Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Social Media Marketing › Social Media Management

Setting Up a Social Media Approval Workflow

How to set up approvals for social posts by risk: who signs off what, a fast lane for replies and reactive posts, scheduled queues, records and review.
Tutorialintermediate12 min readUpdated

What an approval workflow is for on social

Approval in marketing usually means a campaign asset passing through brand, legal and accessibility checks before it is released. The approvals and compliance reviews guide sets out an approval matrix by risk for all marketing, and the creative approval SOP is the procedure for a finished creative asset. Both apply to social when a campaign asset is posted. But most of what a social account publishes is not a campaign asset. It is a weekday post written that morning, a reply to a comment, a reaction to the news, a customer's photo reshared, a creator's post amplified, a price announced, a competition run, an apology. Each carries a different risk and needs a different speed.

A workflow that treats all of those the same fails in one of two ways. If everything needs the same sign-off, replies arrive late, reactive posts miss the moment and the team learns to post without asking. If nothing does, the post that quotes the wrong price or makes a claim nobody can support goes out because nobody was asked. The social media strategy hub puts the balance plainly: approval chains should be short, and a voice guide people trust removes the need to approve every post.

The workflow this tutorial builds therefore sorts what goes out by risk and gives each level its own route: no sign-off for the routine, a fast lane for the urgent, and the right check for the risky. The fast lane is the level that needs written rules for what may enter it and what must leave it. It sits inside the routines of the social media management section and links to the general matrix rather than repeating it.

Before you start

  • A content calendar. The social content calendar tutorial builds the calendar with owners, statuses and an approval column; this tutorial sets what goes in that column.
  • A voice guide. The brand voice and cadence guide is what makes it safe to publish routine posts and replies without sign-off.
  • The organisation's approval matrix. If one exists, from the approvals and compliance reviews guide, social approvals should map to it rather than invent new reviewers.
  • A list of who is involved. The social team, the people who approve claims, prices and partnerships, whoever speaks for the organisation under pressure, and any agency that posts on the brand's behalf.
  • A RACI. A RACI matrix names, for each level, who is responsible for the post, who is accountable for approving it, who is consulted and who is informed, with one accountable person per level.
  • A few recent examples. Posts that waited too long, posts that should have been checked and were not, and replies that went wrong. They test the levels once they are drafted.

The steps

  1. List everything the accounts publish

    Start from what the accounts actually put out, not from an org chart. Go through a recent month on every account and list each kind of thing that went out: planned posts by pillar, replies to comments and messages, reactive posts, reshared customer posts, creator and partner posts, campaign assets, posts that mention prices, offers or competitions, posts that make claims about the product, paid boosts of organic posts, and statements made when something went wrong. Add what the accounts would publish if something serious happened. The list is usually longer than the team expects, and it is the scope of the workflow: anything not on it will be decided under pressure on the day it first appears.

  2. Sort it by risk

    For each kind of content, ask what could go wrong if it went out with a mistake and how easily the mistake could be undone. The one-way and two-way doors framework is the test. A routine post or a reply that can be corrected or deleted quietly is a two-way door: a mistake costs little and can be reversed. A claim, a price, a partnership, a competition or a statement on the organisation's behalf is closer to a one-way door: once it is seen, screenshotted and quoted, deleting it does not undo it. Sort the list into a small number of levels, from routine to sensitive, and write next to each the reason it sits there. The approvals and compliance guide's matrix does the same for marketing generally, triggered by something observable in the work, and the social levels should use the same triggers where they overlap.

  3. Name an approver for each level

    Give each level one accountable approver, by role, in the RACI. The routine level usually has none beyond the person posting, working to the voice guide. Posts that touch claims, prices or competitions go to whoever owns that risk in the organisation's matrix. Partnership and customer-content posts go through their checks, covered below. Statements under pressure go to the person who speaks for the organisation. Name a deputy for each approver, so that one person's absence does not stop the queue, and keep the number of approvers per post to the one the risk requires. A post that needs everyone's agreement waits for the slowest reviewer and is shaped by the most cautious.

  4. Write the fast lane for replies and reactive posts

    Replies and reactive posts are where a slow workflow does most damage, so give them their own lane. For replies, write a bank of pre-approved lines for the questions and complaints that recur, built with the approvers of the risky topics, so the person replying can answer a question about prices, terms or service without waiting; the community management guide sets the response standards and escalation path these replies follow. For reactive posts, write down what counts as one and who can approve it on the day, as the content calendar tutorial recommends, because a reactive post that waits days for sign-off is no longer reactive. Anything in the fast lane that touches a claim, a price or a sensitive subject leaves the lane and goes through its normal check.

