The commerce tooling on the major social platforms is built around a product catalog, and a rebrand is not a catalog row. Where prohibitions exist they are specific and worth quoting. Meta's commerce policy states that commerce content may not promote the buying, selling or trading of downloadable digital content, digital subscriptions or digital accounts. TikTok Shop's US prohibited-products list carries a category named Services and Virtual Products. LinkedIn's own Ads Guide contains no shop, catalog, product-tagging or checkout format at all. Even on the product side Meta has moved off native checkout: its commerce developer documentation describes buyers discovering products on Facebook and Instagram and completing checkout on the seller's website. So every sale in this category ends somewhere the platform does not own: a DM, a call, a scope document, an invoice you send yourself. That is the shape of the category rather than a gap to engineer around, and it changes what the content is for.
The second oddity is timing. A business that needs a brand identity today did not need one last month and will not need one again for years. Demand here is event-driven: a funding round, a product launch, an acquisition, a marketing hire who quit. Your posting does not create that event. What it decides is whether you are the name that surfaces when the event fires. Content in this category is a deposit against a trigger you cannot see coming, which is why the signal that matters most looks like failure on a dashboard: saves from people who do not reply for four months.
Third, the audience you want is small and the audience you will get is mostly peers. Craft tips attract other practitioners. Work that names a business constraint and shows the decision that resolved it attracts the person who owns that constraint. The distance between those two audiences is where a lot of creative practices spend years with healthy engagement and flat revenue.
Who is buying, and how they decide
The buyers split into four groups that decide very differently. Founders and owner-operators at small companies decide alone, fast, and largely on whether they think you will be easy to work with. Marketing managers at mid-size companies decide with a committee, a budget cycle and often a procurement step, which means your work has to survive being forwarded to someone who never saw your feed. Other agencies and studios buy capacity, and they are buying reliability and discretion rather than distinctiveness. In-house creative leads buy overflow, brief precisely, and judge the work against an internal standard you cannot see. Across all four the decision is rarely made in public. It usually starts when someone with a problem asks one or two people they trust for a name, then checks those names against whatever the shortlisted freelancer has published. Social is the verification layer more often than the discovery layer, and that reframes the job: not only to be found, but to hold up when someone who already has your name goes looking. What the buyer is weighing is risk, not craft, because a buyer outside your discipline usually cannot assess craft. They are estimating the odds of a missed deadline, a defensive reaction to feedback, a deliverable they will have to defend internally, or a freelancer who disappears mid-project. Evidence of process, a clearly stated engagement model, and any sign of having worked with a company that looks like theirs all pull that estimate down. Audience size does little for this. One person with a budget reading a case study on LinkedIn outweighs a large impression count, and once the trigger event fires the window is short, so a scoped proposal sent within a day beats a stronger portfolio that answers in a week.
Where this actually sells
Not every platform is worth your time in this category. These are the ones that are, and why.
The content that works
The path from scroll to sale
- A buyer scrolls past work that looks like their problem rather than your taste. The post names a constraint and an outcome, so a non-specialist can tell what they would be buying.
- They do not act. They save it, or follow, or screenshot it. Treat this as the intended outcome of the first touch, because demand in this category is event-driven and the event has not happened yet.
- Proof accumulates over weeks: process, teardowns, a named client result, work in their sector. This is the part most practices quit before finishing.
- A trigger fires: funding, a launch, a rebrand, a departure. The buyer now goes to your profile, not your feed. Bio, pinned posts and link do the work, and a referral from a colleague often arrives in parallel.
- One low-friction step, and only one: a DM, a short form, or a booked call. For paid traffic, a lead form or a click-to-message ad. Then qualify before the call, not on it: decision-maker, budget band, timeline, scope. A three-question form filters better than a thirty-minute call.
- A scoped proposal with options goes out within a day while the trigger is live, then contract, deposit and kickoff. Money moves off-platform, by invoice or your own payment link: TikTok Shop's US list prohibits Services and Virtual Products, LinkedIn publishes no checkout format, and Meta's and Pinterest's own documentation routes the buyer to your site. Delivery then becomes the next content asset, with written permission.
Rules that constrain this category
Policies and regulations change. Treat this as the starting point for your own check, not as legal advice.
