Agency Content Approval Workflow and Revision Limits

Most delays in agency content work are not about writing quality. They come from approval. A draft goes to three people, each pulls it a different way, the copy gets blander with every round and the publish date slips. By the end nobody can say which version was approved, which one is live or who had the final word. The problem is an undefined process, not weak writers.

Short answer: capping revisions in the contract makes sense for most agencies, but a number alone is not enough. A cap works only when the contract also says what is being counted (rounds or edits), which requests count as scope changes rather than revisions, and who gives approval. The right figure depends on the type of work, the length of the piece and the client's internal chain of approval.

The buyer's side of a delivered article, which means checking it against the brief and writing feedback for the writer, is a separate subject covered in article revision process and delivery check. This guide stays with agency and client: the approval chain, version records, the revision cap and the final sign-off. The contract suggestions are a working framework; for binding wording, consult your legal adviser.

Capping revisions: what counts and what falls outside

Should revisions be capped in the contract? For most agencies yes, but how the cap is defined matters more than the figure. A badly defined cap either does nothing or starts a fight.

What a cap protects

Unlimited revisions make the agency's cost unpredictable and, for the client, the project never has a finish line. A cap helps both sides: the client sends feedback in considered batches, and the agency can plan the work. There is no standard number. A short piece may need two rounds, and a long guide with many stakeholders three. Those are illustrations only, not a recommendation from any documentation.

Rounds are the better unit

Counting rounds is usually fairer. The client receives a draft, sends one consolidated list of feedback and the agency applies it. That is one round. Three separate emails of comments on the same draft should still be one round, provided only the coordinator's consolidated list is accepted. Counting individual edits is heavy, because a comma change would use up a revision.

Sorting requests: revision, scope change or new round

Run each incoming request through three questions, in this order.

1. Does the approved brief already cover it? If the text falls short of the brief, as with a typo, a wording fix or a statement that conflicts with a source, it is a revision and the agency's responsibility.
2. Does it change something the brief fixed? A different tone, a new section, another target reader or a new product feature that needs a new information source is a scope change. It needs a decision from whoever approved the brief, and possibly a new schedule or quote.
3. Does it touch text that is already locked? Editing approved text during publishing opens a new round and goes in the log.

When a request is unclear, log how you classified it and why.

The round after the last one

Two options can be written into the contract: an extra round at a price or time agreed in advance, or extra rounds limited to typos and accuracy fixes. Choose one at the start. Raising the issue for the first time once the cap is used up strains the relationship. This is a working framework, not contract language; ask a legal adviser for binding terms.

Where a recorded exception is wise

Not every limit needs to be enforced rigidly. On a first engagement, or when a correction exists because of the agency's own mistake, flexibility protects the relationship. Record it: "this round accepted as a courtesy despite being out of scope". Unrecorded flexibility turns into an expectation on the next project.

A nonprofit asks for a different audience

Example scenario: A nonprofit approved a brief for an article aimed at donors. In round two it asks to rewrite the piece for volunteers. The agency does not process this as a revision. It reminds the nonprofit that the target reader changed and offers two options: finish the donor article as briefed, or plan a separate volunteer article as new work. The nonprofit picks the second, and the log shows the request as out of scope. The record keeps the relationship intact.

Client response time as a limit

Define how long the client has to respond, and what follows silence: a schedule shift, deemed approval, or a reminder and extra time. No standard exists; set it per project. If you plan to treat silence as approval, write that plainly in the contract, because it can surprise a client. For the order-to-publication timeline and revision requests on a sponsored article order, see sponsored article turnaround time and revisions.

The approval chain: roles and who decides

An approval chain says who sees a piece before it goes live and what each person is responsible for approving. Without it, everyone comments on everything and the process never ends.

Who holds which role

Most projects run on five roles. On the client side there is a decision maker who gives final approval, a coordinator who gathers feedback, and sometimes a subject reviewer for legal, product or technical accuracy. On the agency side there is an editor responsible for content quality and a project lead who tracks time and records. In small jobs one person can hold two roles. What matters is that the role has a name and that the decisions each one owns are written down.