  5. Set approval states in the calendar or scheduler

    Write the approval route into the tools the team already uses rather than into a separate process. The content calendar tutorial uses a short set of statuses, such as draft, in review, approved, scheduled and published, and the management hub asks whether a scheduler can hold posts in approval states so that nothing unapproved can go out. Use those states, add a column for the risk level and the approver, and make the rule simple: nothing moves to scheduled until it is approved at its level. Campaign posts that arrive with an approved Creative Brief and assets through the campaign workflow carry that approval with them; the social team still checks the caption and the timing.

  6. Route claims, partnerships and customer content to their checks

    Three kinds of content have checks of their own, and the workflow sends them there rather than inventing new ones. Claims, prices and competitions go to the compliance check: the approvals and compliance guide notes that claims must be capable of substantiation before they are made and that in the UK the ASA, listed in this page's sources, publishes the codes and rulings that apply, and the social post publishing checklist asks for claims and competition terms to be checked against advertising rules before a post goes out. Partnership and creator posts go through the disclosure check in the influencer contracts and disclosure guide. Reshared customer content goes through the permission check the user-generated content guide describes. Accessibility rides along with every level rather than as a separate gate: the accessible social media posts guide sets the standard for descriptions and captions, and the approver checks them as part of the post. This tutorial does not interpret any of those rules; it routes the post to the person who applies them.

  7. Set rules for out-of-hours and the queue

    Scheduled posts go out whether or not anyone is watching. Decide who may pause the queue out of hours and how they are reached, and write down when they should: news that makes a scheduled post look tone-deaf, a complaint wave, an incident affecting customers. Decide too what can be posted out of hours and by whom, which is usually only fast-lane replies and holding lines, with anything else waiting for the next working day or for the named person. When the organisation faces a serious problem, the workflow switches to the route set for statements under pressure, with one approver and one voice; the routine queue pauses until that approver says it can resume. The social media crisis plan tutorial is where that route, its holding lines and its approver are written.

  8. Record each approval

    An approval nobody can find later is only slightly better than none. The approvals and compliance guide describes the audit trail as a record attached to the work item of what was reviewed, which version, by whom, when and what they decided, with any evidence relied on. For social, the natural home is the calendar row or the scheduler's approval history, next to the post itself. Approvals given in a chat thread or a hallway are copied onto the item. Fast-lane replies need no individual record, because the bank of pre-approved lines is the approval; record instead when the bank was approved and by whom, and update it when it changes. Those records are what the tutorial on auditing marketing approvals samples when it checks that live posts had the reviews their risk level required.

  9. Review what was held and why

    Look back at the workflow on a regular cycle. List the posts that were held, returned or escalated and the reason each time, the posts that went out and caused a problem, and any approvals that took so long the post lost its point. Held posts with good reasons show the workflow working. Posts held for no clear reason show a level set too high, and problems that slipped through show a level set too low or a check nobody ran. Adjust the levels, the reply bank and the approvers accordingly, and move kinds of content between levels as the team's experience and the voice guide mature.

When the steps are done, most teams end up with something like the table below. The levels and approvers will differ; the principle that the check follows the risk does not.

LevelWhat it coversWho approvesRoute
RoutinePlanned posts on the voice guide, everyday replies, resharing the brand's own contentThe person posting, working to the voice guideNo sign-off; checked against the publishing checklist
Fast laneReactive posts, replies to recurring questions from the reply bankThe named approver on the day, or the reply bank itselfSame-day approval or pre-approved lines
CheckedClaims, prices, offers, competitions, partnership posts, reshared customer content, paid boostsThe owner of that risk in the organisation's matrixCompliance, disclosure or permission check before scheduling
StatementResponses on the organisation's behalf when something has gone wrongThe person who speaks for the organisationOne approver, one voice; the routine queue pauses

A pet insurer's approval levels (illustrative)

Illustrative example — not a real result

A pet insurer's approval levels

Velverra Pet Insurance, an invented pet insurer, sends every social post and every reply about cover through its legal team. Over a quarter, replies to customer questions take an average of 26 hours, 3 posts about a seasonal campaign go out after the season has started, and the community manager starts answering claims questions from memory to keep up, which is how a reply describing the wrong excess comes to be posted.