- None of the major platforms' commerce surfaces will transact a service for you, so build the off-platform checkout first and treat social as the surface that gets someone to it. The prohibitions that exist are specific. Meta's commerce policy states that commerce content may not promote the buying, selling or trading of downloadable digital content, digital subscriptions or digital accounts, which rules out listing a template pack, a course or a membership in a Meta Shop. TikTok Shop's US prohibited-products list carries a category named Services and Virtual Products. Meta's own commerce developer documentation describes offsite checkout, with buyers discovering products on Facebook and Instagram and completing the purchase on the seller's website. Watch the scope of that policy: it governs Shops and Marketplace listings rather than ordinary posts, so posting about your service is not what it restricts. Two regional points for Canadian practitioners: TikTok Shop is not available in Canada, with no official launch or public roadmap, and TikTok's US affiliate eligibility policy requires creators to be based in the United States, so neither is a route to build on from here.
- The category clears no certification gate, unlike supplements, alcohol, cryptocurrency or financial services, which all appear by name on the platforms' restricted lists. Do not read that as freedom in paid. Google's Other restricted businesses policy states that technical support for consumer technology products and online services provided by third-party providers is not allowed, and that promotions for call directory, forwarding and recording services are not allowed, which catches a real slice of freelance web, IT and automation work across Google and YouTube. Treat every written allowance as a ceiling rather than a description of what you will experience: automated enforcement on these platforms is often stricter than the policy text, so expect rejections on ambiguous creative and keep a version of the ad that makes no claim about outcomes.
- If you run lead ads on Meta, a Facebook Page is mandatory even for an Instagram-only funnel: Meta's documentation states that you need an Instagram account to run a lead ad on Instagram and that the leads generated still belong to the Facebook Page. If you pull leads by API rather than reading them in the Leads Center, Meta's documentation states you cannot retrieve leads while your app is in Development mode, and App Review covering the leads_retrieval and pages_manage_ads permissions plus completed Business Verification comes first. Meta's docs advise against user access tokens and recommend Page access tokens, on rate-limit grounds.
- DM-led selling runs on a clock if it goes through the Messenger Platform. Meta's policy states that businesses have up to 24 hours to respond to a user, that promotional content beyond that window requires either an approved message tag or a sponsored message, that a sponsored message is annotated in the conversation with the word Sponsored, and that the Human Agent tag permits manual human responses within a 7-day period. Meta's own documentation also sets a hard channel boundary: a person logged in to Facebook cannot message your Instagram professional account, and a person logged in to Instagram cannot message your Facebook Page using the Messenger Platform.
- LinkedIn is a lead-capture surface, not a selling surface, and three of its rules bind this category directly. LinkedIn's help documentation states that Lead Gen Forms allow up to three custom questions and twelve selectable fields, that there is no additional cost beyond the associated ad sets, and that cost-per-lead bidding is not available, so you buy on CPC or CPM and work out cost per lead yourself. LinkedIn's ads policy states that ads related to affiliate advertising are prohibited, so a referral-commission model cannot be run as LinkedIn ads at all. If you target EU members with Sponsored Messaging, only those who agreed to see Sponsored Messaging in their inbox will receive it, so model EU inbox volume conservatively. One planning caveat: Sponsored Messaging now centres on conversation ads, and while Message Ads still appear in LinkedIn's Ads Guide, their availability in Campaign Manager has changed, so confirm in Campaign Manager before building a campaign on them.
- Testimonials, reviews and social proof are a directly regulated part of this category in the United States. 16 CFR Part 465, the Rule on the Use of Consumer Reviews and Testimonials, took effect 21 October 2024 and prohibits fake or false consumer, celebrity and insider testimonials, buying positive or negative reviews, insiders creating reviews or testimonials without clearly disclosing their relationships, company-controlled review sites that falsely purport to be independent, certain review-suppression practices, and buying or selling fake indicators of social media influence. Conditioning compensation or an incentive on a review expressing a particular sentiment is covered. Unlike the Endorsement Guides, Part 465 is a trade regulation rule, so the FTC can seek civil penalties against knowing violators, and the FTC announced warning letters to ten companies on 22 December 2025 leaning on exactly that incentivised-review point.