A single coordinator on the client side

Feedback breaks down most often when several people see the same draft and send comments without seeing each other's. Ask for one coordinator to collect every comment and send a single list. When comments conflict, that person resolves them, not the agency. Otherwise the agency becomes the referee in an internal argument and takes responsibility for each call.

Limiting what subject reviewers weigh in on

Legal, product or technical reviewers check accuracy in their own area. Disliking the headline or rewriting the tone is outside that role. Writing "accuracy in their field only" into the role description removes many unnecessary style rounds.

Mapping Google's who, how and why onto the chain

Google's guidance on helpful, reliable, people-first content (Opens in a new tab) suggests assessing content through who created it, how it was created and why. It says readers should be able to tell who made the content, with bylines where they would expect them and links to author background, and it treats fake creator details such as AI-generated headshots, made-up names or false credentials as a sign of a low-quality page. Your chain can hold a matching record: who the author and editor are, whether automation or AI was involved, and whether the piece was made to help people. Decide in the chain who sets the byline for content published under a client's name.

The sequence from draft to go-live

1. The editor checks the draft against the brief.
2. The agency sends the draft to the client coordinator.
3. The coordinator gathers subject reviewers' comments into one list.
4. The agency applies them and records the round.
5. The decision maker approves the final text.
6. The project lead logs the approval and the version that will go live.

A furniture retailer's buying guide

Example scenario: An agency writes a buying guide for a regional furniture retailer. The retailer's brand manager is the decision maker, a product buyer is the subject reviewer and a marketing assistant is the coordinator. In the first round the buyer wants a competitor's brand removed while the brand manager wants more comparisons. The agency forwards both messages to the coordinator and asks for a resolved list. The coordinator settles it with the brand manager. From then on all comments come through the coordinator and the number of rounds drops.

Brief, drafts and final: one source of truth

Version confusion is the second big problem. If two people work in two files, "the latest" becomes a point of dispute. A simple naming and logging habit removes most of it.

The brief sets the yardstick

The brief is the first document and the standard that revisions are measured against. It should state topic, target reader, tone, sources, what is in and out of scope, word range, how the product may be mentioned and who is in the approval chain. Once the client approves the brief, "we actually wanted something else" is a brief change, not a revision. Writing that distinction down at the start prevents later disputes.

Numbering and dating drafts

Give each draft a number and a date, such as v1, v2, v3. A date in the filename is not enough. Record which comments each version answers. If the team works in one shared document, keep the version history and comment trail, and still assign a version name at the end of every round.

Locking the approved text

When the decision maker approves the final text, that version is locked. A later change becomes a new round, and it is logged. The live page should match the approved version, and any edits made during publishing, such as a changed title or an added link, belong in the log. Otherwise the gap between approved and published text raises the question of who is accountable.

Recording comments you set aside

If a comment was not applied, say why in a short note: conflicts with a source, outside the brief, contradicts another comment and awaits the coordinator. The notes inform the client and also build the agency's own defense. A comment that is silently ignored returns in the next round as a sharper objection.

Noting automation and AI in the record

The same Google guidance says how content was created is a question readers may reasonably ask, and it presents explaining any use of automation or AI as a trust-building practice. It also states that using automation, including AI, mainly to manipulate rankings violates its spam policy on scaled content abuse. If those tools touched the delivered text, note how and where human editing took over in the version log, and say so in the brief and contract. The documentation does not suggest a ready-made disclosure wording.

Light, standard or heavy: scaling the process to the job

The same heavy process for every piece is wasteful. A one-paragraph blog update and a guide that needs legal review should not travel the same chain. Setting up three tiers by type of work saves time on both sides.

Light, standard, heavy

Choose the tier at the brief stage. Making a job heavier midway means adding a rule in the middle of the work, which breeds friction.

