Why dollars-per-million beats pips
The unit you quantify a leak in decides whether it gets fixed. A short argument for the desk's language.
Two ways to report the same problem.
The first: the markup on this group drifted by half a pip. The second: this group is leaking eleven thousand dollars a month. Same finding. One gets nodded at and filed. The other gets fixed before lunch.
The difference is not the problem. It is the unit. And the unit you choose decides whether anyone acts.
Pips are a measurement, not a decision
A pip is a fine way to measure a price. It is a terrible way to motivate a fix. It is abstract, it is per-unit, and it hides the two things that drive action: how much money, over how much volume. Half a pip sounds like rounding. Nobody clears their afternoon for half a pip. The number is technically correct and operationally inert.
The trouble is that most surveillance speaks in pips, because pips are the unit the data arrives in. The spread is in pips. The markup is in pips. So the finding comes out in pips and lands on a desk that has to do the conversion in its head to work out whether it matters. Most of the time the conversion never happens, and the finding dies on the page.
The desk thinks in money
Walk onto any dealing desk and listen to how risk actually gets discussed. Nobody says we are exposed forty pips. They say we are carrying eighty thousand dollars on that book. The desk thinks in money, per unit of volume, because that is what a broker is. A leak is only real when it has a dollar sign and a rate.
Dollars per million is that language. It takes a per-unit price difference and ties it to the thing that scales it, volume, then expresses the result as money. A spread drift of a fraction of a pip becomes a number of dollars for every million of notional that trades through the group. Now it is not a measurement. It is a rate of loss. And a rate of loss is something a desk knows exactly what to do with.
Why the unit is the whole game
Surveillance only matters if it ends in a decision. That is the entire point of watching anything. A finding that does not move someone to act is not a finding, it is a log entry. And the single biggest determinant of whether a finding moves someone is whether it arrives in a unit they already use to make decisions.
Report a leak in pips and you have handed the desk homework. Report it in dollars per million, or in dollars a day across the book, and you have handed them a decision. Same data, same accuracy. One of them gets actioned, because it speaks the language of the person who can fix it.
Take the reverted swap. In points, it is a small number on a financing table. In dollars a day on every open position in the group, it is an emergency. Take a margin tier that loosened. In leverage ratio it is an abstraction. In dollars of additional exposure on the book, it is a limit breach. The finding did not change. The unit made it legible.
The argument, in one line
Measure in whatever unit the data arrives in. Report in the unit the person who can fix it already thinks in. For a broker, that unit has a dollar sign in front of it and a volume behind it. Everything else is a number waiting to be translated, and most numbers that need translating never get read.
Pips are how you measure the price. Dollars per million is how you get it fixed.
Every Broker Intelligence finding lands the same way: the exact symbol, the dollar exposure and the timestamp. Quantified to the dollar, so it ends in a decision and not a log entry. Book a 30-minute walk-through.