Difficulty: Intermediate | Prerequisites: Basic understanding of social media platforms (Twitter/X, Facebook, YouTube, Instagram)
Social media analytics sits at the intersection of marketing strategy, data analysis, and digital communication. This assignment (CO6) tests your ability to move beyond casual social media use and examine platforms through an analytical lens: how public opinion forms and spreads, how brands build and measure their presence, and what risks organisations face when they operate in these spaces. You should already be comfortable with the idea that social media is not just a broadcast channel but a two-way interaction system with measurable outcomes.
This material covers three core competencies: analysing public opinion and detecting manipulation on Twitter/X, auditing brand social media strategies using engagement metrics and interaction types, and evaluating the risks organisations face on social platforms. If you can identify bot-like behaviour, measure a brand's social media effectiveness, and articulate social media risks, you have the essentials.
Sentiment analysis
The computational process of identifying and categorising opinions expressed in text to determine whether the writer's attitude toward a topic is positive, negative, or neutral. In simple terms, it is software figuring out whether a tweet is "for" or "against" something.
Bot account / astroturfing
A bot account is an automated social media profile that posts, retweets, or likes content without direct human involvement. Astroturfing is the practice of using multiple bot or fake accounts to manufacture the appearance of grassroots public support. Think of it as a fake crowd, a hundred accounts all cheering the same thing to make it look popular.
Engagement metrics
Quantitative measures of how audiences interact with social media content: likes, shares, retweets, comments, click-through rates, follower counts, and impressions. In simple terms, the numbers that tell you whether anyone is paying attention.
Broadcast interaction
A one-way communication model where a brand publishes content to its audience without soliciting or responding to feedback. Company announcements and press releases are typical examples. Think of it as a megaphone: the brand talks, the audience listens.
Request for input
A social media interaction where the brand actively solicits audience responses, such as polls, questions, or calls for user-generated content. In simple terms, the brand asking "What do you think?" and meaning it.
Direct interaction
Two-way, personalised communication between a brand and an individual user, typically in reply threads, direct messages, or customer service exchanges. Think of it as a conversation rather than a broadcast.
Social media risk
Any threat to an organisation arising from its social media presence or the broader social media environment. This includes reputational damage, data breaches, employee misconduct online, regulatory non-compliance, and crisis escalation. In simple terms, the ways social media can hurt a company rather than help it.
Content repetition / coordinated inauthentic behaviour
A pattern where identical or near-identical posts appear across many accounts in a short time frame, often indicating automated or organised manipulation campaigns. Think of it as copy-paste at scale, a red flag that real people are probably not behind it.
Tags: Twitter opinion mining, sentiment analysis, bot detection, coordinated inauthentic behaviour, hashtag analysis, echo chambers
Social media platforms are real-time barometers of public sentiment on political issues, bills, elections, and social debates
Understanding how opinion forms, clusters, and spreads is a foundational skill in social media analytics
Platforms like Twitter/X surface opinion through hashtags, trending topics, and reply threads, making them natural laboratories for opinion research
Choose a current issue: Select a live public debate (a bill under consideration, an election, a political controversy) with active discussion
Search and collect: Use the platform's search, hashtag tracking, and advanced filters (date range, language, location) to gather a representative sample of posts
Summarise the opinion landscape: Group posts by stance (for, against, neutral, nuanced) and identify the dominant themes within each group
Assess dominance: Determine whether one opinion clearly outweighs the others in volume, engagement, or visibility. A dominant opinion may reflect genuine public sentiment, or it may reflect algorithmic amplification or coordinated campaigns
Repeated content is a signal: When the same tweet (or near-identical phrasing) appears across many accounts, this is a red flag for coordinated inauthentic behaviour
Distinguishing legitimate repetition from manipulation:
Legitimate: a viral tweet widely retweeted by real users, a campaign slogan repeated by genuine supporters, a news headline shared by many outlets
Suspicious: identical text posted (not retweeted) by dozens of accounts within minutes, accounts with no profile pictures or follower history, accounts created in bulk around the same date
Why it matters for analysis: If a significant portion of the conversation is manufactured, any sentiment analysis drawn from it will be skewed. Identifying bot activity is a prerequisite for trustworthy opinion research
During elections, political campaigns and interest groups use social media to shape public perception. Analysts at news organisations, think tanks, and platform trust-and-safety teams use exactly these techniques to separate organic opinion from manufactured narratives.
