Social Media, Influencer and Digital Marketing: Platforms, Creators and Campaigns That Prove Their Worth

Kaplan and Haenlein (2010, p. 61) define social media as "a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of User Generated Content". The definition is old, but its emphasis still matters: the defining feature of social media is that users, not brands, make most of the content and most of the meaning. A brand is a guest in other people's conversations, and a strategy that treats social platforms as free broadcast channels misunderstands the medium from the start.

Kietzmann et al. (2011) give managers a more practical lens. Their honeycomb describes seven functional building blocks: identity (how much users reveal about themselves), conversations, sharing, presence (knowing whether others are available), relationships, reputation and groups. Platforms emphasise different blocks, and so do audiences. The authors' point is not that a brand should try to excel at all seven, but that it should work out which functions matter to its audience on a given platform and build its activity around those, while monitoring the others.

For Moorfield Botanics, botanical skincare for sensitive and reactive skin, the strongest blocks are conversations, sharing and reputation: people with reactive skin ask each other what worked, share routines and judge products by other users' experience. Presence and identity matter far less. That analysis already narrows the job social has to do: support peer conversation and earn reputation, rather than simply display products.

A social media strategy, then, is a set of choices: which audiences, on which platforms, in what role, towards which measurable objectives, in service of business goals. It is equally a statement of limits. Social can build awareness, community and consideration, and it can drive sales, but a brand controls neither the algorithm nor the conversation; it cannot make people care about a category they don't care about, and it cannot fix a weak product. Felix, Rauschnabel and Hinsch (2017) treat social media marketing as a cross-functional activity that uses social media to achieve organisational goals by creating value for stakeholders, which makes it a matter of strategy rather than a fragmented set of tactics. The rest of this module builds that consistency, one decision at a time.

Platforms are not interchangeable pipes of different sizes. Each has affordances, the actions its design makes easy or hard (resharing, stitching, saving, commenting, searching), and each decides distribution by its own logic. Those differences shape what content spreads, how audiences behave and how much a brand's follower count is worth.

The most important shift of the last decade is from the social graph, in which you mainly saw posts from accounts you followed, to recommendation-led feeds that rank content by predicted interest. TikTok made this explicit: its For You feed ranks videos on user interactions, video information and device and account settings, and TikTok states that "neither follower count nor whether the account has had previous high-performing videos are direct factors" in recommendation (TikTok, 2020). Instagram followed. Mosseri (2023) explains that Instagram has no single algorithm: Feed, Stories, Explore and Reels each have their own ranking, with Stories weighted towards close connections and Reels and Explore built largely on recommendations from accounts the user does not follow. YouTube and Pinterest behave partly as search engines, so content can be found long after it is posted; LinkedIn's audience is there in a professional role, which suits corporate, recruitment and B2B communication more than consumer skincare.

The implication is double-edged. Reach depends less on audience size and more on how each piece of content performs, so a small brand can reach strangers. But followers are a less reliable asset, and organic reach becomes something earned post by post rather than owned. Much sharing also happens out of sight in messaging apps and email, what Madrigal (2012) called "dark social", so public engagement understates how much a brand is passed on.

For Moorfield, this explains a pattern Amara expects to find: the 84,000 Instagram followers do not guarantee that a post is seen, while a TikTok from the 41,000-follower account can reach people who have never heard of the brand.

Platform audiences and features change quickly. Any usage figure you use should be cited and dated; Ofcom's annual Online Nation report (Ofcom, 2025) is the authoritative UK source, and its December 2025 edition is the latest at the time of writing. Where you can't cite a figure, describe direction and relative differences instead.

A platform portfolio is the set of platforms a brand chooses to be active on, with a distinct role for each. The decision is a trade-off, because every platform has to be resourced properly to work: content made natively for its formats, someone to reply to comments, budget to amplify what performs. In a recommendation-led feed, a thinly run account is not a cheap presence; it is a stream of under-performing posts. Fewer platforms done well usually beats many done thinly.

A weighted scorecard makes the trade-off explicit. Five criteria cover most decisions. Audience presence asks whether the people the brand needs to reach use the platform, in that role, often enough. Objective fit asks whether the platform's logic serves the objective: recommendation feeds suit discovery, search-led platforms suit education and consideration, close-connection formats suit community. Content capability asks whether the team can produce what the platform rewards. Cost to resource, scored so that a higher score means more affordable, asks what it would take to do the platform properly. Competitive opportunity asks whether there is room to stand out. Weights reflect the brand's priorities; changing them and seeing whether the ranking holds is part of the analysis.

The scores rank platforms; they do not make the decision. The manager then assigns roles. A primary platform gets sustained investment and is expected to deliver against a main objective. A secondary platform has a narrower job, such as search and education, with lighter, lower-frequency investment. A limited-role platform serves a specific non-marketing purpose, such as corporate communication or recruitment, and is resourced accordingly. Dropping a platform is a legitimate outcome, provided the account is handled responsibly rather than left to decay.

Resource is where most portfolio decisions are really made. Moorfield Botanics has 38 employees, and its 2027 digital marketing budget of £720,000 allocates £110,000 to content production and £260,000 to paid social. Those sums have to cover every platform the brand keeps. Spread across four platforms equally, they would buy four mediocre presences; concentrated, they can buy two strong ones and a modest third. Felix, Rauschnabel and Hinsch (2017) treat scope as one of four strategic choices in social media marketing, ranging from defenders, who use social mainly to broadcast to customers, to explorers, who collaborate with many stakeholders. How widely to spread across platforms is a related choice, and it should follow from strategy and resources, not from what competitors happen to be doing.

Objectives connect social activity to the reasons the business invests in it. A useful structure is a cascade: business goals (revenue, customers, margin), marketing objectives that serve them (awareness among a target segment, acquisition, retention), and social objectives that state what each platform contributes. Every social objective should be traceable up the cascade. If it can't be, it is an activity measure, not an objective.

Doran (1981) gave managers the SMART test: objectives should be specific, measurable, assignable, realistic and time-related. The letters are now often read as achievable and relevant, but the discipline is the same. "Grow engagement on TikTok" fails; "increase the number of new customers who first discover Moorfield through TikTok, measured by post-purchase survey, by December 2027" passes most of the test, once a baseline is measured and a target is set against it.

The second distinction is between brand-building and activation. Binet and Field (2013), analysing campaigns in the IPA Effectiveness Databank, found that long-term brand-building and short-term sales activation work differently and that the most effective campaigns balanced the two, with roughly 60% of investment on brand-building. Treat this as a guideline from their analysis, not a law: their later work showed the right balance varies by category, brand size and context (Binet and Field, 2018), and their evidence is about advertising budgets in general rather than organic social specifically. The lesson for social is that reach-based, memorable brand activity and conversion-focused activity need different objectives and different measures, and a plan made only of activation is likely to under-invest in future demand.

That links to reach versus engagement. Engagement rates are easy to report but can rise while reach falls, and a small, highly engaged audience may be people who would have bought anyway. Follower counts, likes and impressions are vanity metrics when reported alone, because they do not show whether anything changed for the business.

For Moorfield, the business goals for 2027 are explicit: revenue from £6.2M to £7.5M (+21%) and new customers from 48,000 to 60,000 (+25%). Any social objective Amara sets has to show how Instagram and TikTok contribute to those numbers. The social objectives set here are the ones Module 7 later tests against evidence.

Most marketing plans still describe their audience demographically: women aged 25–44, ABC1, urban. On social platforms that description does less work than it seems. Recommendation-led feeds, which you met in Module 1, distribute content by predicted interest rather than by who follows whom, so two people with identical demographics can see entirely different versions of a category. What separates them is what they come to social media for and what they do once there.

Uses and gratifications research offers the starting point: people actively choose media to satisfy needs. Applying it to brands, Muntinga, Moorman and Smit (2011) interviewed social media users about their brand-related activity online, which they termed consumers' online brand-related activities (COBRAs), and proposed three levels of activity. Consuming is watching, reading and following without visible response; contributing is liking, commenting, rating and joining conversations; creating is producing brand-related content such as reviews, tutorials and routines. They linked motivations to each type of activity: consuming was driven mainly by information, entertainment and remuneration (rewards); contributing by personal identity, social interaction and entertainment; and creating by those same motives plus empowerment. The proportions matter: in most categories far more people consume brand content than create it, so a strategy that assumes everyone wants to "join the conversation" will usually overestimate participation.

A need-based segment combines three things: the need a person brings to social media in this category, the behaviour that need produces (where they look, whether they consume, contribute or create), and the evidence that the segment exists. "Reactive-skin researchers" — people who search platforms and forums for proof that a product will not trigger a reaction — is a need-based segment; "women 25–44" is not. Such segments are not a replacement for the segmentation and targeting work a brand already does; they are a social-specific layer that tells content and community teams what each group is looking for.

Two cautions. First, motivation research is largely interview- and survey-based, and self-reported motives can flatter respondents; observed behaviour, which the next lesson collects, is a useful check. Second, need-based segments must still be sized and reached. A vivid segment that is tiny, or invisible to the platforms' targeting, may be better served by community than by paid media.

Social listening is the systematic collection and analysis of public conversation about a brand, category or topic in order to generate insight. It differs from social monitoring, which tracks individual mentions so that someone can reply, fix a delivery or spot a complaint. Monitoring is operational and runs continuously; listening is research, framed by questions and judged by whether it changes a decision.

