18 June 2026

Reply counts beside spend, without forcing them onto one axis

Baht and reply counts do not share a natural scale. Drawing them on one axis asks the reader to trust a coincidence.

Two horizontal ivory strips on a lacquered table, one with bars and one with small copper marks

A dual axis is tempting when a counter brand wants to see live-selling hours, spend, and paid orders in one glance. The lines will cross somewhere. The crossing looks like a finding. It is usually the place where two unrelated scales happened to meet.

Paired rows, shared time

The sheet we use splits the story into rows that share time buckets and nothing else. Spend is a row of bars in baht. Replies, or paid orders, sit on the next row as marks in their own unit. The eye travels down a Tuesday and sees both facts without being told they are the same kind of fact.

Time buckets have to be agreed. For a live-selling counter, an hour may be the true bucket, but only for the hours the seller was actually present. Empty hours, drawn as honest gaps, stop a quiet Thursday from being averaged into a busy Sunday. A weekly total is what flattered that Thursday in the first place.

What the note is allowed to say

The written note may say that orders clustered in two hours and that spend was spread across six. It may not say that a particular hour caused the orders, unless the client’s own log supports that and the comparison hour is fair. Causation is a separate argument, and most of the tables we receive are not built for it.

If you are bringing a sitting or a single chart, send the reply log with its clock times, not a weekly sum. The weekly sum is how a dead hour keeps its place on the roster.

Back to the journal · Ask for a reading date