Don't Just Read Comments and Move On—The "Feedback Ledger" That Turns Questions, Corrections, and Counterarguments into Future Assets
When you publish a video, comments arrive.
"Who is this system for?"
"That figure has changed in the latest documentation."
"Isn't there another way to look at this?"
You read them on the spot and reply if necessary. But a week later, they are washed away by comments on new videos.
The next time you cover the same topic, you get the same questions again. You forget the points raised last time and end up searching for materials from scratch.
This means the information received from readers isn't feeding back into your production process.
The comment section is not just a place to see reactions, but also a place to discover gaps in your research.
In the previous article, I organized how to choose one piece to reactivate this week from the ledger as Before Creating New Work, Put Past Work to Use—The 30-Minute "Content Asset Review" method.
This time, we will create a "Feedback Ledger" to return questions, corrections, and counterarguments received after publication back into your research pack, connecting them to your next article or video.
You don't need to save every comment
If there are a hundred comments, transcribing all hundred into a ledger is not sustainable.
This operation will not last.
Comments like "That was interesting" or "I'm rooting for you" provide the strength to keep creating. However, their role is different from material that updates your research pack.
What you return to the ledger is information that has the potential to change future production or fact-checking.
The explanation was insufficient, and the same question is being repeated
There is a specific correction regarding numbers, dates, names, or systems
Materials or counterarguments that challenge the premise of the article are presented
Information not found in public documents has been provided by stakeholders or practitioners
Multiple readers have brought up the same topic they want to know about next
Do not choose based solely on the magnitude of the reaction.
A single point of criticism accompanied by a source URL is more valuable for research than a definitive statement that has received many 'likes'.
Categorizing feedback into four types
Before adding them to the ledger, categorize the role of each comment.
Questions
These are points the reader could not understand or conditions that were not fully explained in the article.
If it is just one instance, it might be an individual question. If the same question persists across different media, it may indicate a lack of clarity.
Do not just end with a reply; use it as a candidate for additions to the original article, an FAQ, or supplementary material for the next video.
Corrections
These are verifiable points such as numbers, dates, titles, citations, or the scope of a system.
Check these with the highest priority.
However, do not accept the content written in the comment as the absolute truth. Verify the original source, the latest version, the relevant page, and the publication date; if the comment is correct, fix both the master document and the published material.
Counterarguments
These are not factual errors, but rather disagreements regarding evaluation, causality, or premises.
Do not dismiss them by saying 'we have different opinions'; instead, distinguish whether there are facts we haven't addressed or if there is a difference in how the same facts are evaluated.
If there is new evidence, return to the research phase. If it is a difference in value judgment, do not force a single conclusion.
Signals of Demand
These are voices that lead to the next project, such as 'I want to know more about this part,' 'I want you to compare it with other municipalities,' or 'Please teach me how to read primary sources.'
Do not decide on a project based on a single request; prioritize it by looking at search trends, related news, and connections to existing assets.
Eight items to include in the feedback ledger
If you add too many items, you will run out of time just organizing comments.
Eight is enough to start with.
1. Original Asset ID
Record which research pack, video, or article the reaction is regarding.
If you save comments individually, you won't know what they were about later on.
2. Source and Date Received
Write down the medium, such as YouTube, note, X, or email, and the date it was received.
For public comments, also keep the URL. Since they might be deleted or edited, record the original text to the extent necessary.
3. Key Point
Instead of the full comment, summarize what needs to be checked in a single sentence.
'The numbers in the document are wrong' is not enough.
Turn it into a form that can be acted upon, such as: 'Pointed out that the article uses 2025 fiscal year figures, but there are updated values in the document published in July 2026.'
4. Classification
Choose one from: question, correction, rebuttal, or signal of demand.
If it applies to multiple categories, prioritize the reason for the initial action. If it is a project request that includes a numerical error, prioritize the correction first.
5. Status of Evidence
Record in four stages.
Unverified
Partially verified
Evidence verified
Cannot verify the basis
This is to avoid conflating "a viewer pointed this out" with "confirmed as a fact."
6. Reflection in the Master Copy
Write down where you made changes across the review packet, the corrected SRT, and the confirmation memo.
If you only reply without correcting the master copy, the same error will return in the next production.
7. Handling Public Content
Choose one course of action.
Reply
Add an update to the article
Issue a correction
Fix the description or pinned comment
Address it in the next article or video
Take no action
Even when "taking no action," leave a reason. Examples include being unable to verify the basis, the same claim being repeated, or the point being outside the scope of the article.
8. Completion Date and Public URL
Record the date of the correction and the page where readers can verify it.
Leave the URL of the public page, not the editing screen.
Verify corrections in three stages before accepting them
The more specific the feedback, the more you want to fix it immediately.
However, the person providing the feedback might be looking at outdated information. They might also be confusing similar systems.
Check corrections in the following order.
Verify the basis upon which the original article was written.
Confirm that the cited material actually exists and supports the relevant section.
Check which is newer, and whether the subject, time period, and definitions are the same.
