A delivery dashboard can be correct and still leave the important customer consequence difficult to see. In a 2023 cohort at our leaflet business, ninety-six of 120 orders reached customers by their agreed dates. That produced an 80% on-time order result. The arithmetic was straightforward and the measure was familiar. Commercial colleagues nevertheless said the review sounded much more reassuring than their customer conversations. Several large orders remained late. Their objection was not that the dashboard had invented completed jobs, but that it had given equal numerical weight to orders with very different quantities.
The ninety-six on-time orders contained fourteen million leaflets. The twenty-four late orders contained sixteen million. Across thirty million units, approximately 46.7% were on time. Both percentages described the same reviewed cohort. One counted commitments as orders; the other weighted the result by leaflet quantity. Neither alone described every consequence, but the difference was too large to leave out of the discussion. I asked for both views with their units clearly stated, rather than choose a replacement percentage simply because it would now demonstrate that management was being more demanding.
A valid percentage can give an incomplete account
The operating team had a reasonable defence of the original report. It used an established measure and had counted the orders correctly. Changing measures whenever the result became uncomfortable could undermine comparability and encourage another form of selective reporting. I agreed with that concern. The answer was not to erase the existing series. It was to add the quantity view and explain what each measure represented. A consistent measure can remain useful while an additional view reveals a concentration the original was never designed to show.
The distinction also prevented us from treating large orders as the only ones that mattered. A small urgent job may have a serious effect on a customer's packing plan. Unit weighting cannot tell us that consequence by itself. If we used the volume percentage as the sole replacement, we could improve the headline by prioritising a few large jobs while allowing many smaller promises to slip. The order view remains a check on that possibility. The management task is to examine the distribution and the commitments, not to discover one number that will remove the need for judgement about customers.
I wanted the discussion to identify which late jobs accounted for the sixteen million units and what their customers were waiting to do. That is an investigation into operating and commercial context, not a mathematical adjustment to the result. Some causes may sit within production; others may involve an approval or an agreed change. Those distinctions need evidence. They should not be inferred from order size or used to excuse every delay. The paired measures tell us where to ask more carefully. They do not establish the cause of lateness or decide which explanation a customer should accept.
Keep the commitment fixed while widening the view
We preserved the original agreed dates in the cohort. A revised internal schedule after a delay may be necessary for managing recovery, but it should not quietly replace the date against which the original promise is being assessed. Otherwise, a report can improve without any leaflet reaching a customer sooner. I wanted recovery planning and commitment performance to remain distinguishable. One tells colleagues what should happen next; the other tells them whether the business did what it said it would do. Both can be useful without being collapsed into the same field.
Partial quantities also require care in a real operating account. Our illustrative cohort states the on-time quantities and orders used in the comparison; it does not offer a universal rule for every split delivery or amended contract. The terms of a particular commitment still need to be understood. Where management changes a definition, the change should be visible and its effect explained. A director should be able to follow what happened to the customer promise without having to guess whether an apparently improved rate came from production, a different cohort or a revised counting convention.
This is why I resist treating measurement as a purely technical dashboard exercise. The person choosing the unit shapes the conversation about responsibility. An order count may make a team with many small completed jobs appear stronger. A quantity count may make a team handling a few large jobs appear stronger. Neither effect proves misconduct or invalidates the measure. It does mean that the leadership account needs enough context for another person to challenge the interpretation. Commercial and operating colleagues should be able to disagree about the conclusion while working from the same disclosed quantities and dates.
A better report is the beginning of the work
The immediate result of our correction was a less comfortable review. We had to discuss why so much of the leaflet volume sat in the late group. That was useful, but it was not a delivery improvement. The sixteen million late units did not become timely because the report described them more accurately. I have not attached an invented recovery rate to this case. The account ends with the corrected view and the need for investigation. A leadership story can be complete about a decision while remaining honest that the next operating result is not yet part of the evidence.
For a board, the relevant contribution is to test how performance measures connect with the obligations they are meant to illuminate. Directors need not redesign every dashboard or allocate the next production shift. They can ask what is being counted, which important distribution the average obscures and whether the original commitment has been retained. They can also ask whether management is confusing improved information with improved performance. Those questions help the board assess the enterprise without taking over the work of the people who must diagnose and correct the operational problem.
For me, the case changed the tone of an operating review. A challenge from commercial colleagues was not an attack on the accuracy of the production team's arithmetic. It was evidence that the chosen view did not carry enough of the customer consequence. Keeping both percentages allowed us to preserve the valid measure while widening the account. That is the kind of leadership I want the profile to demonstrate: the willingness to make a familiar report less flattering when the business needs a better question. Ninety-six completed orders are a real achievement, but they should not make the larger unfinished promise disappear.