A ratio is worth exactly what its definitions are worth. Most of the ratios reported in this line of work are defined widely enough that two people can read the same figure and reach opposite conclusions, both of them honestly. The number pipera reports is narrow on purpose, and the narrowness is the point of it.
The measure is revenue divided by directed spend. Directed spend is everything pipera moves: media budget on every platform, production of everything that ships, and the monthly fee. The fee is the part most operators leave out, and leaving it out is the single easiest way to flatter a result. A measure that excludes the cost of the operator producing it is not measuring the operation. It is measuring the operation with its own price removed.
Revenue is read from the commerce platform's own order record. The server side conversion API is a secondary source and comes in when platform data is unavailable. When the two disagree, the order record wins, and the difference goes into the report instead of being smoothed away. Only revenue that leaves a record someone can go and read is counted. Cash taken across a counter, money closed on a phone call, sales on a channel pipera does not run, and orders that predate the engagement all sit outside the number. A business that closes most of its revenue offline is told on the call that the measure will describe a slice of it rather than the whole.
The reason to report this instead of a dashboard is simple and slightly unflattering to everybody involved. A dashboard with enough tiles on it always contains at least one tile going up. Reach becomes the story when sales are flat. Engagement becomes the story when reach falls. A platform's own attributed value becomes the story when orders fall. None of this requires anyone to lie. It only requires a choice about which number to open with, and that choice gets made, week after week, by the party with the most to lose from the wrong one. A single ratio removes the choice. It is the same number when it is good and when it is bad, and it goes into the report in both cases with a note saying what is being changed because of it.
It is reported weekly because a week is short enough that something can still be changed and long enough that a change has produced a result. It is reported per client because there is no cross client benchmark that survives contact with reality. Two businesses with different margins, different repeat rates and different price points do not share a threshold. A ratio that is healthy for one is a slow failure for the other. The only meaningful comparison available is the same business against itself over time, which is why the series matters more than any single reading of it.
The ratio is not attribution science and does not pretend to be. It does not prove that the spend caused the revenue. It states what the business took in over a period, and what was spent under this operation to be present during that period. Whether one produced the other is a separate argument, carried by the decision log rather than by the ratio: what was changed, when, on what evidence, and what moved afterwards. Anyone presenting a ratio as proof of causation is selling something else.
The obvious way to game it is to cut spend. The denominator shrinks, residual demand the operation did not create keeps arriving, and the ratio rises while the business gets smaller. Three things sit against that. The numerator and the denominator are reported next to the ratio rather than behind it, so a rise produced by falling spend against falling revenue is visible in the same view as the rise itself. Every cut is a decision and every decision enters the log with its reason, so a quiet withdrawal from the market gets written down at the moment it happens rather than reconstructed later. And the monthly fee sits inside the denominator, which means the operator's own cost does not shrink when the media budget does, so the cheap version of the trick works less well here than it would elsewhere.
The last guard is definitional. The ratio is an accounting definition, not a target. Nothing is optimised toward it directly, because a measure that becomes a target stops describing the thing it was built to describe and starts describing the effort to move it. The ratio can sit flat for a period. It can be negative. It goes into the report either way, and the report says what is being done about it.
None of this makes the number sufficient on its own, and it was never meant to be. It is one honest reading of a business, bounded on both sides, published on a schedule, with its own limits printed next to it. What it replaces is not analysis. It is the weekly negotiation about which metric everyone has agreed to be pleased with. If a single ratio you can argue with, including in the weeks it argues against the operator, is worth more to you than a screen of metrics, book a call.