A listening programme starts with questions, not keywords: what do people with this need say about the category, what language do they use, which competitors do they compare, what tensions recur? The questions then drive the build. Queries combine brand, product and category terms with Boolean operators and exclusions (a brand name that is also a common word will drown the sample in noise). Sources need deliberate choice, because venue shapes what people say. Schweidel and Moe (2014) found that sentiment differed systematically between blogs, forums and microblogs, and that listening to one venue, or pooling venues without adjusting, could give misleading brand sentiment. Platform access also varies: some platforms limit what third-party tools can collect, so a tool's coverage should be checked rather than assumed. Sampling decisions — period, language, geography, how spikes are handled — should be written down so the study can be repeated.

Netnography, developed by Kozinets (2002; 2020), adapts ethnography to online communities and social media. Its contribution is rigour of interpretation: immersion in the community's context, reading whole threads rather than isolated posts, attention to insider language and humour, and reflexive notes on what the researcher brings. That matters because automated sentiment is the weakest part of most tools. Hartmann et al. (2023), in a meta-analysis of 272 datasets, found that newer transfer-learning models classify sentiment substantially more accurately than the dictionary-based lexicons many tools have used — but no method is reliable with sarcasm, slang or mixed feelings. Treat a sentiment score as a rough indicator to be read against the posts behind it, never as a precise measurement.

Ethics is part of the design. The Association of Internet Researchers' guidelines (franzke et al., 2020) reject a simple public/private split: people posting in a support forum may not expect their words to be quoted in a brand presentation. Under UK GDPR, personal data remains personal data when it is publicly visible, so collection, storage and retention need a lawful basis and proportionate handling. In practice: collect only what the questions need, report themes rather than identifiable individuals, and paraphrase rather than quote sensitive posts. This is not legal advice; involve your data protection lead.

A listening study produces thousands of posts; a decision needs a handful of insights. The bridge is coding. Thematic analysis (Braun and Clarke, 2006) offers a disciplined route: become familiar with the data, generate initial codes, search for themes, review them against the data, then define and name them. In listening work, a first pass usually codes each post for topic, sentiment, the need it expresses and who is speaking (customer, prospect, creator, competitor's customer). Two people coding the same subsample and comparing results is a cheap check on whether the codes mean the same thing to everyone.

Counting codes gives share of conversation — the proportion of mentions in each theme — but share is only one signal. A small theme can matter more than a large one if it is intense (strong emotion, long posts, repeated across venues), new (rising over the period) or commercially consequential (it blocks purchase). Noise needs removing before anything is counted: bots and spam, competition entries ("tag three friends"), news spikes driven by a single story, and sarcasm that a tool has scored the wrong way round. A spike is a question, not a finding — find out what caused it before treating it as a trend.

An insight is not an observation. "31% of mentions concern reactive skin" is an observation; it says what, not why or so what. A useful insight statement has three parts: an observation grounded in the data, the human tension or motivation that explains it, and an implication for a decision. Reading the reactive-skin posts behind Moorfield's figure revealed people describing their own routines in detail and discounting brand claims, which produced the insight "People with reactive skin trust other users' routines more than brand claims". Its implication is clear enough to act on: show real users' routines, and give those users somewhere to share them.

Listening has blind spots. It hears only those who post publicly, over-represents strong opinions and under-represents customers who are satisfied and silent. A good insight is therefore triangulated with first-party data — reviews, customer-service logs, search queries, purchase data — before it drives product or budget decisions. These insights become the raw material for content pillars in Module 3 and for judging audience fit in Module 4.

Followers are an audience; a community is something more. Muniz and O'Guinn (2001), studying owners of Ford Bronco, Macintosh and Saab, defined a brand community as a specialised, non-geographically bound community based on a structured set of social relationships among admirers of a brand. They identified three markers. Consciousness of kind is members' sense of connection to each other, and of difference from non-members. Rituals and traditions — shared stories, greetings, celebrations of the brand's history — keep that consciousness alive. Moral responsibility is a felt duty to the community and its members, seen when experienced members help newcomers or defend community norms.

Schau, Muñiz and Arnould (2009), drawing on nine brand communities, showed how communities create value through practices that members perform with little or no help from the firm. They grouped twelve practices into four categories: social networking (welcoming newcomers, empathising with members' problems, governing behaviour), impression management (evangelising for the brand and justifying the time spent on it), community engagement (staking out areas of expertise, marking milestones, badging them and documenting experiences) and brand use (grooming, customising and commoditising). The managerial point is that value comes from members doing things together, so a community concept should design for those practices rather than for brand broadcasting.

Community membership has consequences for the brand. Algesheimer, Dholakia and Herrmann (2005), studying European car clubs, found that identification with a brand community led to greater engagement and loyalty intentions — but also to normative community pressure, which could produce reactance. Communities develop their own norms and may resist a brand that tries to direct them.

Two design choices follow. First, purpose: a community needs a shared need that members care about more than the brand itself — for Moorfield, living with sensitive skin rather than "loving Moorfield". Second, location. A platform-hosted community (a Facebook group, a Discord server, a subreddit) brings existing users and low cost but leaves the brand subject to the platform's rules, algorithm and data access. An owned community (on the brand's site or app) gives control over data, moderation and experience, but every member must be recruited and given a reason to return. Neither is right in general; the choice depends on where the audience already is, what data the brand needs and how much moderation it can resource. This is a different decision from the platform roles set in Module 1: Instagram remains where Moorfield builds public community and can invite people into the Circle, while the Circle is where members help one another.

A community is not self-running. Even the most active communities rely on a few people who welcome newcomers, answer questions, enforce norms and notice problems early. Typical roles are a community manager (accountable for purpose, health and the link to the business), moderators (staff or trained volunteers who apply the guidelines) and member champions (respected members given recognition and early access in return for leading by example). Roles should be written down, with hours and escalation routes, before launch, because an unmanaged moderation backlog is a common way for new communities to fail.

Community guidelines translate purpose into rules members can see. Good guidelines are short, specific about what is not allowed and explicit about what happens when a rule is broken. In a health-adjacent category such as skincare, they must also draw lines the law and member safety require: members can share their own experience, but no one — member or brand — should give medical advice or make medicinal claims. Moderation can be pre-moderation (posts checked before they appear), post-moderation (checked after) or reactive (checked when reported); many brand communities combine reactive moderation with keyword alerts and a fast route for anything involving health or safety.

Negative conversation is not a failure. Members criticising a product inside the community is useful listening and, handled openly, a sign of trust. What needs managing is escalation: the same complaint from many members, a thread migrating to public platforms, members turning on each other, or a moderator decision becoming the story. Each should trigger a defined response — acknowledge, investigate, report back — and a named person to own it. A full reputational crisis needs the brand's crisis plan, which this module does not cover.

Fournier and Lee (2009) warn against common myths, among them that a brand community is a marketing strategy rather than a business strategy, that successful communities are tightly managed and that online social networks are the key to community strategy. Health metrics should therefore look past headcount. Useful measures include active members as a share of total members, member-to-member replies (not just brand-to-member), the proportion of questions answered by members, retention of new joiners and the sentiment of community conversation, read qualitatively as well as scored. Set a target and a review date for each, and connect at least one measure to a business outcome so that the community can justify its budget. The Social Listening and Community Plan brings listening and community plans together in one document.

Most brand content fails quietly: it is not disliked, simply not watched, saved or passed on, and in a recommendation-led feed, as you saw in Module 1, content that fails to hold attention is shown to fewer people. So before planning what to post, ask why people share anything at all.

The most cited evidence comes from Berger and Milkman (2012), who analysed every New York Times article published over a three-month period and which ones made the paper's most-emailed list. Positive content was shared more than negative content, but valence was not the main story. Sharing was driven partly by physiological arousal: content evoking high-arousal emotions, whether positive (awe) or negative (anger, anxiety), was more likely to be shared, while content evoking low-arousal emotions such as sadness was shared less. Practically useful content was also shared more. The study concerns news shared by email, not branded TikToks, so treat its findings as a hypothesis for social content, not a rule.

Why people share is as important as what they share. Berger's (2014) review of word-of-mouth research identifies five functions it serves for the sharer: impression management (what sharing says about me), emotion regulation, information acquisition, social bonding and persuading others. In his popular synthesis, Berger (2013) groups the drivers as social currency, triggers, emotion, public visibility, practical value and stories. For a manager, the useful question is which function a piece of content performs for the person who passes it on. A saved routine is practical value; a shared sustainability post may be identity signalling, saying "I am someone who cares about waste".

Spread is not the same as value. Akpinar and Berger (2017) compared emotional and informative ad appeals using sharing data and experiments: emotional appeals were shared more, but informative appeals did more for brand evaluations and purchase intent, and the best performers were emotional content in which the brand was integral to the story rather than bolted on at the end. Their phrase "valuable virality" is the test this module applies. A post that travels while nobody recalls whose it was has bought attention for someone else.

That is why "going viral" is a poor objective. Virality is rare, unpredictable and largely outside a brand's control. What a brand can plan for is content that reliably performs one of the sharing functions for its audience, with the brand embedded in it, published often enough for the occasional outlier to happen. For Moorfield Botanics, the listening insight from Module 2 — "People with reactive skin trust other users' routines more than brand claims" — already suggests which function matters most: practical value delivered by peers, not announcements delivered by the brand.

Without structure, social output drifts towards what is easiest to make: product shots, offers and announcements. A pillar is a recurring theme that a brand's content is organised around, each tied to an audience need and a brand objective. Pillars keep a plan coherent over months, focus idea generation and make performance comparable, because posts in the same pillar can be judged against each other.

