Brand Guidelines for Social Media: A Practical UK Guide

Brand guidelines for social media are the rules that set how your brand looks, sounds and behaves on every channel you publish to. Most versions of that document contain colour values, a font pairing and four tone-of-voice adjectives saved as a PDF. Helpful for a designer. Not much use at 4pm on a Tuesday, when a founder wants to post from their phone, a creator has sent back content with no ad label, and a scheduled post has become tone deaf after a news event.

That gap is operational rather than aesthetic. A style guide describes what the brand should look like. A social guideline has to tell a named person what they can publish without asking, what needs a second pair of eyes, and who they contact when a post creates risk. The first is a design asset. The second is a working document that gets opened mid-task.

This guide covers what belongs in each of the three layers, how to build the parts teams usually skip, the visual specifications a production team actually needs, and which UK advertising rules have to sit inside your approval workflow rather than in a footnote.

Why do most social media brand guidelines end up unused?

Because they answer a question nobody is asking at the point of publishing.

A traditional brand book covers logo clear space, colour swatches and approved typography. All still necessary. None of it helps when a customer service manager needs to know whether they can reply in the brand voice, or when a creator asks whether their own editing style is acceptable, or when someone has to decide if a promotional claim can go live without review.

The failure mode is predictable. The document gets built once, presented once, saved to a shared drive nobody opens, and then quietly replaced by whatever the person publishing that week happens to think is right. Six months later the LinkedIn account sounds like a consultancy, the Instagram account sounds like a different company, and nobody can say when that happened. We covered the compounding effect of that drift in our piece on why social media consistency builds brand awareness more reliably than posting volume does.

Rigid rules fail differently, and it is worth being honest about that. A firm that forces the same formal caption structure onto a LinkedIn analysis, an Instagram Story and a reply on X will stay technically compliant while becoming irrelevant on two of those three channels. Total freedom produces the drift described above. The workable middle is to sort every rule into one of three buckets: things that must be followed, things publishers can adapt, and things that need performance evidence before they become rules at all.

Working rule: if a publisher has to ask the same judgement call twice, the answer becomes a guideline, a template or an approval threshold.

What should social media brand guidelines include?

Three layers. Identity, channel and operating. The table below is the version we work from when building a brand bible for a new client, and it doubles as an audit: if you cannot name an owner for a row, that row is not really in place.

‍

‍

Most businesses already have the identity layer, usually inherited from a website or logo project. The channel layer tends to exist as informal habit rather than documentation. The operating layer is the one that rarely exists at all, which is why guidelines get ignored: without it, the document has no consequences attached to it.

Within the identity layer, be specific about what “required” means. Core promise, approved claims, disclosure rules, logo integrity, accessibility standards and escalation procedures are non-negotiable. Caption length, opening hook, editing pace, level of informality and creator presentation are adaptable. Content themes, visual variations, calls to action and recurring series are testable, meaning they stay provisional until you have enough data to promote them into rules.

How do you audit your channels before writing any rules?

Start with evidence rather than preference. Choosing adjectives before auditing output is how businesses end up with guidelines that describe a brand they do not actually have.

Build an account register first. List active profiles, dormant profiles, regional pages, product accounts, customer service handles and any account an employee still controls personally. For each one, record the username, the administrator, the recovery email, the publishing owner, the audience purpose and the date of the most recent post. Forgotten profiles confuse customers and create an access risk long after they stop appearing in the marketing plan.

Then read the content itself, and not just the best-performing posts. Sample ordinary posts, replies, captions, graphics, short videos and campaign assets. You are looking for patterns in five areas: language (sentence length, jargon, contractions, recurring phrases), emotional register, the relationship the brand takes with its audience, visual signals (cropping, text density, colour balance, faces, motion), and response behaviour when handling praise, criticism, complaints and misinformation.

Voice stays fixed, tone moves

Voice is the stable character of the brand. Tone is the adjustment you make for context.

A SaaS company might define its voice as clear, capable and generous with information. On LinkedIn that becomes analytical. In a product walkthrough on Instagram it becomes conversational. In a complaint reply it becomes measured and accountable. Those are not three brands. They are one brand behaving appropriately.