Reviewers in parallel or in turn

If subject reviewers see the draft at the same time, the schedule shortens but the risk of conflicting comments rises. Sequential review takes longer, but each person sees earlier comments. A practical middle path is to let reviewers work in parallel while requiring the coordinator to merge their comments into one list. The coordinator, not the agency, settles conflicts.

The agency's internal check

Before anything reaches the client, the agency editor should check the draft against the brief, the source list and the style rules. If the client has to deal with avoidable mistakes, trust drops and rounds are wasted. Add a line to the version log recording that the internal check was done.

A review after the first three pieces

With a new client, apply the process as written for the first three pieces, then hold a short review: how many rounds there were, which comments repeated and which role turned out to be unnecessary. Carry the findings into the next contract or brief template. A review after the first three pieces is a reasonable working suggestion; the documentation gives no interval.

Final sign-off before the page goes live

After the final text is approved there is one short, specific check. Its purpose is not to collect new comments. It confirms that the page going live matches what was approved and that every confirmation has a named owner. Text-level checks on titles, headings and links belong earlier, at hand-off.

Same text on the page as in the locked version

Moving text into a content management system can break formatting. Compare the live page with the locked version and log every change made during publishing. Name who does the comparison; the project lead is a natural owner, and the decision maker signs off only on the list of differences.

Accuracy and sources: who answers for what

Google's helpful content guidance asks whether informational content has easily verified factual errors. For topics that could affect health, finances, safety or well-being, it says content should be highly accurate and consistent with established expert consensus. If the client operates in a sensitive field, confirm that an expert approved the piece. Each statistic, product claim and quotation should be checked with its source. In the chain, the editor ties each statistic and quotation to a source, and the client's subject reviewer confirms product and field accuracy.

Paid or sponsored links

If the agency's text will appear on another site with paid or exchanged links, Google's spam policies (Opens in a new tab) list, among link spam examples, sponsored articles with optimized anchor text that are not qualified with nofollow or sponsored. They also describe scaled content abuse as generating many pages mainly to manipulate rankings. At the last look, ask whether a paid or sponsored link carries the right attribute, whether the anchor text reads naturally and whether the text was written for readers rather than for rankings. None of this guarantees any ranking.

Dates

The guidance warns against changing page dates to appear fresh when content has not substantially changed. Tie the publication date to the real launch day and the update date to genuine updates only. If a client wants an old article to look new, suggest actually updating it instead.

Pre-publish checklist

The project lead confirms

• The live text matches the locked version, with any differences listed.
• The publication and update dates reflect reality.
• Someone is named to check the page after launch.

The editor confirms

• Every statistic, claim and quote traces to a source.
• Any use of automation or AI is recorded, and shown on the page where appropriate.
• Paid or sponsored links carry the right attribute and the anchor text reads naturally.

The subject reviewer confirms

• For a sensitive subject, an expert has approved the piece.

The decision maker confirms

• The byline is accurate and tied to a real person.
• The list of differences, if there is one, is acceptable.

Related reading on hand-off and responsibility

For quality checks on the text at hand-off, the content quality checklist before you publish applies; it works at text level, while this guide follows the agency and client flow. For how responsibility is split among brand, agency and publisher, see the sponsored article approval and responsibility matrix.

The approval log: fields and wording

The approval log is the proof that the process happened. In a dispute, or on the next project, it answers "what happened". Keep it simple, since a complicated log does not get kept.

What the log holds

Wording for the approval message

Get the decision maker's written approval through a brief message: "I approve version [name] of [title] for publication. I understand that change requests after publication will be handled as a new round." Attach the email to the log. A verbal approval invites a later "I never said that".

Corrections once the page is live

Write into the contract how post-publication corrections work: how long a request is accepted and which kinds of errors are fixed without charge. For sponsored posts, the guide to sponsored post correction requests and disputes goes deeper, and the writer and editor workflow in an SEO article writing service shows how to organize the production side.

