A social media contract prevents the month three argument when it puts numbers where the verbal agreement put adjectives: how many pieces, in what format, how many revision rounds, what the approval deadline is and what happens when nobody replies. The rest, account ownership and a clean way out, is written once and never discussed again.
Why the argument lands in month three
It is neither coincidence nor bad luck. Month one is carried by goodwill on both sides: there is enthusiasm, extra work gets done and nobody is counting. Month two brings routine and the edges start to show, but there is still slack. In month three three things happen at once: the client has formed an expectation out of the extra work you did, you have spent the buffer you started with, and the first results arrive without being spectacular.
That crossing point turns every empty box in the agreement into a negotiation, and month three negotiations are won or lost on what is written down, not on who is right. The templates in circulation do not help, because almost all of them describe the service with adjectives: full management, quality content, reasonable revisions. None of those three can be checked, and what cannot be checked gets argued about.
A useful contract is short and boring. It says how many units of what, by when, with how many rounds, and what happens when one of those does not occur. Everything else, the confidentiality clause and the rest of the legal scaffolding, matters less than it looks: hardly any relationship breaks there.
Deliverables you can count, not describe
The test is simple. If somebody outside the relationship can read the contract and say, without asking anything, whether the month was delivered, it is well written. If it needs interpretation, it is not.
| What usually gets written | What ends the argument |
|---|---|
| Social media management | Twelve pieces a month, split across two named networks |
| Quality content | Eight carousels and four vertical videos of up to 45 seconds |
| Community management | Replies to messages and comments on working days, within 24 hours |
| Reporting | One monthly report with the same six metrics, on the first working day |
| Reasonable revisions | One round of changes per piece, two on the flagship piece of the month |
| Photo and video shoots | One four hour shoot per quarter, travel billed separately |
The right hand column has an advantage nobody advertises: it protects you as well. A contract with numbers in it is the only way to say no to an extra request without sounding unwilling, because the conversation stops being about attitude and becomes about units.
It is worth counting pieces by format and not only by quantity, because the cost is not the same. A ten image carousel and a single photo take up the same line in a proposal and very different afternoons to produce. How all of this translates into a monthly figure is in the social media pricing guide.
Revisions, rush jobs and what sits outside the retainer
Revisions are the chapter that costs the most money and the one fewest people write down. Three things need fixing: how many rounds are included per piece, what counts as a round, and what happens beyond that. A round is a block of comments, not a comment; if the client sends three emails over three days, that is either one round or three, and that sentence belongs on the paper.
A change of mind deserves its own line too. Redoing a piece because the brief turned out to be something else is new work rather than a revision, and the difference between those two decides whether your month adds up. The same goes for rush jobs: publishing something off calendar on the same day can be accepted, but then say how many times a month that sits inside the retainer and what gets billed from the next one onwards.
The third block is what is excluded, and listing it beats leaving it implied. Ad spend and its management, external production, music or stock licences, brand design, web development. None of that is included by virtue of sitting near the content, and finding out in month three is precisely the argument this document exists to prevent.
The approval clock
A content calendar that depends on the client replying is not a calendar, it is a wish list. So the contract sets a review deadline, counted in working days, starting the day the piece becomes available to look at rather than the day somebody opens it.
The important half of the clause is the second one: what happens when that deadline passes. There are two reasonable outcomes and no third. Either the piece goes out as it stood, which protects the frequency you committed to, or it moves to the following month and consumes that month's allowance, which protects the relationship with very careful clients. Choosing one of the two in writing saves half a year of emails.
The path to the yes should be genuinely short. In an agency workspace each client carries its own approval switch: with it on, the piece waits for one yes, your team's from the app or the client's from their portal, whichever arrives first, and one is enough. With it off, which is the default, the piece is scheduled straight away when it is saved. The client can reject with a reason, every decision stays in the post's history, and a piece whose hour passed while it was waiting is parked for rescheduling instead of going out late. Which access credentials to ask for so that circuit works from day one is in the client onboarding guide.
Everything in this article happens in one place in GoFeed.
Try it freeWhat gets recorded, and what is not a signature
This deserves its own paragraph because it is constantly misread. A quote is sent as a link, the link records when it was first opened and when it was last opened, and somebody in your workspace with the right permission marks it accepted. The client does not accept anything through the portal, and nothing here is signed: there is no electronic signature in this product and the portal's own quotes screen does not exist yet.
That is not a problem as long as the paperwork accounts for it. An agreement accepted in writing by both parties holds, and what needs care is traceability: which version of the document was accepted, who said so and when. Keeping the email thread next to the accepted document is worth more than any stamp.
Everything administrative hangs off the contract afterwards, and that is worth setting up properly from the start: your own numbering series, credit notes in a series of their own as Spanish law requires, the VAT regime that applies to each operation, and the IRPF withholding (the income tax retention a Spanish client applies) where it is due. How that chain works, from brief to invoice, is on quotes and invoices and in more detail on the billing feature page.
Who owns the accounts and the material
This is the clause that costs the most when it is missing, and it never surfaces until the day you part company. The healthy rule is that the accounts belong to the client, always, and that your team works on them with permissions granted from the client's own business manager. Creating a profile in your name on day one for convenience is a time bomb: when the relationship ends, either somebody hands over something that is not theirs to give, or somebody holds on to something that was never theirs.
The material is a separate conversation. Produced pieces are usually assigned to the client once paid for, and it is worth saying with what scope: use on social, use in paid media, use on other channels. Where licensed photography or music is involved, the licence governs, and the contract should say whose name it is in and how long it lasts.
Then there is a third group almost nobody writes anything about: the data. Historical metrics, previous reports, audiences you built, reply templates. Defining what of that gets handed over at the end, and in what format, avoids having to negotiate it exactly when neither side is in the mood.
How it ends
Every contract ends, and the one that says so in writing ends better. Three lines will do. A notice period, which on a monthly retainer is usually thirty days and on a committed term matches whatever is left of it. What gets delivered in the final month, which is not the same as a normal month. And how access is handed back, with an actual date.
A minimum term is a business decision, not a moral clause. If you set one, it has to be justified by real up front investment, and it is better kept short: a tied client who wants to leave does more damage in reviews and in conversations than they would have done by leaving.
Last, and this is what separates a clean exit from a bad one: put in writing that the final month is paid in full even if the relationship ends on the 12th. Not out of harshness, but because the alternative is arguing about it, and arguing with somebody who is already leaving is time you never get back.