The problem with tone-of-voice adjectives is that “be approachable” gives a publisher nothing to act on. A tone matrix does, because it forces a decision:

‍

‍

Finish the audit with three short lists: voice pillars, approved examples and red lines. Then add before-and-after pairs taken from your own channels, showing how a stiff sentence gets clearer and how an unsupportable claim gets rewritten. People learn a brand voice faster by comparing real copy than by memorising adjectives.

What visual specifications does a production team actually need?

“Use the brand colours” is not a visual rule. A production team needs colour values in a format non-designers can paste, a stated purpose for each colour, combinations that meet contrast requirements, and a type hierarchy covering headlines, supporting copy, in-graphic captions and subtitles. If a font is unavailable in a platform-native editor, name the fallback rather than leaving the publisher to pick one.

For hard numbers, the Office for National Statistics publishes its full social media design specification openly, which makes it one of the more useful reference documents available to UK teams. Its artboard sizes:

‍

‍

The spacing figures matter more than the dimensions, because platform interfaces crowd and cover text. ONS uses 140px of padding at the top, 110px at the left and right, and 230px at the bottom, with the publication credit sitting 85px from the bottom of the artboard. Text boxes get 40px of padding on all edges and a 20px corner radius, with 30px gaps when they are stacked as a list. Chart text is never set smaller than 28 point. Copy is left-aligned rather than justified, sized to read on both desktop and mobile, and checked for contrast against its background.

Treat those as a starting point rather than gospel, then test every template in the live interface. Stories and Reels need their own safe zones because interface controls, captions and account elements land in different places depending on placement.

Beyond specs, a good template removes the choices that cause errors. Lock the grid, logo treatment, text hierarchy and export settings. Leave the content idea free to change. Build separate templates for educational slides, customer proof, announcements, quotes, product demonstrations and vertical video covers, and store them where publishers work rather than in a folder only designers can open.

For photography, document subject matter, lighting, cropping and retouching. For illustration, specify line weight, texture, perspective and whether icons can sit alongside photography. For motion, define opening-frame behaviour, subtitle treatment, logo timing, transitions and whether sound is assumed. Do not require every video to use an identical animation. Give editors a recognisable motion language and let them apply it. The distinction between a recognisable system and a repeated template is also where designer-made content pulls ahead of AI-generated posts: a system can be adapted, a template can only be refilled.

How do you adapt across platforms without losing brand identity?

By deciding in advance which brand signals have to survive adaptation, and writing that decision down.

Message. The problem you solve and the promise you make stay fixed. The opening hook and the order you explain things in can change freely.

Visual identity. Core colour relationships, type principles and logo integrity stay fixed. Cropping, composition, pace and creator framing can change.

Voice. Vocabulary, values and level of honesty stay fixed. Formality, humour and emotional intensity can move.

Behaviour. Disclosure, moderation and response standards stay fixed. Reply length and platform-native interaction can vary.

Content. Approved facts and positioning stay fixed. Format, duration and narrative structure are open.

In practice that means an e-commerce brand can run polished Instagram carousels alongside informal Stories filmed in the packing area. The Stories do not need studio lighting. They do need accurate product information, respectful staff presentation, consistent product naming and a visual cue tying them to the main account.

A professional services firm faces a harder test on TikTok, where the temptation is to adopt a trend with no connection to the expertise being sold. The version that tends to work is narrow: one subject-matter specialist answering one question people actually ask, filmed directly and edited lightly. Native in execution, unchanged in substance.

External creators need their own boundary sheet rather than a copy of the full guidelines. Give them the campaign objective, mandatory product facts, prohibited claims, disclosure requirements, usage permissions, an approval contact and a crisis route. Then leave the delivery alone. Rewriting every sentence in the brand voice produces content that audiences correctly read as a corporate script.

Who approves what, and what do UK rules require?

If everyone can publish, nobody owns the risk. Missing ownership is one of the most common reasons a business’s social presence stalls, which we set out in more detail in our breakdown of why small business social media isn’t working.