The result is not always a binary choice between "correct" and "incorrect."
The numbers in the article may be correct, but the explanation of the target period may be insufficient. The cited material may be newer but based on preliminary figures, while the original article uses final figures. In such cases, you need to add conditions rather than correct a typo.
Handling corrections is not about winning or losing an argument in the comments, but about returning the content to a state where readers can trace the evidence.
Do not bring personal or internal information into the ledger.
Information from involved parties is valuable. However, just because it is detailed does not mean it can be saved or published as is.
Do not record names, contact information, workplaces, etc., more than necessary.
Do not quote private messages without the sender's permission.
For anything that appears to be internal material, separately verify its source, potential for public disclosure, and whether it has been altered.
Do not determine facts based solely on comments regarding whistleblowing, crimes, or allegations of misconduct.
Handle personal experiences that could lead to identification if published with caution, even if summarized.
The feedback ledger is not a repository for personal information.
What is needed is "what should be verified." If there is no need to keep track of who said it, record only the anonymized key points.
This is what happens when a single piece of feedback is fed back into the production process.
For example, suppose a comment to the following effect is received on a video covering a local government support system:
"The video explains that the target households are A, but in the document released this month, B has also been added to the target."
At this point, it is still unverified information.
The ledger processes items as follows.
Original Asset ID: Support System Research Pack
Category: Correction
Key Point: Pointed out an additional change to eligibility criteria
Status of Evidence: Unverified
Verification Task: Check the municipal system page, guidelines, and update date
Verification Result: Confirmed addition of Target B in the new guidelines
Reflection in Master Copy: Updated verification memo and system conditions
Action for Public Content: Added to the summary section and article, and replied to the comment with thanks and the update location
By doing this, a single comment performs four jobs.
It prevents misunderstandings. It updates the master copy. It fixes public content. It leaves an update history that can be used for the next system explanation.
This is the difference compared to just replying to comments.
Twenty minutes a week, look at only three unprocessed items
The feedback ledger will also stall if you try to process everything every day.
Once a week, choose only three items from the unprocessed list.
The priority order is as follows.
There is a possibility of factual errors in published articles or videos
The same question has been received multiple times
There is a rebuttal with evidence that affects the premise of the article
It can be answered in a short time using an existing research pack
Spend the first five minutes selecting three items. Spend the next ten minutes confirming the location of the materials. In the final five minutes, decide whether to route them to updating the master copy, correcting public content, or planning the next project.
You do not need to reply to every item on the spot.
Once you have decided what to check and where to return the information, the ledger is working.
The number to watch is not the number of comments, but the number of items returned to the system
A video with many comments is not necessarily being turned into an asset.
The number to look at is the following:
The number of questions, corrections, and rebuttals registered in the ledger
Number of cases where evidence was verified and incorporated back into the original document
Number of cases where published articles or summary sections were updated
Number of FAQs or future project ideas created from the same questions
Number of cases where a response was provided to the reader
Number of cases not adopted due to lack of evidence
There is meaning even in records that were not adopted.
Even if a point is made with strong language, if the evidence cannot be verified, it is not used as fact. By documenting this decision, you can avoid being swayed by the same information later on.
After publication, the next investigation begins
Articles and videos are not finished the moment they are published.
Readers have questions. They point to other materials. Involved parties point out what is missing. Opposing views arrive.
From there, you can see the weak points in your explanation and the issues that need to be investigated next.
Turn questions into FAQs. Turn corrections into original document updates. Turn counterarguments into a re-verification of premises. Turn requests into candidates for future projects.
With this feedback loop, the comment section becomes an input that improves future production rather than just a place for fleeting reactions.
To start, pick one recent video or article and try picking up just three comments that require verification.
Next time, we will address the problem of readers not knowing "where and why" an article changed even after it was corrected, and consider a "Public Update History" that turns corrections and additions into trust.
Have the questions or corrections you sent to a creator ever been reflected in their subsequent articles? Please let me know in the comments.
I explain the "mechanisms of money and power" behind political and economic news.
▼Read also
・Before creating new work, put past work to use—The 30-minute "Content Asset Review"
https://note.com/poliplus/n/n11699fa6661b
・Don't just "make it and forget it"—The gateway to the PoliPlus-style "Content Assetization" series
https://note.com/poliplus/n/n0b78ceaa230c
・How do critical comments disappear from a public figure's official YouTube channel?—A record of verification using multiple accounts
https://note.com/poliplus/n/ncf80115028f1
・You made it, but it can't be found—The "Content Asset Ledger" that manages videos, SRTs, and articles in one place
https://note.com/poliplus/n/nb66864e6de5c
・Those who make one thing and stop, and those who leave seven behind—PoliPlus-style "One Output, Multi-Use"
https://note.com/poliplus/n/n1a42855360b6
▼Summary of this theme
・Summary of Politics, Money, and SNS Verification
いいなと思ったら応援しよう！
よろしければ応援お願いします！ いただいたチップはクリエイターとしての活動費に使わせていただきます！