Good pillars are derived, not brainstormed. Each should trace back to an insight about what the audience needs (from listening, research or first-party data) and to an objective the brand has set (from the cascade in Module 1). A pillar that answers an audience need but serves no objective is a hobby; one that serves an objective but meets no audience need is advertising that people will scroll past. Three to five pillars is a practical range: fewer gives too little variety, more spreads production thin.

The pillars should also be balanced across jobs. Some content builds the brand (memorable, distinctive, reaching people who are not yet customers), some serves the community (giving existing customers reasons to participate) and some converts (moving interested people to buy). The research suggests the balance matters. Ashley and Tuten (2015), in a content analysis of top brands' social content, found that most of it was functional, while experiential, image and exclusivity appeals were the ones associated with higher engagement; the design was correlational, so it shows association rather than cause. Lee, Hosanagar and Nair (2018), coding over 100,000 Facebook posts from 782 companies, found that content expressing brand personality, such as humour and emotion, was associated with more engagement, while directly informative content such as prices and deals reduced engagement when used alone but increased it when combined with brand-personality content — and deals did drive click-throughs. The lesson is not to avoid conversion content but to stop it crowding out everything else.

Each pillar needs a success measure that matches its job, and none of them should be a follower count. A community pillar might be judged on saves, shares and replies; a brand-building pillar on reach among non-followers; a conversion pillar on tracked visits to product pages. Measures should be set against a baseline, as the strategy module argued, and reviewed: a pillar that consistently under-performs its measure is a candidate for rethinking or retirement.

For Moorfield Botanics, the key insight points directly at one pillar. If people with reactive skin trust other users' routines more than brand claims, then a pillar built on real customers' routines will do more than one built on Moorfield's own product claims — and it connects naturally to the Skin Circle, where members already share those routines.

Platform-native creative is made in the format, length, tone and grammar of a specific platform rather than produced once and cross-posted. Each platform rewards different behaviour: a TikTok is watched full-screen with sound, often by someone who has never heard of the brand; an Instagram carousel is swiped through and saved by someone who wants to come back to it; a YouTube video may be found months later through search. One asset resized for all three is usually second-best on each; an idea can travel, but its execution should be rebuilt for each platform.

On video platforms the first moments carry most weight. The hook is the opening of a piece of content — typically the first one to three seconds of a video — designed to stop the scroll. Teixeira, Wedel and Pieters (2012), tracking viewers' facial expressions, eye movements and decisions to skip internet video ads, found that surprise concentrated attention and joy retained viewers — the level of surprise mattered most for attention, while rising joy kept viewers from skipping — so an emotional payoff saved for the end reaches only those who are still watching. It was a small laboratory study of conventional ads, so take the principle, not a benchmark: earn attention at the start, then give people a reason to stay. Hook devices include showing the result first, naming the viewer's problem in their words, or posing a question the video answers.

Brand prominence needs handling with care. Viewers skip content that looks like an advert early, but content in which the brand never appears does nothing for the brand, as the "valuable virality" argument showed (Akpinar and Berger, 2017). The resolution is to integrate the brand into the story — the product in use within the routine — rather than opening with a logo or closing on a pack shot nobody stays for.

Several other choices make creative native. Sound is central on TikTok and Reels but many people watch without it, so on-screen text and captions carry the meaning either way and make the content accessible. Format follows purpose: short vertical video for discovery, carousels for step-by-step information people save, live sessions for questions and community, and long-form video for depth and search. A lo-fi, creator-style look often reads as more authentic among ordinary users' posts, but polish suits content that needs precision. Finally, social platforms increasingly act as places to search, so the words people actually use (from listening) belong in captions, on-screen text and titles. Format specifications such as video lengths and carousel limits change often; check each platform's current guidance.

These principles apply equally to creators' content, which Module 4 takes up.

User-generated content (UGC) is content about a brand made by its customers: reviews, routines, unboxings. For an audience that trusts peers more than claims, it is the most credible content available. Goh, Heng and Lin (2013), linking a Facebook fan-page community's posts to members' purchases, found that engagement in the community was associated with higher purchase spending, and that both user- and marketer-generated content influenced purchasing through the information and persuasion they carried, in different ways.

UGC has to be encouraged, curated and cleared. Encouragement means giving people a reason and a format, such as a routine-sharing prompt; curation means choosing what to amplify against the pillars. Clearance is the step brands most often skip. A customer owns the copyright in what they post; platform terms license the platform, not the brand. Reusing it therefore needs the person's explicit, recorded permission, covering where, for how long and with what credit; use in paid advertising needs wider permission than an organic repost. And once a brand shares a customer's words, it adopts them: a customer's claim that a product "cured" their eczema becomes a medicinal claim made by the brand. Claims rules for cosmetics and the evidence needed to support them are covered in Module 5.

Generative AI is genuinely useful for ideation, caption variants, versioning one idea into several formats, transcription and alt text. It harms where it substitutes for what the audience values — real people and real experience — or introduces errors no one checks. An AI-generated "customer" with flawless skin is not just inauthentic; it misleads about what a product does.

The UK advertising codes are media-neutral: ads made with AI face the same rules on misleadingness, harm and social responsibility as any other. CAP's guidance on disclosing AI in advertising, published on 29 May 2025, sets no blanket requirement to label AI use; instead it asks whether the audience would be misled without disclosure, and warns that a label cannot cure an ad that is misleading in itself (CAP, 2025). The ASA has ruled against beauty ads whose filters exaggerated a product's effect (influencer posts for tanning and cosmetics brands, 2021); AI imagery raises the same risk. Platforms also have their own labelling rules: TikTok requires creators to label realistic AI-generated content, YouTube has required disclosure of realistic altered or synthetic content since March 2024, and Meta applies an "AI info" label (introduced in July 2024) — all as at the time of writing and changing often, so check current rules. A sound policy states the principle in-house now: AI may help make content, but it may not fake the people, results or evidence the content depends on.

A content plan turns pillars and creative principles into a schedule a team can deliver week after week. Cadence is a resourcing decision before it is a creative one. Every post on the plan needs an idea, production, editing, captions, approval, publishing and someone to reply to the comments it draws. A cadence the team cannot sustain produces rushed, under-performing posts — exactly what recommendation-led feeds punish — so the right question is not "how often should we post?" but "how often can we post content that meets our standard, with the people and budget we have?"

A plan also has to integrate two rhythms. Always-on content runs every week across the pillars; campaign content is concentrated around a launch or moment. A useful lens is the hero, hub and hygiene model popularised by Google's YouTube creator guidance for brands in the mid-2010s (hygiene is sometimes called "help"): hygiene content answers what the audience is searching for and is always available, hub content is the regular programming that brings people back, and hero content is the occasional big moment designed for reach. Most of a calendar is hygiene and hub; heroes are planned around them, not instead of them. Paid amplification of the best organic content belongs to Module 6; the content plan should simply flag which pieces are candidates.

Workflow is what makes the cadence real. A simple workflow names each stage (idea, brief, production, edit, approval, scheduling, community management), who owns it and how long it takes, and works backwards from publishing dates. In a regulated category the approval stage carries real weight: in skincare, any statement about what a product does needs evidence, so the plan must leave time for claims to be checked by the people who hold that evidence, and the creative team needs a list of what can and cannot be said so that most content passes first time.

Test-and-learn replaces opinion with evidence. Choose one variable at a time — the hook style, the format, the length, the posting time — hold the rest as steady as you can, decide in advance what result would change your practice, and run the test long enough to see past the noise of individual posts. Organic results vary a great deal from post to post, so treat a single comparison as a clue rather than proof; the rigorous testing methods in Module 7 show what stronger evidence looks like. Finally, schedule a review: each quarter, compare each pillar against its success measure, and retire or rework what consistently fails. A plan that never drops anything is not being managed.

Ohanian (1990) built the standard measure of a spokesperson's source credibility from three dimensions: expertise, trustworthiness and attractiveness. The model was developed for celebrity endorsers, but it transfers well to influencers, with one addition. Lou and Yuan (2019), surveying social media users who followed at least one influencer, found that the informative value of an influencer's posts, together with the influencer's trustworthiness, attractiveness and similarity to followers, increased followers' trust in branded posts, and that trust in turn raised brand awareness and purchase intention. Similarity matters because influencers are, in their followers' eyes, people like them who happen to be followed, not distant stars.

The second mechanism is relational. Horton and Wohl (1956) described parasocial interaction: the one-sided sense of intimacy that audiences develop with media figures who address them directly. A creator who talks to camera every day, answers comments and shares her own routine invites exactly that. Followers feel they know her, so her recommendation lands more like advice from a friend than an advertisement. The relationship is an asset the creator has built over years, and every sponsored post draws on it.

That is where the tension lies. Followers value influencers for apparently intrinsic, non-commercial motives, yet brand partnerships pay them to speak. Audrezet, de Kerviler and Moulard (2020) studied how influencers manage this and identified two strategies: passionate authenticity, in which the creator only takes on brands that genuinely fit their own interests and voice, and transparent authenticity, in which they are open about the commercial arrangement and their honest view of the product. The implication for brands is that over-scripted content makes both strategies harder to sustain.

