Platform appeals
A structured, evidence-based response to a suspension or rejection.
Shopify AUP suspensions, Merchant Center disapprovals, Meta rejections and app review refusals: what the notice actually points at, what to change, and an appeal built on evidence. The decision belongs to the platform, so no outcome is promised.
Fixed price
From $850
That covers one notice: the diagnosis, the fix and a single submission. A suspension with several causes, or an app review, runs to $3,500. No outcome is promised and the fee does not depend on the decision.
Send the brief and get a number back.
Describe the projectThe notice names a rule, not the thing that tripped it
A suspension email quotes a policy and a category. It does not say which page, which product, which screen or which request failed. The account, the store or the app is off in the meantime, and the reply box is right there, which is how most first appeals get sent: fast, written about intent, arguing that you are a real business selling real things. Platforms do not weigh intent. They re-check, and a re-check needs something to look at. An appeal that gives a reviewer nothing to verify gets closed in the time it takes to read it.
What the work covers
Reading what the notice actually points at
The useful detail is rarely in the letter. It is in the item-level issue list behind an account-level banner, in the exact URLs a reviewer hit, and in the screencast attached to an app review rejection, which shows the failing screen frame by frame with the console open. That material gets read before anything is written.
Fixing the cause, then making it checkable
The fix comes first and the evidence is built for a reviewer with a minute of attention: the URL that now answers correctly, the policy page that now exists at a findable address, a test account with credentials, a screenshot with a timestamp. If verifying the claim takes effort, the claim does not count.
One submission, in the platform's own vocabulary
The appeal cites the clause they cited, states what changed, and says exactly where to look. It does not explain your business model, dispute the decision or ask for understanding. Attempts and cooldowns are limited on every platform, so the first one is written as though it is the only one.
Why it is done this way
Appeals are decided on verifiability, not on argument
Reviewers work through queues against a checklist, usually without the context you have. Anything they cannot confirm in the interface in front of them does not exist for the decision. This is why appeals fail for procedural reasons far more often than for the underlying problem: the cause was fixed but not shown, or shown somewhere the reviewer never looked. Rewriting the same case as a checkable claim is most of the work.
We sit on the other side of review ourselves
Our own products go through the same gates. Stellaforge is published in the Shopify App Store, its taxonomy appeal was argued and approved, and access to Google's Merchant API was granted eighteen days after submission, held up by a detail in a privacy policy rather than by anything about the software. App review rejections here get read from the reviewer's screencast rather than the summary email, which is how one of them turned out to be failing requests on the platform's own admin domain and not on the app at all. That habit is the service.
How it runs
You forward the notice with everything attached to it, plus read access where the platform allows it. Nothing is drafted before the underlying issue list is read.
A written diagnosis: what the platform is most likely reacting to, what is fixable, what is not, and whether an appeal is worth filing at all.
The fixes are made and then verified from outside, the way a reviewer would see them rather than the way the admin shows them.
The appeal is written and sent once, with the evidence attached and the clause addressed directly.
If it is refused, the refusal is read for what it adds, and the decision on whether a second attempt is worth it comes with reasons rather than optimism.
Background
How platform appeals actually get decided
Most appeals fail for procedural reasons
The common failure is not that the merchant was in the wrong. It is that the appeal was sent before the fix, sent through a channel that does not open a review, or written as an explanation of intent rather than a claim someone can check. Reviewers are working a queue against a checklist and have less context than you do. Anything they cannot confirm from the interface in front of them does not enter the decision. So the appeal that works reads like a bug report: here is what was wrong, here is what changed, here is the address where you can see it.
The evidence a reviewer can actually use
Useful evidence is short, addressable and dated. A URL that now returns the right thing. A policy page that exists at a findable path rather than inside a modal. A test account with working credentials, where a login gate blocked the review. A screenshot with a visible timestamp. Useless evidence is a paragraph of context, a promise about future behaviour, or a claim about how long you have been trading. Nothing about your history is checkable in the two minutes the case gets.
The same shape, different rules per platform
Google Merchant Center separates item-level disapprovals from account-level suspensions, and only the second kind needs a request at all. Shopify enforces its acceptable use terms against the whole store, often on the product mix or on claims made on the page. Meta reviews commerce assets and tends to react to the catalogue and the domain together. App review, on either the Shopify or the Google side, attaches a screencast, which is the single most useful artifact any platform hands you and the one most often left unwatched. The craft transfers between them, the specific channel and vocabulary do not.
What is not promised here
No outcome. Reinstatement, approval and re-listing are decisions made by the platform against rules it can change, and any offer that guarantees one is selling certainty nobody holds. What you get regardless is a written diagnosis of what the enforcement is reacting to, the fixes carried out, and a case file you can hand to anyone else, including your own team, if you decide to carry it further alone.
Describe the project
Tell us what you have and what should change. Within two working days you get a written calculation: the scope broken into parts with a price against each, or the questions needed to write one. No discovery call in between.
Questions
Appeals, asked and answered
No, and neither can anyone else. The decision belongs to the platform. What is controllable is that the cause is correctly identified, actually fixed and presented so a reviewer can confirm it in under a minute. If the honest read is that the case is unwinnable, you get that in writing instead of a bill for attempts.
That is common and not fatal on most platforms, though it narrows the room. The first thing is to read the refusal for what it added, because a second notice often names something the first one did not. Then the question is whether anything material has changed since. Sending the same case again with different wording is the fastest way to close the remaining attempts.
The form, in almost every case. Replies to a no-reply notice land nowhere, and support chat rarely reaches the team that made the decision. Each platform has one channel that actually opens a review, and using the wrong one costs days while you believe the case is in progress.
Different platform, same shape. Shopify's acceptable use terms are enforced against the store as a whole, often triggered by the product mix, by claims on the page, or by a payment risk signal rather than by anything you did that day. The work is the same: find what the enforcement is reacting to, fix it visibly, and answer in the channel that reopens a review.
The preparation is days rather than weeks, because most of it is fixing and verifying rather than writing. The decision is the platform's timetable and it varies wildly: some come back within days, and our own Merchant API access took eighteen. Anyone quoting you a decision date is quoting a date they were not given.