. Measure the tail, not the mean. Look at your slowest tenth of conversations. That decile is where the lost revenue is.
2. Close the out-of-hours gap first. It is the biggest single block of late replies in almost every consumer business, and it is the one a human rota cannot fix economically.
3. Put every channel in one queue. A message nobody can see cannot be answered quickly, and the fix costs nothing but configuration.
4. Automate the repetitive half. Opening hours, availability, price, order status — answering these instantly is not a customer experience compromise, it is the opposite. It also frees the people for the conversations where a person changes the outcome.
5. Only then consider hiring. By this point you will know what the queue actually needs, rather than guessing.

The trap in all of this

Fast is not the goal. Answered is the goal. A bot that replies in two seconds with something unhelpful has improved your response-time chart and damaged the thing the chart was a proxy for.

If you automate, insist that the automation answers from your real information — your prices, your availability, your policies — and hands over to a person the moment the question goes outside what it was given. Speed without accuracy just moves the loss somewhere you are not measuring.

---

Fellix answers the repetitive half of every conversation instantly, on every channel, at any hour. See what that is worth on your numbers.