The persuasion knowledge model (Friestad and Wright, 1994) explains why. Once people recognise a message as an attempt to persuade them, they draw on what they know about persuasion tactics and discount, scrutinise or resist it. Influencer content works partly because it does not look like advertising; the moment it reads as a brand's script, persuasion knowledge activates and the borrowed credibility drains away. Disclosure, which the law requires and which Module 5 covers, makes the commercial nature explicit, so brands cannot rely on concealment. They have to choose creators whose credibility survives being seen as paid.

For Moorfield Botanics the implication is direct. A reactive-skin audience that trusts other users' routines more than brand claims will be most persuaded by creators who are, or credibly understand, people like them.

The creator economy is the ecosystem of independent creators, the platforms that distribute and monetise their work, and the brands, agencies and tools that pay them. Understanding its structure tells a brand what it is actually buying and from whom.

Creators are usually grouped into tiers by following. Campbell and Farrell (2020) distinguish celebrity, mega, macro, micro and nano influencers and argue that the tiers differ not only in reach but in what they offer a brand: smaller creators tend to offer closer audience relationships and more niche credibility, larger ones reach and cultural visibility. The thresholds, however, are conventions, not standards. Agencies, platforms and academics draw the lines in different places, and a "mid-tier" band between micro and macro is common in practice. Moorfield, for instance, treats 10,000 to 50,000 followers as micro and 50,000 to 150,000 as mid-tier. Always state the thresholds you are using.

Creators earn money in several ways: fees for sponsored content, affiliate commission on sales, platform revenue sharing and creator funds, subscriptions and tips, and their own products. The mix shapes their incentives. A creator who depends on affiliate income has reason to push conversion; one with their own product line may see a brand as a competitor; one who earns mainly from an engaged community will protect it from unconvincing sponsorships. Haenlein et al. (2020) note that influencer marketing has professionalised quickly, with creators operating as businesses and expecting to be treated as such.

Between brands and creators sit intermediaries: talent agencies that represent larger creators and negotiate their fees and exclusivity; influencer agencies that plan and run programmes for brands; and discovery and management platforms, including the platforms' own creator marketplaces. Intermediaries add speed and scale, but each takes a margin and each has incentives, such as favouring their own roster, that a brand should recognise.

The edges of the economy are expanding. Virtual influencers are computer-generated personas run by companies; they offer control and less risk of off-brand behaviour, but raise questions of authenticity and disclosure. At the other edge, employee advocates and customer advocates are people with small followings whose credibility comes from first-hand experience rather than reach. For a brand whose audience trusts other users' routines, customers who already post about the product may be a creator tier in their own right.

Before choosing creators, a brand must decide what influencers are for. There are four main roles. Awareness: putting the brand in front of people who do not know it, especially through recommendation-led feeds. Credibility: borrowing trust the brand cannot claim for itself, which matters most in categories with high perceived risk. Content production: creators make platform-native content faster, more cheaply and more convincingly than many brands, and that content can be reused. Conversion: driving sales directly through links and codes. Most programmes want more than one, but they should rank them, because the role decides the criteria. A credibility role puts audience fit and expertise first; an awareness role weighs reach and audience overlap with non-customers; a conversion role weighs audience purchase intent and trackability.

The role must follow from the strategy already set. Module 1 gave each platform a role and set objectives that ladder up to business goals; Module 2 turned listening into insight. An influencer strategy that contradicts either is a separate plan, not part of the strategy.

Three partnership models are common. A campaign model buys a burst of content around a moment such as a launch: efficient for reach, but every relationship starts cold and ends abruptly, which can look transactional to followers. An always-on model keeps a rolling roster of creators posting throughout the year, building familiarity and steady content. An ambassador model signs a small number of creators to longer relationships, often with product input or exclusivity; it builds the deepest credibility because repeated, consistent advocacy reads as genuine preference, but it costs more per creator and concentrates risk on a few people. Many brands combine them: a campaign for the moment, ambassadors for continuity.

Finally, a brand should ask whether influencers are the right tool at all. Paid social offers control over targeting, message and frequency and is easier to measure; owned content builds an asset the brand keeps. Influencers earn their place when the job needs someone else's voice, as when the audience discounts brand claims. Creator content can also feed paid social, but that depends on usage rights agreed in the contract, which belongs to Module 5.

For Moorfield, the key insight that people with reactive skin trust other users' routines more than brand claims points to credibility as the first role and TikTok discovery as the second, which is why Amara chooses a launch campaign plus always-on ambassadors for the Barrier Repair Serum.

Discovery and vetting are where most poor influencer decisions are made, often invisibly: a creator is found because someone in the team follows them, and vetted by glancing at a follower count. A disciplined process separates the two steps and records evidence for each.

Discovery should cast a wide net. Routes include searching hashtags and keywords on the target platforms, social listening (the creators already talking about the category or the brand), platform creator marketplaces, agencies and discovery tools, and the brand's own customers and followers. Fans who already use the product are often the most credible micro and nano creators available, and listening is the cheapest way to find them.

Vetting then tests each candidate against the role. Audience fit comes first: who follows the creator, by age, gender and especially location. A UK brand paying for an audience that is mostly overseas is paying for reach it cannot convert. The best evidence is platform audience data supplied by the creator or accessed through an approved tool, ideally first-party exports or live screen-shares rather than static screenshots, which are easy to edit.

Audience authenticity asks how much of the following and engagement is real. Warning signs include sudden follower spikes that do not match a viral post, engagement that is high but generic (emoji strings and "love this!" from accounts with no content of their own), engagement concentrated in the first minutes from the same accounts on every post, which suggests an engagement pod, and follower bases dominated by inactive or irrelevant accounts. Third-party audit tools estimate these patterns, but their methods are proprietary and their estimates differ, so treat them as prompts for judgement rather than verdicts.

Brand safety and values review means reading the creator's back catalogue, not just the last few posts: past controversies, claims they make (in skincare, medical claims are a particular risk), tone, and whether their values align with the brand's. The risk runs both ways; the creator is also judging whether the brand is safe for them.

Past partnerships show how often the creator posts sponsored content, whether they have recently worked with direct competitors, and how their audience responded. A feed dominated by sponsorships signals a creator whose recommendations followers may already discount.

For Moorfield, vetting Jas Kaur meant checking platform audience data (81% UK), reading a year of her content for claims and controversies, and reviewing her skincare partnerships.

Selection turns vetting into a decision. The method this course has used since Module 1 applies here too: choose criteria that follow from the role, weight them, score each candidate on a defined scale, write down why, and test whether the ranking survives a change of weights. A scorecard does not make the decision objective, but it makes the judgements visible, comparable and open to challenge, which is what an auditable process means.

The research gives good reason not to let follower count dominate. De Veirman, Cauberghe and Hudders (2017) found in experiments that Instagram influencers with more followers were seen as more popular and more likeable, but that for divergent products, a large following could reduce the brand's perceived uniqueness and so harm brand attitude. Wies, Bleier and Edeling (2023), using 802 Instagram campaigns with more than 1,700 influencers plus eye-tracking and experiments, found an inverted-U relationship between follower count and engagement with sponsored posts: engagement rises with following and then falls, because larger audiences signal a weaker relationship with each follower. The decline was weaker where creators had more freedom over content and where the brand was less familiar.

Leung et al. (2022), analysing 5,835 sponsored posts by 2,412 influencers on Weibo in October 2018, found that larger followings, more original content and more salient sponsor mentions made posts more effective, while new-product announcements made them less so; influencer activity, follower–brand fit and post positivity all had inverted-U effects, meaning moderate levels outperformed extremes. Two cautions apply. The data come from one platform in one country at one point in time, so generalising to UK TikTok in 2027 is a judgement, not a finding. And the finding on product launches is a warning for any launch campaign: followers are reluctant to vouch for something unknown, which is a further argument for credible creators who use the product before they post.

Taken together, the evidence argues for selecting on fit, authenticity and content rather than size, and for a tier mix rather than a single bet. A tier mix pairs a small number of larger creators, who provide reach and a visible anchor, with many smaller ones, who provide closer relationships, more varied content and spread risk. Budget follows the role: a credibility-led programme should not spend most of its fees on the largest name available.

Moorfield's launch mix is one macro creator, six mid-tier creators and 24 micro creators, a structure Amara justifies in the worked example.

A creator brief turns a selection decision into content. It must make sure the content serves the objective, says nothing untrue and meets the rules, while leaving the creator room to make something their audience will believe. Briefs fail more often on the second job than the first.

The research explains why. Martínez-López et al. (2020), studying followers' responses to sponsored Instagram posts, found that trust in the influencer was the main route to a post's credibility, that trust fell when a post looked commercially oriented, and that visible brand control made followers less willing to seek more information about the product. Audrezet, de Kerviler and Moulard (2020) showed that creators protect their authenticity by working with brands that fit their own voice and being open about the deal; a script makes both harder. Wies, Bleier and Edeling (2023), discussed in Module 4, found that engagement fell less for larger creators who had more freedom over content. The rule follows: control what you must, and nothing more.

What the brand must fix is short. One objective, taken from the influencer role already chosen. The audience, described by need. No more than three key messages, written as substance to convey, not lines to read. Mandatories: the disclosure label and where it goes, the product name, link or code, tags and deadlines. The claims that may and may not be made, with permitted wording given exactly, because a paraphrase can turn a lawful cosmetic claim into an unlawful medicinal one (see Lesson 5.4). Deliverables, dates and a short list of things to avoid.

Everything else belongs to the creator: format, hook, tone, setting, how the product fits their routine, and their honest view, including what it does not do. A section headed "Yours to decide" makes the boundary explicit. Campbell and Farrell (2020) argue that part of what brands buy is the creator's skill as a content producer; an over-written brief pays for that skill and then discards it. Co-creation goes further: ask lead creators how they would tell the story, and send product early enough that they have genuinely used it before posting.

