A dashboard waits to be opened

Every team that loses track of work eventually asks for a dashboard. A view of everything open, grouped by owner, with the overdue items in red. It is a reasonable request, and it gets built.

Then the work that is going stale stays stale. The dashboard is accurate. It is also sitting in a tab nobody has opened since Tuesday.

We decided Servanda would not rely on anyone opening it.

A dashboard is a pull, and stale work is a push problem

A dashboard answers a question, but only for someone who already thought to ask it. That is the weakness. The work most likely to go wrong is the work nobody is thinking about, and the person least likely to open a view is the one who has not noticed there is anything to look at.

There is a second problem, quieter than the first. A dashboard turns "this is late" into a fact that is available to everyone and the responsibility of no one. It is visible, so it feels handled. Anyone could look. Therefore, in practice, nobody does until something has already gone wrong.

None of that is an argument against seeing your work in one place. Servanda has a task list, and you should be able to read it. The argument is about who has to make the first move. For stale work, we do not think it should be the person who is already busy.

What Servanda does instead

Servanda nudges the owner of a task. If the work then stays stale, it escalates to the owner's manager.

The manager does not have to remember to check. The work comes to them, with the context needed to act on it, at the point where it has become a management question rather than a reminder. We want a stale task to find the person who can unblock it, reassign it or decide it no longer matters, without waiting for someone to go looking.

The boundary is the one we already use elsewhere: a manager sees their direct reports and nothing wider. Escalation follows the reporting line your organisation already has. We did not invent a second one.

What this costs us

An escalation is an interruption, and we are choosing to make it on someone else's behalf. We should say plainly what that risks.

It can feel like surveillance. Nobody likes learning that a task they had half in hand has been flagged to their manager. Done carelessly, escalation teaches people to resent the tool and to keep their real work somewhere it cannot see it, which is the opposite of the goal. A dashboard never does this. It is passive, and being passive is exactly why it is easy to live with.

It can also be wrong. A task may look stale because the real work happened in a conversation nobody recorded, or because it was quietly overtaken by events. An escalation about work that was in fact fine costs the manager's attention and the owner's goodwill, and we will not pretend we can always tell the difference.

So we keep the order deliberate. The owner is nudged first, and has the chance to act before anyone else hears about it. The manager is the second step, not the first. That ordering reduces the cost. It does not remove it, and we would rather say so than promise that nobody will ever be escalated unfairly.

We think the trade is still right. A manager who is interrupted occasionally about something that turned out to be fine has lost a minute. A manager who learns about a stalled commitment after the person waiting on it has already given up has lost something much harder to recover.

This is not really a decision about notifications

The same question sits under a lot of what we build: where does the burden of noticing fall?

Most tools leave it on the person with the least spare attention. They store the information faithfully and wait to be consulted. That is easy to build and easy to defend, because nothing can be blamed on a system that only ever did what it was asked.

But the cost of coordination work is mostly the cost of noticing: who has seen that this slipped, and whose job it was to say so. If a tool can take on part of that, it should, as long as it does so through the people and the structure the organisation already has, and as long as it is honest about when it might be mistaken.

A view you have to open and a prompt that reaches you are different products. The first assumes the problem is a shortage of information. We think it is usually a shortage of someone's attention at the right moment, and that is a harder thing to build for.

Back to the red column

The dashboard in the first paragraph was not wrong. It was just waiting.

We would rather build the thing that goes and finds the person who can act. Servanda is where it lives.