Start with access. Name the people who can draft, approve, publish, moderate comments, manage paid activity and respond during an incident. Use role-based permissions where the platform supports them, remove access the day responsibilities change, and keep the account register from your audit current. Then set approval tiers:

1. Routine content. A trained publisher prepares and schedules using approved templates and factual sources. No second review needed.

2. Commercial content. A designated reviewer checks pricing, offers, product claims, targeting and disclosure language before scheduling.

3. Sensitive content. A senior owner reviews complaints, regulated topics, personal data, employment matters, political subjects and any reference to competitors.

4. Incident content. One named decision-maker controls the response. Everyone else pauses scheduled activity when instructed.

The UK Government’s social media playbook is worth reading for the operational detail here. Government Digital Service runs a standard responses document covering its most frequent queries with pre-approved wording, plus a traffic-light escalation process with risks reviewed quarterly and scenario-based refresher training. It also targets an initial response to all enquiries within two working days, and defines three to five content pillars that every post has to align with. Those are all rules a five-person marketing team can copy.

Disclosure belongs in the workflow, not the appendix

The Competition and Markets Authority is direct about where responsibility sits. Its guidance for brands states that businesses must take responsibility for ensuring paid-for endorsements promoting them are properly labelled as ads, and that unlabelled paid endorsements may breach consumer protection law. Under the Digital Markets, Competition and Consumers Act 2024, whose review and endorsement provisions came into force on 6 April 2025, the CMA can now enforce consumer law directly rather than through the courts.

The Advertising Standards Authority publishes the practical half of this, and its rulings are specific about what fails. The ASA’s advice on recognising ads on social media records repeated cases where a brand tag or an “@ mention” alone was found insufficient to identify a post as advertising, with brands and influencers told to use a clear and prominent identifier such as “#ad” instead. On platform tools, the ASA notes that its 2026 research found labels like Instagram’s paid partnership tool can help people recognise ads but may not be enough by themselves, so an additional “Ad” label is advisable where the platform label is obscured or unclear.

So your guideline needs to state where the disclosure appears, who verifies it before scheduling, how gifted products and affiliate links are handled, and what happens when a creator submits content without the label. “Follow advertising rules” is not a rule. Approved wording and example placements are.

Apply the same discipline to user-generated content: record how you request permission, where you store evidence of it, how you credit the creator and what you do with a withdrawal request. Copyright verification is a publishing gate, not an assumption.

Check whether the Online Safety Act applies to you

This is worth being precise about, because the point is widely muddled. Online Safety Act duties fall on the provider of a user-to-user service. Your company page on LinkedIn or Instagram is not the regulated service; the platform is. What can bring a business into scope is hosting user-to-user features on its own property, such as a forum, a community space, a review section or comments where users can respond to each other.

Ofcom estimates that over 100,000 online services are likely to be in scope, from the largest platforms down to the smallest community forum, and it publishes guidance and a scope checker aimed specifically at small services. If you run anything like that, establish scope first, then document controls: a route for flagging potentially illegal content, a process for child-safety concerns, records of moderation decisions and a named person who seeks specialist advice when a case exceeds the team’s competence.

How do you keep the document current?

Guidelines earn their place by improving decisions, not by getting longer.

Measure whether publishers actually use the approved templates, whether posts clear the accessibility and disclosure checks, and whether recurring themes produce useful audience responses. Then set a review cycle: content (which formats are hardest to produce correctly), voice (which phrases land and which confuse), visual (which templates stay legible across devices), compliance (what has changed in disclosure or moderation requirements), access (does every current role still hold the right permissions), and learning (what belongs in the do and don’t examples now).

On benchmarks, the government playbook gives a defensible standard: if you post every day, three months of data is enough to set benchmarks, and if you post weekly, collect six months or more first. That is a useful brake on rewriting a brand because one post underperformed.

