How to prove a client approved a post (and what actually counts as proof)

By Grig K. · Updated June 2026 · 7 min read

Most managers think they have proof because they have a "yes." The yes is the easy part. The part that bites comes weeks later, when a post underperforms or a new stakeholder objects, and you're asked to show that the client agreed to this exact post, on a date, in a form the other side can't wave away. I watched a manager lose that moment with an approval she actually had: the client had okayed the campaign, but the only record was a chat line that didn't name the specific image that ran. "I approved the idea, not that picture," the client said — and there was nothing to put next to that sentence. The yes existed. The proof didn't.

What actually counts as proof a client approved?

Proof is a record that survives the scrutiny of a person who wasn't in the room. And the moment that record gets demanded is common, not exotic: in Adobe's 2024 State of Creativity report, 34% of stakeholders said they face long waits getting work reviewed and approved, and 22% said they struggle to keep versions straight at exactly that stage — the two places where a "yes" gets lost. Run any approval through four tests, in this order — the last one is the one almost everything fails.

1. Dated — there's a timestamp you didn't write yourself.
2. Tied to the exact post — it points to the specific caption and image that ran, not "the campaign" or "the plan."
3. Shows who approved — it's attributable to the client, not your note that they said yes.
4. Independently checkable — someone who wasn't there can confirm it without trusting you.

How do common approvals score on the four tests?

Almost everything solo managers rely on clears one or two tests and quietly fails the rest. Only the bottom row passes all four.

The chat rows fail for reasons you can read in the platforms' own documentation. WhatsApp lets a sender edit a sent message for 15 minutes (since 2023) and delete it for everyone for two full days; iMessage allows up to five edits within 15 minutes and unsending within two; Telegram lets either side delete any message in a one-on-one chat, at any time, leaving no mark; in Slack, editing and deleting are workspace settings, and deletion is permanent. A "yes" that either party can quietly rewrite or erase isn't a record. It's a draft.

The email row is where most careful managers stop, and it's why disputes still happen: an email is dated and attributable, but it isn't tied to the exact post, and an edited reply pasted into a long thread is easy to argue about. The screenshot is worse — it fails on its face. Fact-checkers at Poynter called fake screenshots "a double deception" back in 2020: they falsify both the information and its source, and people lean on them precisely because the original messages get edited or deleted. The first thing a determined client says about your screenshot is "that's doctored" — and you can't prove otherwise from the screenshot itself.

How to capture an approval that actually counts

You don't fix this in the dispute. You fix it at the yes. Four steps:

1. One approval per post, in writing, before it publishes — never a verbal okay or a reaction in a thread.
2. Tie it to the exact version: the record has to point to the specific caption and image the client saw, so "I approved the idea, not that picture" has nothing to stand on.
3. Capture who and when automatically, so the record is the client's yes — not your note about it.
4. Keep it somewhere a third party can open and check without your login. A record only you can see is an audit trail, not proof.

None of this adds a stage to your workflow, because review is already where the time goes: in a 2018 inMotionNow/InSource survey of 400+ in-house creatives, 48% said work takes two to three days to get reviewed and approved, and for 29% it takes a week or longer. The days are already being spent. In my experience building an approval tool, the hard part was never capturing the yes — it's capturing it somewhere the other side can open.

This is the gap a no-login approve link is built to close: you send one link, the client taps approve with no account, and a dated record of the exact post — and a page anyone can open to confirm it — exists the moment they say yes. See how it works.

Why "the tool already logs it" isn't the same as proof

Most approval tools keep a record — but it lives behind your login, as an internal audit trail. That clears three of the four tests and fails the one that decides a dispute: the person you need to convince can't open it. A new account lead, a brand's finance team, or a client who's now claiming they never agreed can't independently confirm a line in your dashboard — they'd have to trust the party they're arguing with. Proof is the version they can check themselves, on a page that doesn't require your account or your goodwill. That's the difference between a yes you logged and a yes you can prove.

The stakes of losing that difference are measured, not theoretical. In the Freelancers Union's Costs of Nonpayment survey (2015, 5,000+ US freelancers), 71% had struggled to collect payment they were owed at least once, and the average unpaid freelancer lost about $6,000 a year — 13% of income. A lot of those fights have this exact shape: work delivered, agreement remembered differently, and no record both sides accept.

Grig K.

Writes about client approvals, sign-off and disputes in social-media work — the everyday spots where a 'yes' goes missing and someone gets blamed. More from Grig K. →