Shopify content approval workflow for a 2 or 3 person team

Why small teams need a workflow at all

With two or three people, the usual argument against an approval process is that everyone talks to each other anyway. The trouble is that product content is rarely changed in a conversation. It is changed in the admin, often late in the day, often in bulk, and more and more often from AI drafts. A second look is the cheapest protection you have against a wrong size, an unsupported claim or a broken translation reaching customers.

The goal is a workflow light enough that nobody wants to route around it. That means clear roles, a short list of what needs approval, and an agreed way to handle the exceptions.

Who proposes, who approves

Start by writing down roles, even if one person holds two of them. A typical split for a small store:

Person	Proposes	Approves
Founder or owner	Rarely	Yes, final say on claims and pricing text
Content or marketing lead	Yes, most copy and SEO	Yes, for the other person’s work
Freelancer, assistant or AI agent	Yes	No

The one rule worth keeping in every configuration: nobody approves their own substantive change when someone else is available. In a team of two, that means each person reviews the other. In a team of three, the owner does not need to see everything; they can approve only what touches claims, prices mentioned in text or legal wording.

What needs approval and what doesn’t

A workflow that asks for approval on every comma dies within a month. Draw the line by risk, not by field.

Needs approval:

• New products and new descriptions
• Any change to materials, dimensions, compatibility or care instructions
• Health, safety, certification or legal wording
• Titles, handles, SEO titles and meta descriptions
• Translations, in every language
• Anything drafted by AI or imported in bulk from a spreadsheet

Can skip approval:

• Typo and punctuation fixes that don’t change meaning
• Restoring a previous version after a mistake
• Internal-only fields that don’t appear on the storefront

Write this list down and share it. When someone is unsure, the default is to propose.

Turnaround and actionable change requests

A review queue only works if proposals don’t sit in it. There is no correct number here, so set an expectation that suits your pace. A reasonable starting point for a small team is a decision within one working day for normal changes, and a fixed moment each day (for example after the morning orders) when the reviewer clears the queue.

The other half is the quality of the feedback. “Not quite right” sends the author back to guess. A useful change request names three things:

• The field: “Description, second paragraph” or “SEO meta description, German”.
• The problem: “States the jacket is waterproof; the supplier sheet says water-resistant.”
• The expected fix: “Use ‘water-resistant’ and add the rating from the spec sheet.”

If a fix is a one-word correction, it is often faster for the reviewer to make it than to send it back. Agree that reviewers may correct small issues themselves, and keep “request changes” for problems the author needs to rethink.

Urgent changes and the Shopify permissions gap

Sooner or later something wrong is live and needs fixing now. Decide in advance who may make an urgent change directly (usually the owner and the content lead), and that every urgent change is looked at by the other person the same day. The fix goes first; the review still happens.

Then there is the gap in Shopify itself. Shopify’s store permissions have a products permission called “Create and edit”, and according to Shopify’s documentation it lets a user create, edit, import, publish, share and archive products and collections. Blog posts and pages work the same way: one permission covers creating, changing, publishing and deleting them. There is no separate “submit for review” permission. If a product is already active, saving an edit in the admin changes the storefront immediately.

So the admin can’t enforce the workflow you just designed. You have two levers:

• Narrow Shopify permissions. Give edit access to products and content only to the people who approve, and let others propose through a process or tool that doesn’t need it.
• Hold changes outside Shopify until approved. Proposals live somewhere else, and only an approval writes them to the store.

How Vetta implements it

Vetta is the second lever, built for this case. Versions of products, collections, pages, articles and their translations are proposed by a person, a CSV import or an AI agent, and reach the storefront only once a reviewer approves them.

The roles map to the table above. Authors propose and follow their proposals; they can’t approve. Reviewers approve, request changes with a note, or reject. Admins do the same and also manage the workspace, members and imports. Users are unlimited, so the freelancer and the founder don’t cost a seat. AI agents connected over MCP are authors only and have no way to publish.

When proposing, the author can add a note and pick the reviewer, who is notified. The reviewer sees a field-by-field diff of live versus proposed content, with one tab per translation, and can correct a small issue and save it as a new version instead of bouncing it back. A daily digest email lists what is waiting and what was decided.

For urgent changes, a reviewer or admin can propose and approve in one sitting. And if someone edits directly in the Shopify admin, Vetta records the edit as a dated version marked “Edited in Shopify” in the version history, so the direct path stays visible and reversible.

Setup checklist

1. Write down who proposes and who approves
2. List what needs approval and what doesn’t
3. Agree on a daily review moment and turnaround
4. Define who may make urgent changes
5. Review Shopify staff permissions for products and content
6. Remove edit access from people who only propose
7. Install your approval tool and invite everyone
8. Choose which products and pages are tracked
9. Run one week, then adjust the list