Tags: brand audit, social media strategy, engagement metrics, interaction types, customer service on Twitter, cross-platform presence, social media KPIs
Most companies maintain accounts across multiple platforms: Facebook, Twitter/X, Instagram, YouTube, LinkedIn, TikTok, Pinterest, and sometimes niche platforms
The first step in a brand audit is identifying every active account. Brands often have separate accounts for different markets, product lines, or functions (e.g. @NikeSupport vs @Nike)
Not all platforms serve the same purpose for a brand. Instagram skews visual and lifestyle-oriented; LinkedIn targets professional audiences; Twitter/X suits real-time conversation and customer service
Follower count: Total audience size per platform. A baseline, but not a quality indicator on its own
Engagement rate: (Likes + comments + shares) / followers. A better signal of content resonance than raw follower numbers
Post frequency: How often the brand publishes. Daily, weekly, or sporadic posting tells you about resource allocation and strategy consistency
Response time: How quickly the brand replies to customer queries or complaints. Especially relevant for customer service accounts
Reach and impressions: How many unique users see the content (reach) vs how many times it is displayed (impressions)
Share of voice: The brand's proportion of the total conversation in its category compared to competitors
Broadcast: One-to-many. The brand posts announcements, promotions, or content without inviting or expecting dialogue. Example: a product launch video posted to YouTube
Request for input: The brand asks its audience for opinions, votes, or content. Example: a poll on Twitter asking "Which flavour should we bring back?"
Direct interaction: One-to-one conversation between the brand and a user. Example: replying to a customer complaint on Twitter with a personalised resolution
Combination: Most effective strategies blend all three. Broadcast builds awareness, requests for input build community, and direct interaction builds trust and loyalty
When evaluating a brand's approach, ask:
What is the brand doing well, and what evidence supports that? (High engagement, fast response times, consistent voice)
Where are the gaps? (Inactive platforms, ignored complaints, inconsistent posting, no community engagement)
What specific, actionable improvements would you recommend, and why?
Many large companies use Twitter/X as a public-facing customer service channel (e.g. @AmazonHelp, @AppleSupport)
Analysing the 10 most recent interactions reveals patterns: average response time, tone, resolution quality, and whether complaints are resolved publicly or redirected to DMs
Public customer service is a double-edged sword: good responses build reputation, but slow or dismissive ones are visible to everyone
Brand strategists, social media managers, and marketing analysts run exactly this kind of audit to benchmark performance, justify budgets, and identify competitive gaps. The metrics and frameworks here are the same ones used in industry.
Tags: social media risk assessment, reputational risk, data breach, employee social media policy, crisis management, regulatory compliance, brand safety
Reputational risk: A single viral post, whether from a disgruntled customer, a poorly worded brand tweet, or an employee's personal account, can damage a brand's reputation in hours. Crises that once took days to build now unfold in minutes
Security and data risks: Social engineering attacks, phishing via social platforms, data leaks through oversharing, and unauthorised access to brand accounts. Compromised accounts can post damaging content before the breach is even detected
Legal and regulatory risk: Posting content that violates advertising standards, copyright, privacy regulations (e.g. GDPR), or industry-specific rules. User-generated content on a brand's page may also create liability
Employee-related risk: Staff posting controversial content on personal accounts that gets linked back to the employer, or employees inadvertently leaking confidential information through social media
Operational risk: Over-reliance on a single platform (algorithm changes, account suspension, platform outages), lack of crisis response plans, or insufficient moderation of brand communities
Competitive intelligence risk: Competitors monitoring a brand's social activity for strategic information, hiring plans, product hints, or internal culture signals
Establish a clear social media policy for employees covering acceptable use, disclosure rules, and consequences
Implement multi-factor authentication and access controls on all brand accounts
Develop and rehearse a crisis communication plan with clear escalation paths
Monitor brand mentions and sentiment continuously to catch issues early
Conduct regular audits of active accounts and permissions
Train staff on recognising phishing and social engineering attempts
Risk managers, compliance teams, and social media directors use risk frameworks to protect organisations. High-profile incidents (hacked brand accounts posting offensive content, employee tweets causing PR crises, data scraped from public profiles) illustrate why this is not a theoretical exercise.