Approvals need the same discipline. Agree in advance what the brand will review (facts, claims, disclosure, safety), what it will not (voice and style), how many rounds and how fast. Approval does not transfer responsibility: if an approved post breaks the rules, the brand is still answerable, as Lesson 5.3 explains.

For Moorfield Botanics, Amara's brief for the Barrier Repair Serum fixes the objective (credible discovery among reactive-skin audiences), three key messages and the claims Dr Siân Lloyd, Head of Product and Regulatory, has evidence for; format and tone belong to the creators.

The contract turns the brief into obligations both sides can rely on. As the selection module noted, creators now run businesses (Haenlein et al., 2020): many have agents, and they expect clear terms. Informal deals agreed by direct message are where most disputes about rights, payment and disclosure begin.

Deliverables come first: platforms, formats (feed post, short video, Stories, live), number, length, dates, and whether posts must stay live for a minimum period.

The fee structure should follow the creator's role. A flat fee gives certainty and leaves the creator free to be honest, which suits a credibility role. A performance fee (per click or sale) moves risk to the creator but pushes towards hard selling and pays nothing for the trust they lend. Affiliate commission through a unique link or code rewards conversion directly, and is still advertising that must be labelled. Hybrids combine a flat base with commission. Gifting is not free either: a product sent in hope of a post is an incentive, with consequences covered in Lesson 5.3.

Exclusivity should be narrow (sensitive-skin skincare, not "all beauty"), time-limited and paid for, because it removes income the creator would otherwise earn.

Usage rights are the term most often forgotten. In the UK the creator normally owns copyright in what they make, so a brand that wants to reuse it needs a licence stating where (reposts, paid ads, website, email, retailers), for how long, in which territories and with what edits. Running ads from or with a creator's handle, often called whitelisting and offered through tools such as Meta's partnership ads and TikTok's Spark Ads at the time of writing (2026), also needs the creator's account permissions. Paid usage is usually priced separately. Moorfield's contracts grant it 90 days of paid usage rights for amplification, on which the creator amplification in Module 6 depends.

Approval rights record the review process agreed in the brief. The disclosure obligation should be explicit, with the brand able to require corrections. A morality clause lets either side end the deal if the other's conduct seriously damages its reputation; making it mutual is fairer, because creators take risk from brands too. Payment terms fix when fees are paid (for example, part on signature, part on posting) and how quickly invoices are settled. Termination covers notice, payment for work done, and what happens to the content and licence when the deal ends.

Finally, the contract should provide for measurement: a unique code and tracked link per creator, and an agreement to share platform performance data, which Module 7 relies on.

Influencer content persuades partly because it does not look like advertising; the persuasion knowledge model (Friestad and Wright, 1994) predicts labels will trigger scepticism. Evans et al. (2017) found that clearer disclosure language on Instagram increased advertising recognition, which lowered brand attitudes and intention to share; Boerman, Willemsen and Van der Aa (2017) found that disclosure on a celebrity's Facebook post raised ad recognition and reduced intention to pass it on. Disclosure has a cost, but the law forbids blurring it, so the answer is the one Module 4 gave: choose creators whose credibility survives being seen as paid.

In the UK two regimes apply (as at October 2026). The CAP Code, enforced by the Advertising Standards Authority (ASA), requires marketing to be obviously identifiable (rule 2.1); influencer content is an ad when the brand paid for it, in money or in kind, and had some control over it, and affiliate content always is. Consumer law, enforced by the Competition and Markets Authority (CMA), goes wider: content must be disclosed whenever the creator received payment, a gift, a discount, commission or another incentive, whatever their following, even without brand control.

The joint ASA/CAP and CMA Influencers' guide to making clear that ads are ads (third edition, March 2023) sets the standard: a prominent label upfront, before anyone engages; "Ad" or "Advert" preferred over terms such as "sponsored", "gifted" or "collab"; every post and Story labelled; affiliate content identified as advertising. A brand tag, a code or a label buried in hashtags is not enough; platform "paid partnership" tools count only if the result is obvious. Check for a later edition: the ASA plans an update for new legislation.

That legislation is the Digital Markets, Competition and Consumers Act 2024, whose consumer provisions took effect on 6 April 2025. The CMA can now decide that consumer law has been broken and fine up to £300,000 or 10% of global turnover, whichever is higher, without going to court, and the Act bans fake reviews and concealed incentivised reviews outright. The CMA's Social media endorsements: guidance for content creators, updated on 3 September 2025, reflects the Act; brands must tell creators they pay or gift to label posts, and both can be held responsible.

In the US, the FTC's revised Endorsement Guides (16 CFR Part 255, 26 July 2023) require material connections to be disclosed "clearly and conspicuously": difficult to miss, easily understood, in the same medium as the endorsement and, on social media, unavoidable. Advertisers should brief, monitor and correct endorsers, and can be liable when endorsers fail. The substance matches the UK's; enforcement differs.

Moorfield takes the simplest compliant line: every creator post and Story opens with "Ad".

When a creator makes a claim in content a brand has paid for, it is the brand's claim. The CAP Code requires marketers to hold documentary evidence for objective claims before the ad appears (rule 3.7), and a creator's honest experience ("my skin has never felt calmer") is not evidence for a claim about what a product does. That is why permitted claims go into the brief word for word.

Cosmetics have their own claims law. In Great Britain the UK Cosmetics Regulation (the retained version of Regulation (EC) No 1223/2009) governs cosmetic products, and Regulation (EU) No 655/2013, still in force in Great Britain as assimilated law, sets six common criteria every cosmetic claim must meet: legal compliance, truthfulness, evidential support, honesty, fairness and informed decision-making. Evidence must match the claim: a study showing improved skin barrier function after 28 days supports exactly that, not "repairs damaged skin".

The sharpest line is between cosmetic and medicinal claims. Under the CAP Code (rule 12.1), claims to prevent, treat or cure disease may only be made for licensed medicines or appropriately marked medical devices. Eczema is a medical condition, so "heals eczema" cannot be said of a cosmetic; the ASA advises checking classification with the medicines regulator before any eczema claim. For Moorfield, Dr Siân Lloyd confirms the company holds the evidence for "clinically shown to improve skin barrier function in 28 days"; "heals eczema", or any medicinal claim, is forbidden.

Some sectors add rules. Since 5 January 2026 paid online advertising of identifiable less healthy food and drink has been banned in the UK at any time, including paid creator content, subject to exemptions for brand-only ads and businesses under 250 employees. Since October 2022 the CAP Code has prohibited gambling ads likely to have strong appeal to under-18s, ruling out many creators popular with young people. The FCA's guidance on financial promotions on social media (FG24/1, March 2024) warns that influencers promoting regulated products without an authorised firm's approval may commit a criminal offence, and that firms remain responsible for promotions they cause to be made.

Protecting children runs through all of this. The CAP Code has specific rules for ads aimed at or featuring children; check audience age data, as vetting in Module 4 recommended, and take particular care with creators who are under 18.

Brand safety runs both ways: the brand needs the creator's content to stay consistent with its values, and the creator needs a brand that will not embarrass them. Sign-off should be a named process: who checks each claim against the evidence, who checks sector rules and audience age, who reviews brand safety, and who can stop a post.

A launch campaign ends; a reputation with creators does not. Module 4 set out the partnership models; this lesson is about how a brand behaves within them, which decides whether good creators want to work with it again.

Ambassador programmes turn the best campaign partnerships into longer relationships. Repeated, consistent advocacy reads as genuine preference, supporting the passionate authenticity Audrezet, de Kerviler and Moulard (2020) describe, and lets creators build the product into their content naturally. A programme needs a clear design: who is selected and on what evidence (audience response, content quality, reliability and compliance record, not results alone); how many and for how long; what they receive (a retainer or fee per deliverable, early access, product input); what is asked of them; how the relationship is reviewed; and how it ends. Ambassadors concentrate risk on fewer people, so review points matter.

Paying fairly is both ethical and practical. Duffy (2017) showed how much unpaid "aspirational labour" goes into building a creator career; brands that substitute gifting for fees, pay late or demand unpaid revisions draw on that labour for free. Fair practice is simple: pay a fee for work; set rates by tier and deliverable rather than negotiating stamina, so creators doing the same work are paid the same; pay separately for usage rights and exclusivity; and pay on the date promised.

Feedback makes the relationship two-way. Share results with creators, explain what worked, and ask what the brand could do better: unclear briefs, slow approvals, late product.

Creator wellbeing deserves explicit attention. Sponsored posts can attract abuse, especially when creators talk about their own skin or health. A brand can set realistic deadlines, help moderate hostile comments under campaign content and avoid asking for more personal disclosure than the creator wishes to give.

Things will still go wrong, and a plan made in advance keeps the response proportionate. A post missing its "Ad" label should be found and corrected the same day; UK and US regulators both expect brands to monitor and correct, and Stories, which disappear after 24 hours, need a process for capturing and checking them. Disappointing results are a lesson about the brief or the fit, not a reason to punish an honest creator. A creator controversy calls for establishing the facts, pausing content where needed, acting on the contract and communicating calmly. If the brand is at fault, its creators should hear it from the brand first.

For Moorfield Botanics, the launch is followed by always-on ambassadors, chosen from launch creators whose audiences responded and who were reliable to work with, and offered a longer relationship with product input rather than one-off posts.

