A scenario that plays out constantly on every freelance platform
A client and a freelancer are a few messages into a job. One of them suggests moving to WhatsApp or email — it feels faster, more personal, less friction. Everyone agrees. A few weeks later, the client says the work isn't what they asked for. Or the freelancer says a scope change was agreed to verbally that the client now denies. Neither side has anything to point to. The conversation that would have settled the disagreement in thirty seconds doesn't exist anywhere either party can produce it.
This isn't a hypothetical — it's the single most common story in freelance dispute forums, on both sides. It's also the most avoidable one.
Why this matters for Smiths
If a client stops responding after you deliver, or disputes work you believe you completed as agreed, the job thread is your evidence. Messages, submissions, and timestamps on Smith show exactly what was asked for, what was delivered, and when. A verbal agreement or a text exchanged off-platform doesn't leave anything either side can point to — which means it can't help you if things go sideways, no matter how clearly you remember it.
There's a sharper version of this too: requests to move off-platform, especially combined with unusual payment requests (crypto, "verification fees," wire transfers), are one of the most consistent signals of an actual scam — not a rule technicality, a real pattern. A legitimate client has no reason to need you off Smith to pay you.
Why this matters for Clients
The same logic runs the other way. If a Smith stops delivering, or delivers something different from what was discussed, your job thread is what shows what was actually agreed — the brief, the scope, any changes along the way. An agreement made over a call or a personal message leaves you with nothing to reference if the final work doesn't match what you thought you'd asked for.
It also means you lose the parts of Smith built specifically to protect you as a client — payment held until you approve the work, a structured review window, and a dispute process that has something to actually look at.
The mechanic that makes this concrete
When something needs to be reviewed — whether that's confirming what was delivered, or resolving a disagreement — what gets looked at is the job thread: the messages, the submitted deliverables, the timestamps. That's not a policy preference, it's just how a fair review works. It can only be based on records both sides can see, not on a private conversation only one side remembers accurately.
Every scope change, every "actually, can you also add X," every "yes, that looks good" — if it happened in the job thread, it's part of the record. If it happened somewhere else, it isn't, for either of you.
This isn't about restriction — it's about what actually protects you
Reframe it this way: staying on-platform isn't a rule you're following for Smith's benefit. It's the thing that makes your own position defensible if a disagreement happens. Every message you send in the job thread is evidence working in your favor, whichever side of a disagreement you end up on.
What's actually fine
This isn't about never talking outside the platform. A quick video call to sync on something complex, or a screen-share to walk through a bug together, is fine and often faster than typing it all out. The distinction is simple: anything that needs to be remembered later — scope, price, deadlines, what "done" means, whether something was approved — belongs in the job thread. If you have a call, drop a one-line summary back into the thread afterward. It takes ten seconds and it's the difference between a disagreement that resolves in a minute and one that doesn't resolve at all.
The one-line version
If it matters later, put it in writing, and put it here — not because you have to, but because it's the only version of "what we agreed" either of you will be able to prove.