Two prices for one day

The first cost report of Ramadan says the month is expensive. It says that every year, and it is true the way an average is true.

Look at the hours and the sentence comes apart.

The afternoon is dead and cheap. Almost nobody is bidding into it, almost nobody is buying out of it, and the attention is available for very little because it is worth about that much. The night is crowded and expensive. Every budget in the market arrives inside the same two hours, after the table has been cleared, and the price says so.

The daily average sits between the two and describes no hour that exists. It is a number about a day nobody lived.

From the platform side, what interests me is not the inaccuracy. It is the decision the average produces. The report goes up flat. Somebody senior reads one line, the month looks inefficient against February, and the budget is trimmed for efficiency. The trim then lands evenly across a day that is not even, so the cheap hours are cut alongside the expensive ones, and what is left is a plan spending least in the hours where it could buy most.

The accounts that have a good month are the ones that stopped reading the day as one number by the second morning. They do not need a better model. They need the report cut at sunset. Two blocks, two costs, two decisions, the same money.

For the other eleven months an average is a fair summary of a day, which is why nobody questions it. For thirty nights the day has two halves with almost nothing to do with each other, and the summary is the one line in the report that is hiding something.