Organic social was never free, but for most brands it is now a weak route to new audiences. As Module 1 showed, recommendation-led feeds distribute content on predicted interest rather than on who follows whom, so organic posts reach mainly people already engaged with the account. That is valuable for community and for testing ideas, but unreliable for reaching a defined audience at a defined time. Paid social buys that reliability: delivery to people the brand chooses, in placements it chooses, as often as it chooses. In the paid, earned, shared and owned (PESO) model (Dietrich, 2014), paid media is the part a brand can switch on, scale and schedule.

What a brand should buy with it depends on how brands grow. Sharp (2010) argues that brands grow mainly by recruiting and refreshing large numbers of light buyers rather than by deepening loyalty among a few heavy ones, and that this depends on mental availability (being easily thought of in buying situations) and physical availability (being easy to find and buy). For paid social, reach among category buyers who do not yet know the brand matters more than repeated exposure to existing followers. Frequency still has a role, since one exposure in a fast-scrolling feed may not register, but Krugman (1972) argued that a few exposures can be enough for an advertisement to do its work, which is a reason to cap frequency rather than maximise it. The planning question is the media planner's: how many of the right people, how often, over what period.

Binet and Field (2013) add the distinction between long-term brand-building and short-term sales activation, introduced in Module 1. Paid social can do either, but platforms make activation easy to buy and easy to report, so budgets drift towards it, and a plan made only of conversion campaigns aimed at warm audiences tends to harvest demand created elsewhere rather than create it.

Paid social is the wrong lever when the product, landing page or checkout cannot convert the attention it buys; when the job needs someone else's credibility, the influencers' role set in Module 4; when creative has not yet been shown to stop the right people; and when the margin cannot support the cost of acquiring a customer, a calculation Lesson 6.3 makes explicit. Amplifying a weak idea only makes it fail faster.

For Moorfield, organic Instagram and TikTok reach mainly existing followers, but the FY2027 goal is to grow new customers from 48,000 to 60,000. Paid social, with £260,000 of the £720,000 digital budget, is the line that can reach people with reactive skin who have never heard of the brand.

Every campaign starts with an objective: the advertiser tells the platform what to optimise for (reach, video views, traffic, leads or purchases), and the delivery system seeks the people most likely to take that action. Purchase optimisation finds people likely to buy, some of whom would have bought anyway, so platform-reported conversions can overstate what advertising caused. Gordon et al. (2019), comparing large randomised experiments at Facebook with common observational methods, found the latter often misestimated the true effect, sometimes widely. Choose the optimisation event that matches the job and expect conversion-optimised reporting to flatter itself; Module 7 covers the tests.

The signals behind audience strategy have weakened. Apple's App Tracking Transparency, introduced with iOS 14.5 on 26 April 2021, requires apps to ask permission before tracking users across other companies' apps and websites. Where users decline, platforms lose data behind detailed targeting, retargeting, lookalike audiences and conversion reporting, and lean on modelled results and data advertisers send from their own servers. Targeted audiences shrink and measurement becomes less certain; size the effect only from a brand's own data.

The platforms' answer is automation: broad targeting, often only country, age and exclusions, letting the system find buyers. As at 2025–26, Meta's automated suite is branded Advantage+ (Advantage+ shopping campaigns became Advantage+ sales campaigns in early 2025) and TikTok's is Smart+; names change often. Broad delivery trades control and transparency for scale and cost, putting the weight of targeting on creative.

In a broad audience the ad itself decides who stops scrolling. A hook that names a reactive-skin problem in the viewer's own words selects people with that problem; a generic beauty shot selects no one. Distinct hooks built on the pillars from Module 3 do the segmentation that targeting settings used to do. Specificity should match the decision stage: Lambrecht and Tucker (2013) found dynamic retargeted ads showing products people had browsed were on average less effective than generic ads, unless browsing showed that preferences had narrowed.

Test with structure: a hypothesis, one changed variable (hook, format, message or creator), a fair budget and time per variant, and a decision rule set in advance, because early results are noisy and delivery systems shift spend to early winners.

Creator amplification runs creators' content as ads: Meta's partnership ads run from the creator's handle, or both identities, labelled "Paid partnership"; TikTok's Spark Ads boost an organic post through an authorisation code the creator generates for a limited period. Both need the creator's permission and a contract granting paid usage rights (Module 5): the platform switch is not the licence, and amplified content is still advertising that must be disclosed. Moorfield's 90 days of paid usage covers the six-week launch.

Every paid social budget rests on a chain of assumptions; making it explicit is the best defence against plans that cannot work. Spend buys impressions at a cost per thousand (CPM): impressions = budget ÷ CPM × 1,000. A share of impressions become clicks (click-through rate, CTR), a share of clicks become orders (conversion rate) and each order is worth the average order value (AOV). These yield revenue, return on ad spend (ROAS = revenue ÷ spend) and cost per acquisition (CPA = spend ÷ orders).

ROAS leads most dashboards and is the most often misread, because it ignores what it costs to make and deliver the product. The business needs to know whether the media paid for itself out of margin. Gross profit is revenue × gross margin; contribution after media cost is gross profit minus spend. Two break-even points follow. Break-even ROAS is 1 ÷ gross margin: at a 68% margin, about 1.47x. Break-even CPA is AOV × gross margin: at £42 and 68%, about £28.56. A ROAS of 1.3x is "positive", since revenue exceeds spend, yet at a 68% margin it loses money on every order. If the margin figure excludes fulfilment, payment fees and returns, true contribution is lower still.

The assumptions are fragile. CPMs vary by platform, placement, audience, season and competition; click-through and conversion rates with creative, offer, landing page and product. Published benchmarks differ widely by source, sample and date: use them as a sense check, not a forecast, and prefer your own data and small tests. The model is also linear, while campaigns are not: as spend rises, delivery reaches less responsive people and frequency climbs, so returns usually diminish. And platform-attributed orders are not orders the advertising caused, a gap Module 7 quantifies.

Sensitivity analysis shows which assumption the decision depends on: in a tightly balanced model, a small change in one rate turns profit into loss. Change one input at a time and note how far each must move to reach break-even; the inputs with the smallest margin for error are the ones to test first and to monitor closely in flight.

First-order profit is not the whole case. An acquisition campaign can justify a thin or negative first-order contribution if the customers it recruits come back, but that is a claim about future behaviour, to be stated as an assumption and tested rather than used to excuse any result. Binet and Field (2013) also remind planners that brand-building effects arrive over longer periods than a campaign report covers.

Amara's model for the £60,000 launch budget shows all of this in one line: a ROAS of 1.58x, "positive", leaves just £4,260 of contribution after media cost.

A campaign is integrated when its channels work as one plan rather than run in parallel. Schultz, Tannenbaum and Lauterborn (1993) argued that communication should be planned from the customer outwards, with every contact coordinated, rather than bought channel by channel, an early statement of integrated marketing communications (IMC). Integration is often mistaken for consistency; shared cues do build memory, but the stronger idea is complementarity: each channel does the job it is best at, and the channels reinforce one another. Batra and Keller (2016), reviewing how traditional and digital media combine to affect consumer decisions, argue that plans should be designed around where consumers are in their decision journey and how media effects interact.

Practical integration rests on four shared elements: one idea, which every channel expresses natively; one audience definition; one calendar, so activity is sequenced rather than colliding; and one measurement plan, agreed before launch, so that each channel is judged on its role and the same sale is not counted several times by different platforms.

Two frameworks help assign roles. Chaffey's RACE framework (Chaffey, Ellis-Chadwick and Abed-Rabbo, 2025) organises digital activity around the customer lifecycle: Reach (build awareness and visits), Act (prompt interaction and leads), Convert (turn interest into sales, online or offline) and Engage (build long-term loyalty), often preceded by a Plan stage. Kaushik's (2013) See-Think-Do framework groups audiences by intent: See is the largest addressable audience; Think, those with some commercial intent; Do, those ready to buy now; a later fourth stage, Care, covers existing customers. RACE suits planning objectives by stage; See-Think-Do-Care is a sharper check that content and targeting match intent, and it exposes plans that spend everything on Do.

In a social-led campaign, organic social sustains community and tests ideas cheaply. Influencers supply credibility and native content the brand cannot make for itself. Paid social buys reach at scale and amplifies proven content. Search, especially paid search on brand and problem terms, captures demand the other channels create; it rarely creates it. Email reaches people who have given permission, cheaply. Owned community and the website hold the relationship and the sale. This module treats search and email only in these roles.

Integration fails in known ways: silos with separate budgets and targets, email going out before the creators have posted, platforms each claiming the same order, and messages drifting as each team adapts them. The cure is as much organisational as creative: one owner of the plan and one channel-role map everyone works from.

For Moorfield, the idea is "Strong Skin Season", the audience is people with sensitive and reactive skin, and 1 March to 11 April 2027 is the shared calendar; the worked example shows Amara's channel-role map.

Plans fail in execution more often than conception: assets arrive late, claims go unchecked, budget sits in a weak ad because no one knew who could move it, and a comment thread goes wrong on a Saturday. An execution plan settles, before launch, who does what, by when, and what happens when things change.

The launch plan works backwards from go-live, listing every asset and task with an owner and deadline, and identifies the critical path: the dependent tasks that set the earliest possible launch. In social and influencer campaigns it usually runs through creator content, claims approval and platform ad review, all slower than expected, so build buffer before launch.