Students often think a high follower count automatically means a strong social media strategy. It does not. Engagement rate, content quality, and audience relevance matter more than raw numbers
Students confuse retweeting (legitimate amplification) with content repetition by bots. A retweet is a platform feature that attributes the original author. Coordinated inauthentic behaviour involves posting identical text as original posts across many accounts
Students assume that all social media risk is reputational. Security, legal, operational, and competitive risks are equally important and often more expensive to recover from
Students sometimes treat "broadcast" as inferior to "direct interaction." In practice, effective social media strategy uses all three interaction types for different purposes. Broadcast is not a failure; it is a deliberate choice for certain objectives
⚠️ Be able to distinguish between the three interaction types (broadcast, request for input, direct interaction) and give a concrete example of each. This is a standard exam question format for CO6
⚠️ Know how to identify signs of bot activity or coordinated inauthentic behaviour: identical text across accounts, accounts created in clusters, no profile history, posting at inhuman frequency
⚠️ Be prepared to list and explain at least four categories of social media risk with examples. The assignment explicitly asks you to "evaluate" risks (BTL2), which means you need to judge their severity and likelihood, not just list them
⚠️ Understand the difference between vanity metrics (follower count, likes) and meaningful metrics (engagement rate, response time, share of voice). Exams often ask you to critique a brand's strategy, and citing vanity metrics as evidence of success is a common trap
⚠️ The assignment maps to Bloom's Taxonomy levels 2 and 3 (BTL2, BTL3). BTL2 (evaluate) means making judgements with justification. BTL3 (apply) means using knowledge in new situations. Expect questions that ask "assess," "evaluate," or "analyse," not just "define"
True or False: A tweet retweeted by 500 accounts is always a sign of bot activity. (False: retweeting is a standard platform feature; the red flag is identical original posts, not retweets.)
Fill in the blank: The three types of brand interaction on social media are ______, ______, and ______. (Broadcast, request for input, direct interaction.)
True or False: Engagement rate is calculated as total followers divided by total posts. (False: engagement rate is (likes + comments + shares) divided by followers.)
Fill in the blank: A brand account being hacked and posting offensive content is an example of ______ risk. (Security / data risk.)
True or False: A company that only broadcasts on social media and never engages directly with users has no social media strategy. (False: broadcast is a deliberate strategy choice, though a combination of interaction types is generally more effective.)
Q: A brand posts a Twitter poll asking customers to vote on their next product colour. What type of interaction is this, and why is it strategically useful?
A: This is a request for input. It is strategically useful because it builds community engagement, generates free market research, and makes customers feel invested in the brand's decisions, increasing loyalty.
Q: You notice that 40 accounts have posted the exact same 280-character statement about a political bill within a two-hour window. None of them are retweets. What should you suspect, and what further checks would you perform?
A: This pattern suggests coordinated inauthentic behaviour, likely bot-driven. Further checks: examine account creation dates (clustered dates suggest bulk creation), profile completeness (no photos, generic bios), posting history (only political content, no personal posts), and follower/following ratios (following thousands, followed by very few).
Q: Name three categories of social media risk and provide one specific example for each.
A: Reputational risk (a brand's tone-deaf tweet goes viral and draws public backlash), security risk (a phishing attack through a fake brand DM compromises user credentials), and legal risk (a brand reposts a user's copyrighted photo without permission, violating intellectual property law).
Q: A company has 2 million followers on Instagram but averages only 500 likes per post. What does this suggest about their social media effectiveness?
A: Their engagement rate is approximately 0.025%, which is extremely low. This suggests the follower count may be inflated (purchased followers, inactive accounts) or the content is not resonating with their audience. Follower count alone is a vanity metric; engagement rate is the more meaningful measure.
Q: Explain why a crisis communication plan is essential for managing social media risk, and name two elements it should include.
A: Social media crises escalate rapidly because content spreads in real time and is visible to the public. A crisis communication plan ensures a fast, coordinated response rather than ad hoc reactions. It should include: clear escalation paths (who is notified and in what order) and pre-approved response templates for common crisis scenarios (account hack, PR incident, customer safety issue).
This material connects to data analytics and business intelligence, as the metrics and audit techniques here are a subset of broader BI skills. Sentiment analysis connects to natural language processing (NLP) in machine learning courses. Social media risk management ties into information security and corporate governance topics. If your programme covers digital marketing or e-commerce, the brand strategy section here is foundational to those modules.
Social media analytics, SMA, CSDC8203, CO6, Twitter analysis, sentiment analysis, opinion mining, bot detection, fake accounts, astroturfing, coordinated inauthentic behaviour, brand audit, social media audit, engagement rate, follower count, impressions, reach, share of voice, broadcast interaction, direct interaction, request for input, social media strategy, social media risk, reputational risk, data breach social media, social media policy, crisis communication, customer service Twitter, social media KPIs, Bloom's taxonomy BTL2 BTL3, social media metrics, vanity metrics