Platform changes should trigger a focused update rather than a rewrite. When a new format appears, define its safe zones, caption treatment, approval route and measurement method, and leave the rest alone. When a recurring theme works, document the principle behind it rather than the post, so the team can adapt it. That is the same discipline that separates useful repurposing from recycling: our guide to repurposing old content covers how to tell which ideas deserve a new format and which should stay retired.

What we’ve learned building brand bibles

Socials Sorted is a fully managed social media service that gives businesses a dedicated London-based team of designers, copywriters and strategists handling strategy, design and posting for a flat monthly fee from £399 a month. We build a brand bible for every client during onboarding, from an intake form plus research on the brand, its competitors and its audience, and the client approves it before we produce a single post.

Two things have held up across the 7,000-plus posts we’ve published for more than 200 brands. First, the identity layer is almost never the blocker. Colour and type get agreed quickly. What breaks consistency is the absence of a named owner and a working approval route, which is why every client gets a portal where content is approved before anything is published, and why typical client input runs to around 30 minutes a month rather than a weekly meeting. Second, guidelines that get written once and never touched again decay within about two quarters. Ours get revised continuously against what the account is actually publishing, and revisions to in-plan content are unlimited.

Worth stating the limits plainly. We cover strategy, content creation, copywriting, scheduling and publishing. We do not run community management, creator or influencer programmes, or paid social. If your operating layer needs to cover comment moderation, creator disclosure sign-off or an Online Safety Act escalation route, those stay with you or with another supplier, and your guidelines still need to say who owns them.

Which version of this should you build?

If you publish on one or two channels and one person does all of it, skip the document. Write a one-page working sheet: voice pillars, three to five content pillars, the required rules, template locations and one escalation contact. Anything longer will not get read.

If two or more people publish, or if creators, agencies or client-facing staff post on your behalf, build all three layers properly. The operating layer is where the return is, because approval tiers and named owners are what stop the drift that a colour palette cannot.

If the constraint is production rather than definition, and you have the rules but not the hours to apply them every week, outsourcing the execution is the more efficient route. We break down what that costs against DIY, freelance, in-house and agency options in our guide to social media budgets for small businesses.

If you’d rather hand the whole thing over, Socials Sorted builds the brand bible during onboarding and then produces against it every month, with first content typically delivered within five working days and a published 30-day money-back guarantee. Send us your channels and whatever brand materials you already have.

‍

‍

FAQs

How long should social media brand guidelines be?

Short enough to be opened mid-task. A solo publisher needs one page. A team publishing across four or five channels usually needs somewhere between ten and twenty pages, most of which should be examples, templates and approval rules rather than prose. Length is not the measure. Whether a publisher can find the relevant answer in under a minute is.

Are brand guidelines and a brand bible the same thing?

Broadly, though the labels get used loosely. “Brand guidelines” often means the identity layer only: logo, colour, type, tone adjectives. A brand bible, as we build it, is meant to include the channel and operating layers as well, so it covers who publishes, who approves and what flexes per platform. Check which one a supplier means before agreeing scope.

Who should own social media brand guidelines?

One named individual, not a department. In smaller businesses that is usually the founder or marketing lead. The owner does not have to write every rule, but somebody has to hold the review cycle, approve changes and be the escalation contact during an incident. Shared ownership reliably becomes no ownership.

Do you need separate guidelines for each platform?

No. One document with a channel layer works better than five documents, because separate files drift apart and nobody reconciles them. Give each platform a section covering format, caption structure, pacing, visual treatment and interaction style, and keep the identity and operating layers shared across all of them.

How often should you review them?

Set a scheduled review twice a year for the full document, plus a focused update whenever a platform ships a format you intend to use or a disclosure requirement changes. Do not rewrite the guidelines in response to a single underperforming post. Use enough data to tell a pattern from noise, which is roughly three months if you post daily.

What should you give an external creator instead of the full guidelines?

A one-page boundary sheet: campaign objective, mandatory product facts, prohibited claims, disclosure requirements and approved label wording, usage permissions, approval contact and crisis route. Do not send the visual specification or ask them to write in your brand voice. You are protecting the required information, not the delivery style, which is what they were hired for.