A RACI matrix names, for each task, who is Responsible (does the work), Accountable (owns the outcome and signs off; one person only), Consulted before a decision and Informed after it, exposing two common faults: tasks with no accountable owner, and tasks with so many people consulted that nothing moves.

In skincare, every ad, caption and creator script describing what a product does must be checked against the claims the brand can substantiate, and paid ads face the same advertising codes as any other ad. Agree which claims are pre-approved, who signs off anything new and the turnaround time, so approval is a scheduled step, not a bottleneck.

In flight, a campaign needs a small dashboard and pre-agreed rules. The dashboard tracks leading indicators for each channel's role (reach and frequency for paid social, creator posts live and compliant, branded search volume, email clicks, landing-page conversion) against the plan's assumptions. Optimisation rules set in advance what triggers a change (a creative losing budget, spend shifting between platforms, creative rotating as frequency rises), guarding against two opposite errors: reacting to daily noise, especially during delivery systems' initial learning period, and leaving money in something clearly failing. Base them on the economics of Lesson 6.3, such as CPA against break-even CPA, remembering that platform attribution overstates causal effects.

A risk register lists each risk with its likelihood, impact, owner, preventive action and response. Typical launch risks include a creator post that breaks disclosure or claims rules, rejected ads, stock-outs, site failure under traffic, and criticism or misinformation in comments. Community management is part of execution: a launch brings questions, complaints and claims into comments, so the team needs staffed hours, escalation routes and approved answers, including for anyone asking whether a product treats a medical condition. The moderation principles in Module 2 apply with extra force while ads drive traffic.

For Moorfield, the partnerships audit found two gaps, no process for checking micro creators' Stories and no written affiliate-code policy, and both belong on the launch risk register until closed.

A measurement plan starts with what the activity is for, not with what the platforms can count. Hoffman and Fodor (2010) argued that social media return is usually measured the wrong way round: brands count what they spend and what platforms report, when they should start from why customers use social media and measure the investments customers make in the brand (attention, conversation, content, advocacy) before tracing those to sales or cost savings. The point survives: the metrics a dashboard offers by default are rarely the ones an objective needs.

The communication measurement profession has codified the same discipline. The Barcelona Principles, from AMEC, the International Association for the Measurement and Evaluation of Communication, updated as version 4.0 in June 2025, open with setting clear and measurable objectives, and call for reporting outputs, outcomes and impact, using qualitative as well as quantitative analysis across all relevant channels (AMEC, 2025). AMEC's Integrated Evaluation Framework, launched in 2016, turns this into stages from objectives: inputs (audience, plan, budget), activities (what was done), outputs (what audiences were exposed to: posts published, reach), outtakes (how they responded: attention, engagement, clicks), outcomes (what changed in them: understanding, preference, intent, trial) and impact (what changed for the organisation: sales, customers, reputation). Most social reports stop at outputs and outtakes; budget decisions turn on outcomes and impact.

Each stage needs a metric, a data source, a baseline and a reading date. Outputs and outtakes are leading indicators: available within days and useful for managing a campaign in flight. Outcomes and impact are lagging indicators: slower, noisier and harder to attribute, but they are what the objective is about. Each target waits for a baseline, as Module 1 insisted.

Time horizons matter too. Binet and Field (2013) distinguish short-term sales activation, whose effects show within weeks, from long-term brand-building, whose effects accumulate over months and years. A six-week campaign report can see the first clearly and the second barely at all, so a framework should say which effects it will read when, and which it cannot see.

For Moorfield Botanics, Amara returns to the objectives set at her platform audit: new customers whose first recorded touchpoint is TikTok or Instagram; reach among people outside the customer base; and growth in Instagram saves, shares and replies. For the Strong Skin Season launch she maps creator posts live and reach as outputs; saves, shares and link clicks as outtakes; branded search and visits from new visitors as outcomes; and first orders, new customers and their contribution to the FY2027 goal of 60,000 new customers as impact, with repeat purchase read later.

Most social metrics are easy to collect, hard to interpret and narrower than their names suggest.

Reach counts the unique accounts that saw content; impressions count every view, so one person seeing a post three times is one reach and three impressions. Both are platform estimates.

Engagement rate has no single definition. It can be engagements divided by followers, by reach or by impressions, and "engagements" may include likes, comments, shares, saves, clicks or video views depending on the tool, so one post can show three different rates. Rate by followers flatters content shown mainly to followers and misreads content that travels beyond them in recommendation-led feeds; rate by reach measures how the people who saw it responded, usually the more useful question. This module uses engagements (likes, comments, shares and saves) divided by reach; any report should state its definition. Even then, engagement is an outtake, not an outcome: a like costs the user nothing, and saves and shares signal more.

Sentiment analysis classifies mentions as positive, negative or neutral, usually automatically. As Module 2 showed, automated classification varies widely in accuracy and struggles with sarcasm and slang (Hartmann et al., 2023), so trends are more trustworthy than single scores, and samples should be read by hand. Share of voice, a brand's share of category conversation, is useful as a relative measure over time but depends entirely on which competitors and keywords are tracked. Follower growth is a vanity metric when followers are not the objective: in recommendation-led feeds following is a weak predictor of who sees content, and followers can be bought.

Earned media value (EMV) is the most seductive figure of all. Vendors price engagement or impressions at what equivalent advertising might have cost, then add it up. EMV descends from the advertising value equivalent (AVE), long used to value press coverage, and the Barcelona Principles 4.0 (AMEC, 2025) exclude AVEs as an invalid measure. EMV formulas differ by vendor and are rarely disclosed, the rates are assumptions, and the result measures neither sales nor profit nor what would have happened without the activity.

Finally, ask who produced each figure. Platform-reported results come from the company selling the media, under its own attribution rules; agency reports come from the company paid to run the activity. Neither is necessarily wrong, but both are marking their own homework. Independent evidence (the brand's own orders, site analytics, customer surveys and experiments) should carry more weight in any budget decision.

For Moorfield, the agency's launch report credits the influencer programme with £410,000 of EMV against a cost of £100,000. Amara sets it aside: it is not revenue, and it cannot say whether a single extra serum was sold.

Attribution assigns credit for a sale to the marketing touchpoints that preceded it. It is indispensable for running campaigns and dangerous when mistaken for a measure of what caused the sale.

Influencer activity is usually tracked in two ways. Tracked links carry UTM parameters, tags that tell web analytics which source, medium and campaign sent a visitor; unique discount codes tie an order to the creator whose code was used, even when the buyer never clicked. Moorfield set up both per creator in Module 5. Both miss and over-count. They miss buyers who saw a creator's video, later searched for the brand and bought without a code; much influence leaves no click. They over-count because codes leak: shared in group chats, posted on voucher sites and found at checkout by people who were already buying. A leaked code records a discount taken, not a sale created.

Attribution models decide how credit is split when a buyer has several recorded touches. Last-click gives it all to the final touch, favouring channels closest to the sale (brand search, email, retargeting) and starving those that created the demand. Rules-based multi-touch models (first-click, linear, time decay, position-based) spread credit by fixed rules; data-driven models assign it statistically. Google Analytics 4 retired its rules-based multi-touch models during 2023, leaving data-driven and last-click, a reminder that the available models change. Every model shares credit among recorded touches; none says whether the sale would have happened without them. Blake, Nosko and Tadelis (2015) showed the gap at eBay: when brand-keyword search ads were switched off, most of the traffic simply arrived through unpaid search results instead, so much of the attributed return was illusory.

Platforms add their own layer. Each reports the conversions it claims, including view-through conversions from people who saw an ad but did not click, under its own attribution window, so Meta, TikTok and Google can each claim the same order. Since App Tracking Transparency, as Module 6 explained, more of those conversions are modelled rather than observed.

Marketing mix modelling (MMM) works from the other end: regression on aggregate time series of sales, channel spend, price, seasonality and other factors estimates each channel's contribution without tracking individuals. It needs a long history and enough variation in spend, and its estimates rest on the model's assumptions, but it is unaffected by tracking loss. Open-source tools have lowered the cost: Meta's Robyn, and Google's Meridian, made available to all advertisers in early 2025. MMM is strongest when calibrated with experiments, the subject of Lesson 7.4. A cheap complement is a post-purchase survey asking "How did you first hear about us?".

The honest position is a chosen method with its limits stated.

Incrementality is the part of an outcome that would not have happened without the marketing activity. It cannot be observed directly, because no one can see what the same customers would have done without the campaign: the counterfactual. Experiments estimate it by comparing an exposed group with a comparable unexposed one.

Comparing people who saw ads with people who did not misleads, because exposure is not random. Delivery systems show ads to people predicted to buy, and creators' followers are already interested in the category, so exposed people would have bought more anyway. Gordon et al. (2019) tested this with 15 large US advertising experiments at Facebook: common observational methods, even with rich data on users, often failed to reproduce the experimental results and tended to overstate the effect of advertising. Randomisation is the reliable way to break the link between who is exposed and who would have bought.

Three designs are common. A user-level randomised holdout withholds the activity from a random share of the target audience; ghost-ad methods (Johnson, Lewis and Nubbemeyer, 2017) record which held-out users would have been shown the ad, sharpening the comparison. A geo holdout test runs activity in some regions and withholds it from matched regions; it needs no user tracking, suits activity that cannot be randomised person by person, and works with regional sales data (Vaver and Koehler, 2011). Platform lift studies, offered at the time of writing (2026) by platforms including Meta and TikTok, run randomised holdouts inside the platform to measure brand lift (survey-measured awareness or consideration) or conversion lift, subject to eligibility and spend thresholds that change.

