Features
Part of Getting content approvals right the first time
7 details of content approvals strategy people miss
Content approvals strategy: how to design a review system with named decision rights, service levels on both sides, and an escalation path that gets used.
An approval process is a queue with opinions attached. Most of them are designed by accretion: someone was surprised once, so a reviewer was added, and nobody ever removed one. The result is a system where six people can say no and nobody can say yes.
This page is about designing the system rather than running it. What happens inside a single review is covered in content approvals.
What to take away
- Decide who has a decision right and who has an opinion, and write it down. Everything else follows from that one distinction.
- A queue with no service level on the brand's side is not a process, it is a hope.
- Design the system before you have a specific contentious asset, because after that every rule reads as being about that asset.
Decision rights, not attendance
The first question is not who reviews. It is who decides.
There are only three positions worth having: someone who can approve, someone who must be consulted and whose objection blocks, and someone who is informed. Most review panels are full of a fourth kind of person, who is consulted, cannot block, and is nonetheless treated as though they can, because nobody wants to overrule them.
Write the list down, per asset type, before the campaign. Something like a responsibility assignment matrix is the usual formalism, and the formalism matters less than the fact that the answer exists in writing.
Two rules that do most of the work:
- One approver. Not a committee. A committee produces the average of everyone's caution, which is the thing readers can feel in finished work. The failure mode has a name, design by committee, and it is expensive twice over: once in the work and once in the time.
- Blocking objections need a stated reason type. Legal risk, factual inaccuracy, disclosure, or brand standard. "I do not like it" from a non-approver is an opinion, and treating it as a block is how three rounds become seven.
Two lanes, not one
Legal and regulatory review is a different activity from taste review, and running them together is the single most common design fault.
Legal review asks whether a claim can be substantiated, whether the disclosure will be noticed, whether third-party material is cleared, and whether anything said is a promise the company cannot keep. It has correct answers and it should run early, on the concept and the script, where a change is cheap.
Taste review asks whether this is the kind of thing the brand wants to be. It has no correct answers, it belongs to one person, and it should also run early, for the same reason.
Splitting notes lets you tell a creator: this note is a requirement, this note is a preference. A creator who cannot tell the difference treats all notes as requirements and stops arguing, which sounds convenient but produces worse work.
A written creator outreach strategy keeps requirement notes and preference notes from arriving as one undifferentiated list.
Where the substantiation question actually lands is worth reading in the FTC's advertising FAQ for small business, which is plainer than most internal guidance on what a claim needs behind it.
Service levels, in both directions
Contracts routinely bind the creator to a turnaround and leave the brand's response time unstated. That asymmetry is the source of most approval friction, and it is entirely fixable.
| Stage | What to commit to | What it prevents |
|---|---|---|
| Concept or script | A response within a stated number of business days | Production starting on an unreviewed idea |
| First cut | One consolidated set of notes by a stated date | Notes arriving in three waves from three people |
| Final | Approval or a specific blocking reason | Silence that pushes a publication date |
| Deemed approval | Automatic approval after a stated period of silence | An open-ended obligation on one side only |
Consolidation is the part people skip. Notes should reach the creator once per round, reconciled, with contradictions resolved internally first. Sending a creator two opposing instructions from two stakeholders transfers your disagreement to someone who cannot settle it.
The queueing arithmetic is not complicated and it is worth internalizing: total time depends on how long each stage waits, not on how fast each reviewer is. Little's law is the compact statement of it. Adding a reviewer adds a wait even when that reviewer is fast.
Escalation
Every system needs a route for the case it did not anticipate, and the route has to be short.
Name one person who can decide when the panel disagrees, and say so in advance. Give them a deadline. Record what they decided and why, because the next contentious case will be similar and the reasoning is reusable. The same reasoning applies to a creator outreach mistake log, where each recorded reply becomes reusable guidance.
Escalation that requires a meeting will not be used. People will instead work around the process, which is worse than a slow process because it is invisible.
Designing for what actually goes wrong
Three failures are worth designing against specifically.
Late-stage taste objections. Someone senior sees the work for the first time at the final cut. The fix is not a faster process; it is putting that person into the concept stage or accepting that they do not get a vote.
Disclosure caught after publication. Approvals happen on a file; disclosure happens in a published context. Build the check on the live post into the process, with a named checker and a date, rather than treating it as something the contract handles. The detail is in contracts and disclosure.
Scope growing through notes. A note that asks for an extra cut is a purchase, not a revision. Say so at the time. The general pattern is ordinary scope creep and it is the main reason approval processes get expensive without anyone deciding they should.
What the system inherits
An approval process cannot repair an unclear brief, and most review arguments were undecided at the start, and they become visible only when there is something to look at.
Push decisions upstream into campaign briefs, and be honest about which were never made. What the review protects is set out in strategy and objectives.
Bottom line
Separate decision rights from opinions and write the list down. Run legal and taste review as two lanes, both early. Commit to a response deadline on your own side and consolidate notes into one voice per round. Name one escalation owner with a deadline. Then add the one check the process usually lacks: someone looking at the published post.
Common questions
How many reviewers is too many?
More than one approver is too many. The number of consulted parties is limited by the deadline you are prepared to commit to, which is a more useful constraint than a headcount.
What if legal will not commit to a response time?
Ask what the constraint is. It is often volume rather than unwillingness, in which case the fix is triage: which asset types need review at all.
Should creators see the internal disagreement?
No. They should see one consolidated set of notes. Unresolved internal argument in a comment thread tells a creator to wait rather than work.
Is deemed approval risky?
It transfers a risk that currently sits entirely on the creator. Whether it suits you is a question for your lawyer, but the alternative is an obligation with no end date, which is its own risk.
How do you shrink a process that has already grown?
Ask each reviewer what they have blocked in the last year and why. The list is usually shorter than anyone expects, and it tells you who is deciding and who is attending.
