Commenting for Clients: A Social Media Manager's Workflow

Commenting on behalf of a client is a different job from commenting for yourself, and the difference isn't volume. It's that you're writing opinions in public, under someone else's name, in a voice that isn't yours, on posts you didn't choose. Get it right and nobody notices, which is the point. Get it wrong once and it's a screenshot in your client's inbox.

The three failure modes are always the same: every client ends up sounding like you, someone comments from the wrong account, or a comment lands on a post the brand had no business touching. This is a workflow built to close all three, based on how the social media managers using Comment Generator actually run it. One of them describes needing to comment on ninety accounts a day, so the constraints here are real rather than theoretical.

The real problem: your clients start sounding like you

This is the one that creeps up on you. You write well, you write fast, and after two months every client's comment section has the same rhythm: same opener, same way of conceding a point, same closing question. Each comment is fine. Read as a set, they're clearly one person wearing five name tags.

Nobody catches this by trying harder in the moment, because in the moment each comment reads fine. You catch it by making the voice an artifact that exists outside your head, and by checking the set rather than the instance.

Write a one-page voice brief per client

One page, and get the client to sign off on it once so you never have to ask them about individual comments again. What it needs: three real comments they'd be happy to have published, which is worth more than any adjective they'll give you. The formality level, stated plainly. Emoji policy, yes or no, and which ones if yes. Whether the account speaks as a company ("we") or as a person ("I"), because mixing those two is the fastest way to sound off. Any banned words, and every client has some. Whether comments should ask questions. And a sign-off, or explicitly none.

The reason this page pays for itself: most of it goes straight into the extension's standing instructions field, so the voice is applied at the draft stage rather than being something you remember to edit in. Set it once per client, and drafts arrive already sounding roughly right instead of sounding like you.

Audit the set, not the comment

Once a week, open each client's comment history and read the last twenty in one sitting. Repetition that's invisible one comment at a time is obvious in a stack. When you spot a tic, fix it in the standing instructions rather than in your own habits, because the instructions don't get tired on a Friday afternoon.

Use a separate browser profile per client

Not separate tabs. Separate Chrome profiles, one per client, each with its own window. This solves two problems at once, and it's the single highest-value change most people reading this can make today.

The first problem is the disaster scenario: posting a comment from the wrong account. If client A and client B are both logged in as tabs in one window, this is not a question of whether but when, and it's unrecoverable in the sense that the comment was public before you noticed. Separate profiles make it structurally impossible rather than something you're careful about.

The second is that extension settings live in the browser profile. Language, length, tone, closing line, standing instructions: all of it is per profile. One profile per client means each client's voice settings are simply loaded whenever you're in their window, with nothing to switch and nothing to remember. Name the profiles after the clients, pin the extension in each, and the correct setup is the default state.

Batch by client, never by feed

Work one client for a full session, then close the window and open the next. The temptation is to work the way the feed presents itself, hopping between accounts as interesting posts appear, and it feels efficient. It isn't. Voice is a mode you get into, and every switch costs you a few comments' worth of accuracy at the start of the next block. Twenty minutes per client, in one voice, beats an hour of jumping.

It also makes the work measurable, which matters when you're billing. You find out that client A takes twelve minutes for ten comments and client B takes twenty-five, and that difference is real information about how you should be pricing them.

Agree a never-comment list before you need one

This is the item that prevents the bad screenshot, and almost nobody has it written down. Get it agreed in the same conversation as the voice brief, because the time to decide whether the brand comments on a competitor's product launch is not the moment you're looking at a competitor's product launch.

The default list worth proposing: nothing political, religious, or on live news events. Nothing on a post announcing a death, a serious illness, or redundancies, unless the client has personally approved that specific one. Nothing on competitor posts, in either direction, because agreement looks weak and disagreement looks petty. Nothing on a post where one of the client's own customers is complaining, because that's a support conversation and it belongs to the support team. And nothing on posts about pricing, legal matters, or unreleased products, where a throwaway line in a comment can be quoted back later.

The value isn't only in avoiding mistakes. It's that a written list means you can hand the work to a junior team member or a freelancer without them needing your judgment, which is the only way this scales past you.

What should you report to the client?

Not comments delivered. That's an activity number, it looks like effort rather than outcome, and it quietly trains the client to value volume. If your report says 400 comments this month, next month's conversation is about why it wasn't 500.

Report four things instead. Replies received, which is the honest measure of whether the comments were good, since a comment that gets a reply was read. Profile views over the period, which is the leading indicator of commenting working at all. New followers or connections. And conversations started, meaning threads and direct messages that came out of a comment, listed individually with a link. That last one is what gets the retainer renewed, because it's the only one the client can see the value of without you explaining it.

Add a short qualitative note on what the account engaged with and why. It's the part that reads like a strategist rather than a subscription.

Where the review step fits when you're not the client

Every draft still gets read before it posts. The difference in agency work is that you're reviewing as someone else, against a brief, and "would I say this" is the wrong question. The right one is whether the client would be comfortable seeing it quoted with their logo next to it.

Comment-by-comment client pre-approval sounds safer and doesn't survive contact with reality. Comments are time-sensitive: a post is worth commenting on for a few hours, and an approval queue with a one-day turnaround produces comments that arrive after the conversation ended. Approve the rules, not the instances. The voice brief and the never-comment list are what the client signs off on, and then you work inside them. That's the same human-in-the-loop principle covered in why AI-drafted comments still need a human in the loop, with the reviewer wearing a client's brief instead of their own judgment.

Practical notes on running several accounts

Keep the volume per account to something a person could plausibly have read and written. Platform spam detection responds to bursts and to repetition, and both are things you cause by working faster rather than by working more, so the fix is spreading a client's session across the day rather than firing twenty comments in four minutes. What actually raises risk, and what doesn't, is worth understanding properly rather than guessing at.

On cost: pricing is per platform rather than per account, and the Unlimited plan is the one aimed at people running several accounts, so a LinkedIn-focused agency needs one subscription for LinkedIn rather than one per client. If you also run Instagram work, that's a second platform. The extensions are Instagram, LinkedIn, and Threads, and most agency work in practice sits on the first two.

None of the above is about commenting faster. Speed is the easy half and a tool handles it. The hard half is five distinct voices, five risk profiles, and one person accountable for all of them, which is a systems problem rather than a writing problem. Write the briefs, split the profiles, agree the never-comment list, and the writing takes care of itself.