For the agency side

• Name the approval chain with real names at the kickoff meeting.
• Send every draft through one channel to one coordinator.
• Show the round count on each delivery, for example "this is round two".
• Log out-of-scope requests the same day.
• Get final approval in writing and store the locked version.

For the client side

• Let one person collect all comments.
• Send feedback in one considered batch; piecemeal comments use up rounds.
• Read the brief carefully before approving it, since later changes may be out of scope.
• Define the scope of subject reviewers in advance.
• Know that change requests after publication count as a new round.

Stalled approvals and what to do about them

Even a defined process meets trouble. These are the most common cases and practical responses.

Endless rounds

If every round brings new comments, the problem is probably in the brief, not the text. The brief may be unclear or the decision maker uncertain. Stop, return to the brief, write down the missing decision and then continue. Polishing the text further only enlarges the problem.

A comment after approval

A comment that arrives after approval breaks the lock. If it is a small accuracy fix, log it and apply it. If it is a larger change, open a new round. Either way, keep a record.

No reply from the client

A client who does not reply moves the schedule. Put the response time in writing and send a reminder before it expires. Reminders belong in the log too.

Ranking expectations

Some clients see the approval process as part of a ranking guarantee. Google's Search Essentials (Opens in a new tab) says that appearing in results is free and that meeting every requirement does not guarantee crawling, indexing or serving. The point of an approval process is quality and clear accountability. Promising rankings is not its job.

Frequently Asked Questions

Does a change request after brief approval count as a revision?

Usually not. The approved brief is the standard that revisions are measured against. A new target reader, a different tone or extra topics are brief changes and should be handled as out-of-scope requests. Write that distinction into the contract and the log from the start.

Should the contract cap the number of revisions?

For most agencies it makes sense, because unlimited revisions make costs unpredictable and drag out the project. A cap alone is not enough, though. Also write what is counted, which requests are out of scope and what happens when the cap is reached. The right number depends on the type of work. For binding wording, ask a legal adviser.

Does one round equal one individual edit?

No. A round is the agency applying one consolidated list of feedback from the client. Counting individual edits is heavy because even a comma fix would use a revision. Counting rounds and asking one coordinator to collect all comments is usually fairer.

Does silence from the client count as approval?

That depends on the contract. The documentation has no rule on it. If you plan to treat silence as approval, write the deadline, the reminder and the outcome plainly, otherwise it can surprise the client.

Is a post-publication correction a new round?

A typo or a statement that conflicts with a source is normally corrected without charge. A new request that changes the content means altering the locked version and counts as a new round. Write into the contract which kinds of errors are fixed free and within what time.

Must I tell the client if AI helped with the draft?

Google's guidance says how content was created is a question readers may ask and describes explaining automation or AI use as a trust-building practice. Stating the usage in the contract and brief, and showing the human editor's role in the version log, is a sound practice. The documentation suggests no ready-made notice.

Who decides the byline on an article published under a client's name?

Google's helpful content guidance suggests readers should be able to tell who created content and reach author background, and it treats made-up names and false credentials as a low-quality signal. Decide in the approval chain who sets the byline and tie it to a real person.

Related guides: SaaS SEO: What to Do in the First 90 Days, SaaS Integration Pages SEO: Is Scaling to Hundreds Safe?.

Conclusion

Should revisions be capped in the contract? For most agencies the answer is yes, but the number is only part of it. Define the approval chain with names and collect feedback through one coordinator. Treat the brief as the measuring stick, number and date the drafts, and lock the final version. Separate rounds, revisions and out-of-scope requests from the start and decide what happens when the cap is reached. Before publishing, confirm that the live text matches the approved one, that sources are accurate and that sponsored links carry the right attribute, and keep everything in a simple log. No process guarantees rankings, and binding contract wording needs a legal adviser.

If you are looking for a service that works from a defined writer and editor workflow once your process is set, you can look at the content writing service on Akua Adforce; scope and terms are listed on that page.