Reading the reaction

The read has finished and the panel is full of things to look at. What you want out of it is a short list of edits, so here is the order to read it in.

The verdict at the top

The result opens with one word for what this reader would do: Engages, Asks back, Passes, Disengages or Deletes. Beside it is the reply state in plain words, from Would reply through Would reply — curious and Replies to decline to Won’t reply.

Under that sit the score out of 10, a sentence on why they landed there, and Reading arc, which prints the journey through your text as two sentiment words, such as curious then suspicious. The arc is often the most useful line in the block, because it tells you whether you lost them or never had them.

How the score moved

For a sequence of two or more steps, Score trajectory plots the score across the steps. Beneath it, each step has its own stacked readout card with the step and message type, score out of 10, a 0–10 score meter, Reaction, and Would. The score is a numeric model signal, not an engagement rate. Reaction is the categorical read, while Would states the action this reader would take after that step; valid results can differ across all three.

The trajectory always uses the full 0–10 score axis, with a labelled horizontal step axis and grid lines. The same score therefore lands at the same height across runs, so shapes are directly comparable without guessing at a changing scale.

If attention was lost, the heading carries Drop-off at step with the number, and the reason is printed under the chart. A single message uses a labelled Message score scale on the same 0–10 range instead of a one-point trajectory; any drop-off reason also appears in an Attention dropped callout.

What landed and what caused friction

Below that, the positives are collected as a count of moments that landed, each tagged Keep. A caption says where they came from: Across the whole message, Specific trigger in Step with a number, or a list of several steps.

The friction findings are grouped the same way, by step, with anything about the message as a whole gathered under Overall and always listed last. Each carries one severity word, from Must fix through Should fix and Worth fixing to Optional. Work top down, and clear the Must fix findings before you send.

If the read produced neither positives nor problems, you get one line instead of these sections: Your reader got through this without a specific reaction either way. That is a result, not a failure. It usually means the text gave the reader nothing to react to.

The highlights on your draft

Clicking a finding highlights the words it is about in the draft and opens a card there, carrying a Reason giving the one-sentence explanation and a Reader thought box with what the reader was thinking at that line. The card does not repeat the quoted words while the draft is marking them.

A few things about the marks are worth knowing:

• Only one panel's findings are painted at a time. Switching to Quality Check replaces every Simulation highlight with that check's own, and switching back swaps them again.
• A finding can have no highlight at all: the quoted words no longer appear, they appear more than once (a mark is left off rather than risk pointing at the wrong copy), or the finding is about the whole message. Its card then opens beside its panel row and carries the quoted words, unless the finding is about the whole message and so has none.
• Edit a sequence after a run and its Simulation findings become panel-only, because their step can no longer be identified reliably.
• Typing in the draft closes an open card. That is the intended way to close it and carry on writing.

There is no Not a problem control on a Simulation finding. That belongs to Quality Check.

The line-by-line read

See the line-by-line read opens the ordered walk-through, captioned Full sequence for a sequence and Your message for a single body. It is headed HOW THEY READ IT, and each row quotes the exact words first with the reader's thought about them underneath, in the order the reader met them. A thought about the message as a whole has no quote above it and stands alone.

Most thoughts carry no tag. A tag appears only on a thought a finding was grounded in, and where one thought carries two findings the more severe tag wins. So an untagged thought is not a minor one, it is one that did not become a finding.

Their reaction in the header takes you back to the summary. While the walk-through is open, the reader header shows Line-by-line read underneath.

What the run assumed about you

At the bottom, Context used for this read lists what the run took as given about your business, captioned Unconfirmed assumptions, each with a word for how certain it was. Nothing there was confirmed by you. If something is wrong, the reaction above is answering a different question than you asked, and fixing your Knowledge Base beats rewriting the draft.