Channels and categories inside Shopify admin, with direct messages for the things that are genuinely between two people, and roles deciding who can see each room.
A single channel is an excellent tool for three people. Past five or six it develops a specific failure: everyone applies one reading strategy to a stream carrying four different kinds of message, and that strategy becomes skimming.
Skimming does not distinguish between a rota question and a flagged order. The message that needed acting on gets scanned with the same attention as the one about lunch, which is why things start getting missed without anybody becoming less careful.
Rooms fix this by separating traffic that needs different responses. The split that survives contact with a busy week is by required response rather than by topic, because the category is obvious in the moment you are posting.
Rooms solve one problem and create a smaller one: past about eight of them, people are unsure where things belong.
Categories group related rooms so the sidebar stays readable. A fulfilment category holding picking, dispatch and returns reads as one thing rather than three, and collapsing it hides all three at once.
Categories can also be private, which is a different control from a private room. A private room is closed but visible: people know it exists and can ask to join. A private category is invisible to anyone outside it, which is the right tool for the narrow case where the existence of the conversation is itself the sensitive part.
Direct messages exist because some conversations genuinely are between two people. They are also the most common way a room structure quietly fails, because posting in a room feels like interrupting and a DM does not.
The cost is that decisions end up invisible and unsearchable, and the next person to handle the same situation starts from nothing.
A rule that holds: if the answer would be useful to anyone else, later, it belongs in a room. That covers nearly everything about orders, stock and customers, and leaves DMs for what they are actually for.
Search covers rooms and messages you have access to. Pins cover the things that should never need searching for: the returns policy exception, the supplier contact, the thing everyone asks in their first week.
The practical difference is that search rewards knowing what you are looking for, and pins help the person who does not yet know what exists.
Most teams overthink this and end up with fourteen rooms nobody can navigate. A starting set of four covers almost every store, and the split is by how fast something needs a response rather than by subject.
#daily-ops. The room that runs the day. Order flow, staffing questions, blockers. Everyone on shift is in it and reads it.
#fulfillment. Picking, dispatch, returns. High volume and largely irrelevant to anyone not packing, which is exactly why it earns its own room rather than flooding the first one.
#announcements. Things said once that everyone needs: a promotion starting, a policy change, a courier switching. Pinned, low traffic, read by everyone.
A private room for the awkward things. Pay, performance, a supplier dispute. Small membership, and having it means those conversations stay inside the workspace rather than migrating to somebody's personal phone.
Add a fifth only when an existing room is being muted by more than half the people in it. That is the real signal that a room is carrying two audiences.
The failure mode of any migration is a parallel run. Both tools stay open, everyone individually prefers wherever the others are, and after three weeks the conclusion is that the new tool did not catch on.
What works is a dated cut-over with a narrow scope. Pick a Monday, move operational conversation only, and leave the old tool readable rather than deleted. Nobody needs to migrate history: chat older than a month is rarely searched, and the handful of decisions worth keeping are better extracted into a document than bulk-imported.
Then name one person to redirect for two weeks. When something operational lands in the old place they move it and say briefly where it now lives. Every migration that sticks has that person; the ones that quietly revert do not.
The structure that suits six people is not the one that suits fifteen, and the change is worth making slightly before you need it, because the moment you need it is peak season.
Around eight or ten people, categories start doing real work. Grouping picking, dispatch and returns under one fulfilment category means the sidebar reads as three areas rather than nine rooms, and collapsing it hides all of them at once.
Around a dozen, the useful addition is usually a room per function rather than per task, with the general room deliberately kept quiet. At that size the general room is where new people post because they are unsure, so it works better as a fallback than as the main channel.
What does not scale is splitting by topic. Topic rooms multiply without limit and every new one makes the "where does this go?" question harder. Splitting by required response holds up because the answer is obvious while you are typing.
Worth stating plainly, because a feature page that only lists strengths is not much use for deciding.
This is internal chat. It does not talk to your customers, and if you need a storefront widget you want a live chat app instead. It is also scoped to store operations, so if you need one tool for a whole company including people who never open Shopify, an embedded admin app is the wrong shape.
And it will not fix an ownership problem. If nobody knows who handles returns, a better room structure makes that ambiguity more visible without resolving it.
Neither belongs in a starting set, and both earn their place within a month or two.
A room for the week ahead. One post on Monday listing what is happening: a drop, a courier collection moving, somebody away. It replaces the same information being asked five times in five places, and because it is one post a week it stays readable.
A room for things that went wrong. Not blame, just a note when something breaks and what fixed it. After a couple of months that room is the closest thing a small store has to documentation, and it is written by the people who actually did the work rather than by whoever was asked to write a process document.
Usually three or four to start, split by how quickly something needs a response rather than by topic. Topic-based rooms multiply until people are unsure where to post.
Yes. Roles decide which rooms each person can see, and private categories stay invisible to anyone outside them. Both are set per person rather than per invitation.
Yes. Store Huddle runs inside your Shopify admin, so everyone who uses it needs a staff account on the store. The Store Huddle role they are given then controls which rooms and categories they can see in the workspace.
The conversation stays. Removing a person removes their access, not the record, so the room history remains searchable for whoever picks up their work.
The free plan keeps 30 days. Paid plans keep unlimited history, which matters more than it sounds, because the archive is what you need during a dispute or a post-mortem.
Free for teams up to five, and about two minutes to connect your store.