The head of marketing replaces "legal approves everything" with three levels. Everyday posts and replies go out on the voice guide without sign-off. Anything about cover, exclusions, prices or claims goes through compliance, with a reply bank of 40 pre-approved answers to the most common claims questions, written with compliance in 2 sessions. Statements about service problems go through the head of marketing alone. The scheduler gets 4 statuses and a risk column, and nothing moves to scheduled without an approval at its level. In the following quarter, replies to claims questions using the reply bank go out within the community manager's working hours, compliance reviews 31 posts rather than 190, and the review at the end of the quarter adds 9 new answers to the bank and moves posts about vet partner offers from the routine level to the checked level after one goes out without its terms.

Verify it worked

  • Open the list from the first step. Every kind of content should sit at a level, with the reason written beside it.
  • Check the RACI. Each level should name one accountable approver by role, and each approver should have a deputy.
  • Open the calendar or scheduler. Each post should show a status, a risk level and an approver. Nothing should be scheduled without approval at its level.
  • Pick a recent reply about prices, terms or service. It should match a line in the reply bank, and the bank should record who approved it and when.
  • Pick one checked post and find its record on the item. It should show what was reviewed, which version, by whom, when and what they decided.

If it goes wrong

  • Replies still arrive late. The likely cause is a reply bank that lacks the recurring questions, or too much content sitting above the routine level. Add answers with the approvers of the risky topics, and move low-risk content down.
  • A post reaches the queue without approval. The likely cause is an approval given in chat and never copied onto the item, or a scheduler that cannot hold approval states. Copy the approval across and check what the scheduler can enforce.
  • The queue stalls behind one person. The likely cause is no deputy, or several approvers working in sequence. Name a deputy, keep one accountable approver per level and consult the others in parallel.
  • A reactive post carries a claim or a price. The likely cause is a fast lane with no written exit. Add the rule that such posts leave the lane for their normal check, and note the case for the review.
  • A scheduled post goes out into bad news. The likely cause is that nobody was named to pause the queue out of hours. Name who may pause it and how they are reached.

Next steps

Common mistakes

  • One route for everything. Replies wait as long as claims, so the team stops asking. Sort content by risk and give each level its own route.
  • Approval for routine posts. A trusted voice guide is the approval for everyday posts and replies. Save sign-off for what carries risk.
  • No fast lane. A reactive post that waits for the weekly approval session is no longer reactive. Name the approver on the day and build a reply bank.
  • Several approvers in sequence. Each reopens the last, and the post is shaped by the most cautious. One accountable approver per level, others consulted in parallel.
  • Approval outside the tools. Approvals in chat cannot be found later, and unapproved posts slip into the queue. Use approval states and record approvals where the post lives.
  • Forgetting customer and creator content. A reshared photo without permission or a partnership post without a label is a problem whatever the caption says. Route both to their checks.
  • No one able to pause the queue. A scheduled post can go out into the worst possible news. Name who may pause it out of hours.
  • Never reviewing the levels. Levels set on day one are guesses. Review what was held and why, and move content between levels as the team learns.

Frequently asked questions

Does every social post need to be approved?

No. Everyday posts and replies written within a voice guide the team trusts can go out without a separate sign-off, as the site's social media strategy hub argues; requiring approval for everything makes replies late and teaches people to work around the process. What needs approval is what carries risk: claims, prices, competitions, partnerships, customer content and anything said on the organisation's behalf under pressure. Sort what the accounts publish by risk and set approval levels to match.

How do you approve reactive posts without losing the moment?

Decide in advance. Write a short list of what counts as a reactive post, a bank of pre-approved lines for common situations, and the name of the person who can approve a reactive post on the day, as the site's content calendar tutorial recommends. A reactive post that fits those rules goes to that one person; a reactive post that touches a claim, a price or a sensitive subject goes through its normal check, however tempting the moment.

Where should approvals be recorded?

Where the post lives. The site's approvals and compliance guide describes the audit trail as a record attached to the work item of what was reviewed, which version, by whom, when and what they decided. For social, that is usually the calendar row or the scheduler's approval state. An approval given in a chat message should be copied onto the item, so the record can be found when someone asks why a post went out.

Terms used on this page

Frameworks

Where this page is used

Sources

  1. Advertising Standards Authority — ASA

Related