Each has limits. Platform lift studies measure only that platform's ads, on its own data. Geo tests need regions with enough sales and similar trends, and spillover (people travel; creators' audiences are national) shrinks the measured effect. Organic creator posts cannot be withheld from a region at all; only paid amplification can. And sales are noisy: Lewis and Rao (2015), examining 25 large field experiments, found the median confidence interval on return on investment was more than 100 percentage points wide. An underpowered test wastes money.

A small brand can still test well: test one decision at a time; choose activity that can be switched off (amplification, a channel, a creator tier); run it long enough to see past weekly noise; fix the decision rule in advance; and report a range, not a point.

Amara's geo holdout withholds creator amplification from matched UK regions during the launch, and she checks how far creator codes have leaked to voucher sites. Together they suggest that 60% of attributed orders were incremental.

Return on investment for influencer marketing is a calculation anyone can do and few do honestly. Four choices decide the answer.

First, cost. The full cost includes creator fees, paid amplification of creator content, product seeding and shipping, and ideally agency, tool and staff time. Leaving out amplification because it sits in a paid social budget is the commonest way to flatter influencer returns.

Second, which orders count. Attributed orders carry a creator's code or link; incremental orders are the share an experiment suggests would not otherwise have happened. Report both, because the gap between them is the finding.

Third, revenue or profit. Return on ad spend (ROAS, revenue ÷ cost) ignores the cost of the goods sold. Return on spend should use gross profit: (revenue × gross margin − cost) ÷ cost. As Module 6 showed, at a 68% margin the break-even ROAS is 1 ÷ 0.68, about 1.47x, so a ROAS above 1 can still lose money.

Fourth, time horizon. A first-order calculation ignores repeat purchases, which may be much of a new customer's value in a replenished product such as skincare, and the longer-term brand effects that Binet and Field (2013) argue build over time. Lifetime value claims are forecasts, not results: state the assumption and measure it rather than invoking it to rescue a weak first-order result. Halo effects, such as extra sales through retailers or of other products, are real but should be estimated, not assumed.

Reporting is where honesty is tested. A one-page evaluation for senior stakeholders states the objective, the full cost, attributed and incremental results side by side, how the incremental estimate was made and how uncertain it is, what is not yet known, and a recommendation with the decision it asks for. Lead with the result, not the activity list. Do not claim precision the data cannot support: a range with its assumptions is more honest than one exact percentage. And keep EMV out of the return calculation entirely.

Senior readers also need the comparison that makes a number meaningful: against break-even, against the plan, and against the next best use of the money, such as the same sum in paid social or email.

For Moorfield, the influencer launch cost £100,000. On attributed orders it returned +17.1% on spend; on the 60% of orders estimated to be incremental, −29.7%. The same campaign is a success or a loss depending on which question is asked, and Amara's job is to tell Tom Hargreaves which question matters.

A strategy document is not the pile of worksheets that came before it. Rumelt (2011) argues that the kernel of a good strategy has three parts: a diagnosis that names the critical challenge, a guiding policy for dealing with it, and coherent actions that carry the policy out. Plans most often fail at the first step. They open with an ambition ("become the sensitive-skin brand everyone finds on TikTok") and never say what stands in the way. This lesson writes the diagnosis; the rest of the capstone writes the policy and the actions.

The diagnosis starts from the business, not from social: which business goals the strategy serves, over what period, and what is stopping the organisation reaching them. For Moorfield Botanics the evidence of the earlier modules points somewhere uncomfortable. The Strong Skin Season influencer programme returned +17.1% on spend on attributed orders but −29.7% on the 60% of orders estimated to be incremental (Module 7). A strategy for the following year that ignores that gap is not a strategy; one built around closing it may be.

The situation summary gives the reader the facts the diagnosis rests on: platform roles and what each is delivering, the community, the last campaign's results on both bases, the budget and the team. State what is not yet known as plainly as what is. A strategy written in November cannot know the full-year result; it should say when that number will arrive and what it would change, rather than estimate it.

Audience insight comes from Module 2: two or three insights, each with its evidence and its limits. Objectives then follow the cascade Module 1 set out. Every social objective must be traceable to a business goal and labelled brand-building, activation or community, because Binet and Field (2013) found that brand-building and activation work over different time horizons and need different measures. Where the baseline is a number the business has not yet closed, the target waits for it.

The strategy core makes four choices: where to be, how to hold a relationship with the audience, what to say, and who else says it. Each was a module in its own right; the capstone's job is to make them one argument. Porter (1996) argues that a strategy is defensible less because any single activity is excellent than because the activities fit and reinforce one another. A platform portfolio chosen for discovery, pillars chosen for saving and sharing, and creators chosen for reach alone could each pass their own module's test and still fail as a system.

Start with the platform portfolio from Module 1. Re-run it with a year's evidence: has a platform's role changed, has the team closed a content-capability gap, is there a platform to drop? Then the community: is it doing what Schau, Muñiz and Arnould (2009) describe, with members creating value for one another, or has it become a broadcast list with a membership badge? Then the content pillars from Module 3, each tied to an insight and a measure, with any pillar that rested on weak evidence now confirmed or cut.

The influencer model is where earlier results should bite hardest. Module 4 chose a tier mix on weighted criteria rather than follower counts; Module 5 showed why repeated, consistent advocacy reads as genuine preference, supporting the passionate authenticity Audrezet, de Kerviler and Moulard (2020) describe, and how an ambassador programme should be designed. A strategy says which model it keeps, which tiers it cuts and what evidence would make it change course.

Under every choice, write one evidence line: the insight, citation or result it rests on, and what that evidence cannot show. The honeycomb (Kietzmann et al., 2011) is a reason to build around conversation and reputation for a sensitive-skin audience; a single launch evaluation is a reason to change the influencer mix, but not proof that the new mix will work. A leadership team trusts a strategy that knows the difference.

Integration is where good strategies are most often lost. Module 6 defined it as complementarity rather than mere consistency: each channel does the job it is best at, and the channels reinforce one another across the decision journey (Batra and Keller, 2016). Over a year rather than a six-week launch, that means a channel-role map using RACE or See-Think-Do-Care for organic social, influencers, paid social, search, email and the owned community, and an agreed rule that the same sale is not counted by several platforms.

The calendar turns roles into time. A year usually combines always-on activity, which sustains community, content and ambassadors, with a small number of campaign peaks in which creators, paid amplification, search and email concentrate on one idea. Place each peak for a reason (a launch, a seasonal need, a moment the audience cares about) and plan backwards from it: creator content, claims sign-off and platform ad review are the slowest steps, as the critical path in Module 6's launch plan showed.

The budget is the strategy's sharpest statement, because it shows what the organisation is actually prepared to do. Start from last year's lines, show each change and the evidence for it, and make the lines add up to the total exactly. Two tests apply. Does every channel have enough to do its job well, rather than a little everywhere? And is measurement funded? A strategy that promises experiments but leaves no budget or audience for holdouts has not committed to them.

Governance names who decides. A RACI for the main decisions (moving budget between peaks, adding or exiting a creator, approving a new claim, pausing a post) prevents both orphaned decisions and decisions that need everybody. Felix, Rauschnabel and Hinsch (2017) treat structure and governance as strategic choices in social media marketing, not administration.

Compliance belongs inside the plan, not in an appendix. Carry forward the disclosure policy, permitted and forbidden claims, usage rights and monitoring from Module 5, checked against the rules in force when you write: the ASA/CAP and CMA influencers' guide (third edition, March 2023), the CMA's social media endorsements guidance for creators and brands (updated August–September 2025) under the Digital Markets, Competition and Consumers Act 2024 and, for US activity, the FTC Endorsement Guides (revised July 2023). A year-round ambassador programme multiplies the posts, Stories and codes that need checking, so scale the monitoring with it. None of this is legal advice; take it for anything novel.

A strategy asks for money; the measurement plan says how anyone will know whether it was well spent. Build it as Module 7 did: start from the objectives, then map outputs, outtakes, outcomes and impact for each, with a data source, a baseline and a reading date (AMEC, 2025). Separate the leading indicators that manage the year in flight from the lagging ones that judge it.

Then decide where the strategy needs causal evidence rather than attribution. Gordon et al. (2019) found that common observational methods often failed to reproduce experimental results and tended to overstate the effect of advertising, so any large or contested spend line should carry an incrementality test: a geo holdout for paid creator amplification, a platform lift study for a paid social flight, a randomised holdout for an email programme. Lewis and Rao (2015) are the warning about power: advertising experiments need large samples and enough time, so test fewer things properly. Write the decision rule before the results arrive: which result would increase, hold or cut the line.

The risk register lists what could knock the plan off course: a disclosure or claims breach, a creator controversy, a platform change that cuts organic reach, an experiment too small to detect an effect, a mid-year budget cut. Each risk needs a likelihood, impact, owner, preventive action and response. The roadmap phases the year with review points at which evidence can change the plan, which is what turns a document into a management tool.

Finally, the executive summary. A leadership team reads it first and sometimes only. In about a page it states the problem, the strategy in a few sentences, the budget asked for and how it differs from last year, how success will be judged and what remains uncertain. Lead with the decision you want, not the activity list. Hoffman and Fodor (2010) argued that social media return should be measured from the investments customers make in the brand, not only from what the brand spends; a summary that reports both honestly is harder to dismiss than one that promises a return it cannot support.