- Disclosure wording is not portable between countries, which matters as soon as you serve clients abroad or promote tools you earn from. The FTC accepts 'ad', 'advertisement', 'sponsored' and constructions such as 'Thanks to Acme brand for the free product', and 16 CFR 255.0(f) requires a disclosure to be difficult to miss, easily understandable, and unavoidable in interactive media. FTC placement guidance rules out a disclosure that sits only on a profile page, only at the end of a post or video, or only behind a 'more' link, and tells creators not to bury it in a group of hashtags; a disclosure left in the comments is avoidable and therefore not clear and conspicuous. UK guidance diverges: the CMA lists '#spon' and '#sponsored' among terms to avoid, and the ASA prefers 'Ad' placed prominently upfront, with '#gifted', '#aff', '#collab' and a bare brand tag all treated as inadequate. Canada's Ad Standards guidance, updated October 2025, names '#Ad' used on its own and not combined with other words as the gold standard, and adds that AI-generated or significantly AI-altered content should be disclosed, for example #MadeWithAI. Canada's Competition Bureau treats failure to disclose a material connection as potentially false or misleading under the Competition Act and has said brands and agencies share responsibility. Platform disclosure tools are not a safe harbour: the FTC states that responsibility rests with the influencer and the brand rather than the platform, and that a platform's built-in tool might not be an adequate disclosure. Disclosure also fixes only the relationship problem, since the FTC is equally clear that you cannot describe experience with a product you have not used. If AI sits in your deliverables and any part of your client base or audience is in the EU, Article 50 of the EU AI Act has applied since 2 August 2026: deployers must disclose image, audio or video content constituting a deep fake, and providers of systems generating synthetic content must ensure outputs are marked in a machine-readable format and detectable as artificially generated or manipulated, with an exemption where the system performs an assistive function for standard editing and does not substantially alter the input data, and a limited-disclosure carve-out for evidently artistic, creative, satirical or fictional work. The AI Act timetable has been the subject of proposed amendments, so verify the current position rather than relying on a date, and settle the practical question in writing: who carries the disclosure obligation on synthetic voice, face or likeness work before you deliver it.
Selling this without owning it
If you want to sell in this category without making the product, these are the routes that exist. The sourcing directory covers each in more detail.
What to measure
- Saves and shares per work post, and saves relative to reach. A save is a buyer filing you against a project that has not started yet, which is the defining behaviour of this category. It is the earliest honest leading indicator you have, and it will look unimpressive next to likes.
- Profile-visit-to-enquiry rate. Buyers arrive at the profile weeks after the post that caught them, often from a referral, so the profile does the converting. If visits are healthy and enquiries are not, the problem is the bio, the pinned posts or the link, not the content.
- Qualified enquiry volume, defined before you start counting: a named decision-maker, a budget band and a date. Raw DM count is vanity in this category, because a large share of inbound is peers, students and speculative pitches.
- Enquiry-to-call, call-to-proposal and proposal-to-close, tracked as three separate rates. Most creative practices assume they have a closing problem and actually have an enquiry-to-call problem, which is a response-speed and qualification problem rather than a persuasion one.
- Days from first touch to enquiry, per channel. Dormancy here is long and uneven, so a channel that looks dead at thirty days can produce at a hundred and twenty. Judging a channel on a monthly window will make you kill the one that works.
- For paid, cost per qualified call rather than cost per lead, plus repeat and retainer conversion on closed work. On LinkedIn you have to compute cost per lead yourself, because its documentation states Lead Gen Forms do not support cost-per-lead bidding and you buy on CPC or CPM.
The mistakes that cost the most
- Posting the finished artefact with no brief, no constraint and no outcome. A buyer outside your discipline cannot assess craft, so they assess judgment, and a beautiful image with no story attached reads as decoration. This single habit is why strong portfolios produce weak pipelines.
- Optimising for an audience of peers. Craft-tips content reliably performs and reliably fails to sell, because the people engaging are other practitioners. Work that names a business problem in a specific sector performs worse and sells better.
- Building the funnel on DM automation and hitting the policy wall at the worst moment. Meta's Messenger Platform policy gives a business up to 24 hours to respond, requires an approved message tag or a sponsored message for promotional content beyond that window, and covers manual human responses within a 7-day period through the Human Agent tag. A thirty-day automated nurture sequence is not deliverable there, and the leads go quiet exactly when they were warm.
- Launching Instagram lead ads without the plumbing. Meta's documentation states that you need an Instagram account to run a lead ad on Instagram, that the leads still belong to the Facebook Page, and that you cannot retrieve leads while your app is in Development mode. The predictable failure is ads spending normally while leads sit somewhere nobody is reading.
- Expecting a shop surface to do the selling, then blaming the content when it does not. Meta's commerce policy bars downloadable digital content, digital subscriptions and digital accounts from commerce surfaces, and TikTok Shop's US prohibited list includes a category named Services and Virtual Products. Build the off-platform checkout first.
- Manufacturing social proof. Offering anything in exchange for a five-star review, having staff, partners or friends write testimonials without disclosing the relationship, running a review page that presents itself as independent, or buying followers to look established all sit inside 16 CFR Part 465, in force since 21 October 2024, which the FTC can enforce with civil penalties against knowing violators. The FTC announced warning letters to ten companies on 22 December 2025 on the incentivised-review point specifically.
A portfolio does not sell a creative service. It lets a buyer infer your judgment, months later, on a day you had nothing to do with.