One dashboard is the wrong ask.

Lede

The request arrives early in almost every conversation with a group, and it is always phrased the same way. We have several properties. We would like one dashboard.

The request everyone makes

It is a reasonable thing to want. A group office is accountable for properties it cannot walk through, staffed by people it did not hire, running procedures it approved some time ago and has not seen executed since. The instinct to ask for one screen is the instinct to close that gap.

The trouble is that the screen, when built, is almost always disappointing, and disappointing in a consistent way. It is looked at intently for a fortnight and then only before board meetings. Nobody can quite say why. The data is correct. The properties are all on it. And yet no decision has ever been made because of it.

The reason is that the request contains two entirely different questions wearing one coat.

Two questions wearing one coat

The first question is operational: is anything wrong right now. It is urgent, it is specific, and it is property-shaped. The answer is always a place and a person: the arrivals at Riyadh Central are backed up because the desk is short a person on a sick shift, or Desert Rose Resort has an escalation that has been open since Thursday and nobody senior has touched it.

The second question is comparative: are we running the same way everywhere. It is not urgent at all. It is slow, structural, and portfolio-shaped. The answer is not a place; it is a pattern. Marina Bay Hotel closes its housekeeping checklists and Palm Suites does not. Three properties escalate a certain complaint to a duty manager and the fourth handles it at the desk.

One dashboard answers neither properly. Built for the first question it becomes an alarm panel that head office cannot act on, because the person who can act on it is standing in the building and already knows. Built for the second it becomes a set of trend lines that are true but too slow to be interesting, and the operational people ignore it because it never tells them anything today.

Most single dashboards end up as a compromise between the two, which is to say they answer a third question that nobody asked.

The average that describes nobody

There is a specific failure that deserves naming, because it survives every redesign.

Roll a figure up across Marina Bay Hotel, Palm Suites, Riyadh Central and Desert Rose Resort and you produce a number that describes none of them. A resort and a city property do not have the same working day. A property in its first season and one that has been open for years do not have the same problems. The portfolio average is a number that is true of the portfolio and untrue of every building in it, which makes it excellent for reporting upward and useless for deciding anything.

The second failure is subtler. Group dashboards tend to fill with the figures that are easiest to move between systems rather than the figures that matter, and in hospitality the easiest figures are commercial ones. So the screen quietly becomes a finance screen with an operations title, and the operational reality — whether the work is being done the same way, whether escalations are closing, whether anyone read the new procedure — is not on it, because it is harder to extract and harder to draw.

It is worth being blunt about what belongs in the second category. Some of the money figures a group looks at are estimates. In Treema they are labelled as estimates, on purpose, because a number presented without its uncertainty gets treated as a fact by the third person who sees it. A group office that builds its comparisons on those figures is comparing properties on a foundation that was never meant to bear weight.

What a group can actually see

Strip away what a head office cannot know and a surprisingly useful list remains. It can see:

• Whether the procedure is the same document. If four properties have four versions of the same policy, that is visible, factual, and fixable, and it is the single most common finding when a group first looks properly.
• Whether the work has the same shape. Does a task type exist at one property and not at another? Is there a checklist behind it, or is it a title with nothing under it?
• Whether escalations close or merely age. An escalation nobody closed is not a satisfaction problem or a service problem. It is an ownership problem, and ownership is exactly the kind of thing a group office is entitled to have an opinion about.
• Whether a flow is actually running. A flow that was rolled out and is firing at three properties and not the fourth is a fact about the fourth property, and usually a fact about a person there.
• Which departments exist at all. Nine departments on paper and rather fewer in practice is a common shape, and it explains a great deal about why the procedures are not being followed.

Now the other half, which is the harder discipline. A group office cannot see whether a guest was happy. It can see what the guest wrote and how quickly someone answered, which is a different and more honest thing. It cannot see, from a screen, why one property is different from another; it can only see that it is, which is the beginning of a conversation with a general manager rather than the end of one.

The temptation is always to add one more tile that closes the gap. It never does. It only makes the screen slower to read and easier to distrust.

The unit that travels

Here is the argument in one line: the thing a group standardises is not the metric. It is the procedure and the flow.

A metric does not travel between properties, because the properties are different, and everyone in the business knows it, which is why every roll-up gets argued with. A procedure travels perfectly. A flow travels perfectly. “When a guest reports a maintenance fault in an occupied room, a task is opened on the engineering board with the room and the procedure attached, and if it is still open after the stated interval it escalates to the duty manager” is a sentence that means the same thing in a resort and a city hotel.

That is why Treema ships flow templates — 16 of them — rather than a portfolio scoreboard. The group installs the template. Each property runs it with its own people, its own rooms, its own hours. And because every property is now running work of the same shape, the comparison the group actually wanted becomes possible as a side effect: not because a number was normalised, but because the underlying work finally is.

Ask for one place where every property's work has the same shape. That is not a dashboard, and it is worth considerably more than one.

The screen can come afterwards. It will be much better, and much smaller, because by then somebody will know what it is for.