The signs that your team is at its limit, before anyone says so

Nobody gives notice that they are at their limit. They give notice once they have already decided to leave, or once something has broken in front of a customer. Until then, people absorb and compensate, and the operation looks like it works.

But it leaves signs, and they are measurable before the awkward conversation arrives. These six are visible without asking anyone, just by looking at data you already have.

Some work gets done at night and on Sundays

Look at the timestamps. When emails go out, when tickets get closed, when product pages get updated, when the month-end close gets filed. Not the average: the extremes.

If there is regular activity at eleven at night, at seven in the morning or on a Sunday afternoon, that is not commitment. It means a normal working day no longer covers the normal work, and somebody is papering over it with their own free time.

And it always gets papered over the same way: first one evening, then the Sundays at month-end, then every Sunday. It is the most reliable sign of all, since it admits no other reading. Nobody chooses to answer tickets on a Sunday when Monday would do.

Some things have spent months «on next week's list»

Take your backlog and find the date each item was written down, not the date it is due. If something has been sitting there for four months, it is not pending: it has been dropped without anyone deciding to drop it.

What matters is not the list, it is what kind of items they are. Almost always, the things that have waited months turn out to be the tasks with nobody chasing them: updating the catalogue, answering reviews, cleaning up the CRM, documenting the process, reviewing old returns. Nobody complains about them, so they always lose against whatever is urgent.

That set of tasks has an exact name: it is one whole job badly shared out. It does not look like one, since it is sliced across five people and each slice looks small.

One person is the only bottleneck in three processes

List the things that cannot move forward without one specific person. If the same person shows up in three different processes, you have a structural problem, not a workload problem.

The giveaway is not that they work a lot: it is that their inbox is the queue for three teams. When they step into a meeting, three things wait. When they reply late, three things are late. And everyone knows it, including them.

There is also an effect that shows up on no dashboard: that person stops doing the work they were hired for, since their day goes on unblocking everyone else. It is the most common case of expensive talent doing cheap work.

Replies arrive later every month

Do not look at this month's average response time. Look at the last six, in a row.

An average that is high and stable is a decision: replies take two days and everyone knows it. An average that creeps up month by month is an operation degrading without anyone announcing it. That curve is the one to watch, and the one nobody presents in a meeting, since each month on its own looks normal.

The same applies outside customer support: days from a reference arriving to it being published, days to close a return, days of delay on the month-end close. Anything measured in days works, as long as you look at it as a series.

The founder is answering tickets

When whoever sets the strategy spends two hours a day on execution work, the problem is not those two hours: it is that the work nobody else does has become their work, and it is the part that actually gets done.

There is an even more expensive variant: the work gets done well precisely because it is being done by someone who knows the whole business, and that hides the problem. The tickets are answered perfectly. The catalogue is immaculate the week they touch it. And there is no red metric to look at.

The question that exposes it is simple: which of your own work has not moved this month? That is where the cost is, and it is an opportunity cost that appears in no set of accounts.

The concrete test: the last two-week holiday

The signs above are arguable. This one is not.

Think about the last time someone on your team was away for two weeks straight, and answer four questions:

1. What stopped getting done, and is still not done?

2. Who took over their work, and what did that person drop in exchange?

3. How many days did it take to get back to a normal rhythm afterwards?

4. Was there anything only they knew how to do, that had to be solved in a hurry?

If the answer to the first is «nothing, we covered it», look carefully at the second: the work did not disappear, it moved. And if the fourth has examples, you already know where the risk sits.

A two-week holiday is a free rehearsal for a long absence, and almost nobody uses it as one. What happened in those two weeks is exactly what will happen when the absence has no return date, only worse: with a holiday you know when they come back. How an operation degrades when that happens, week by week, is set out on absence and replacements.

What waiting costs

Waiting has a price, and it appears on no line of the accounts.

The first part is churn. Someone who has spent eight months compensating with their own free time eventually leaves, and when they do it costs you the search, the training and the months until the next person is autonomous. They also take with them what they knew and never wrote down, which is precisely what was holding the arrangement together.

The second is quieter: work that nobody chases stops getting done, and nobody notices until everyone notices. Unanswered reviews, a stale catalogue, dirty data in the CRM. None of those causes a crisis on a Tuesday. All of them subtract sales every day.

And the third is the decision you make in a hurry. Hiring with a full queue is hiring badly: you search fast, accept the first person who half fits, and train them with a saturated team. If you also want to see what filling a seat really costs —the part that never appears on a payslip— put your own figures into the real cost of an employee.

What to do with the signs

Three steps, in this order and with no skipping.

Write it down. One week noting who does what and how long it takes. Nothing more sophisticated is needed. At the end of the week you will see two things: how many hours go into repetitive work, and which of those hours are being put in by your most expensive people.

Group before hiring. Gather the loose tasks nobody chases and see whether they add up to a full week. They almost always do. That is what gets delegated first, and it is what hurts least to delegate: nobody loses judgement, only load.

Then decide what fills that week. If most of it is a queue of messages that never goes down, that is customer support and it is the most common place to start. If it is catalogue, data or admin, the job is a different one, and the way to decide is the same: what is piling up, how many hours it is, and who is doing it right now at eleven at night.

Half an hour on a call and you will know whether we fit. If we do not, we will tell you in that same conversation.