Reading the reaction
Turn a finished Persona Reaction into useful edits while keeping its reading conditions and limits in view.
Read the result from the top down: Reaction score and its headline, conclusions to preserve or address, any available message scores, a drop-off when one exists, and then the details you choose to open.
Start with the reaction
A current scored result shows one headline beside its Reaction score out of 10. The number determines the headline:
- 0 to less than 4 — Negative reaction
- 4 to less than 7 — Tentative reaction
- 7 to 10 — Positive reaction
Reaction score is one overall measure of the Persona’s complete simulated reaction. A higher number means a more constructive simulated reaction for the same Persona, reading conditions, and scoring method. It is not a copy-quality grade, a Quality Check score, or a calibrated probability that recipients will reply.
Start with What to preserve and What to address. Use the info control beside the score for its meaning, bands, and comparison limits. See the line-by-line read opens the complete supporting reaction directly. A finished run may have no specific conclusions.
Separate runs can still differ. When unpitch can reuse an exact completed result for the same version, Persona, and reading conditions, simulating again reopens it without another charge. Otherwise, it starts a new 40-credit run.
Older results retain their original score and outcome. Scoring changed in History marks a boundary where scoring methods differ. Compare revisions only when the Persona, reading conditions, and scoring method match.
Warm, unknown, and supplied-history reads are qualitative. They show the reaction and conclusions without scores or numeric rulers. The relationship is a supplied test condition; it does not establish current need or openness.
Read the message scores
A scored single message gets one Reaction score ruler from 0 to 10. A scored sequence gets one independent ruler for each step, with that step's score and reaction. Step scores reflect the reaction when each message arrived; the headline Reaction score describes the complete sequence and is not their average.
Every ruler uses the same 0 / 5 / 10 reference marks. The markers help locate changes in the modeled response across messages; their gaps are not changes in reply probability. Older results retain their original scoring explanation. A scoreless context read keeps its qualitative outcome and conclusions without adding numeric rulers. If attention drops, a separate block below the scores names the step and gives a reason. It is not another clickable finding. Use that point to find the first change that lost the reader instead of rewriting every step at once.
Separate what landed from what caused friction
What to preserve shows the conclusions that helped. Keep those while you revise.
What to address uses Must fix, Should fix, Worth fixing, and Optional for friction. In a sequence, each card puts its scope in the explanation as Step N or Across sequence, so different scopes can stay in one list. A single message does not invent a Step 1 label. Read the full explanation, then decide whether it fits your case.
A finished run can have no specific positive or negative findings. That is still a result; use the outcome, reason, and line-by-line read for context.
Connect conclusions to the words
Select a grounded conclusion to highlight the related wording and open its contextual card on the message. The card gives a Reason and may include the Reader thought behind it.
Some findings cannot be placed on one exact passage, such as a message-wide reaction or wording that appears more than once. Those cards open from the panel without an inline mark.
Persona Reaction conclusions do not have Not a problem or revision requirements. They report a modeled reader's response; you decide what, if anything, to change.
Read the message line by line
Choose See the line-by-line read to open Detailed reading. It follows the reader's thoughts and source quotes in message order. A tag appears only when a thought supports a finding, so an untagged thought is not automatically good or bad. Scored sequences also retain each step's reaction and decision with its actual step number. A single message has no invented step row.
At the end, Reader perspective keeps the full verdict, outcome reason, and reading arc. These are modeled reactions under the saved conditions, not observed buyer behavior.
Use Back to their reaction to return to the same result. History uses Back to Simulation result. Choosing another saved result opens its overview.
The overview keeps the conclusions, scores, and drop-off together so you can choose an edit before opening the supporting read.
Take the next action
The result keeps one primary feedback action pinned at the bottom. A fresh result with revision support shows Revise using feedback; stale and History results, or a result without that support, show Copy all feedback. Current may also offer Mark ready as a separate secondary decision. The modeled reaction does not choose that decision, and revising does not promise a particular score or real-world result.
Check how it was read
The reading stamp stays above both the result and detailed reading. Use its info control to see which conditions were frozen for the run. Score interpretation is available beside the score.
With Cold contact and no added context or reply scenario, the result states its Fixed Persona Reaction assumption: actual need and timing are unknown, while the read assumes a relevant need, openness to considering a fit, and that the recipient is not actively shopping. This is a model assumption for the run, not a fact about the saved persona or a real buyer.
For a context read, the disclosure retains the exact relationship and note frozen at launch. Actual replies remain supplied sender claims; hypothetical replies remain scenario assumptions. Their provenance also appears beside the affected step in the line-by-line read. A saved relative date keeps its original assertion-time reference.
Older results retain their original document-context claims and reading conditions; they are not relabelled as new cold/warm reads.
For a planned sequence, the disclosure explains its no-reply assumptions. A supplied reply applies immediately before its chosen step and is not invented for earlier steps. In a LinkedIn DM Sequence, the original first connection-request item keeps that identity only when it is nonempty and survives into the read. If it is blank or absent, the next item is read as a LinkedIn message; unpitch does not call it a connection request solely because it became first.
The saved persona's exact Known context is also frozen for the run. Changing the Knowledge Base does not alter that context or a completed result, because personas do not read the Knowledge Base.
The outcome is a simulated response under these conditions. It does not show that a real recipient replied or that a reply was qualified.