Customer Engagement Test 2 PDF

Customer Engagement Test 2: =========================== Lecture 7: ---------- **Paid Social Media Metrics** - Paid social media metrics: - Percentage: - Click through rate (CTR): percentage of impressions that turn into clicks. **(Clicks/impressions)** -...

Customer Engagement Test 2: =========================== Lecture 7: ---------- **Paid Social Media Metrics** - Paid social media metrics: - Percentage: - Click through rate (CTR): percentage of impressions that turn into clicks. **(Clicks/impressions)** - Conversion Rate: percentage of clicks that turn into conversions **(Conversions/clicks)** - Cost metrics: - Cost per click (CPC): amount of money spent per click **(Spends/clicks)** - Cost per acquisition (CPA): amount of money spent on each conversion **(Spend/conversions)** - Cost per Mille (CPM): Cost per thousands impression, cost paid for every thousand impressions **(total ad spent/ total ad impressions) x 1000** **A/B Testing** - A/B Testing allows marketers to experiment with different digital options to identify which ones are likely to be the most effective  - Marketers can analyze differences in click-through rates, conversion rates, and cost per click between two versions of the same social media ad  - Various elements can be A/B tested, including post text, link preview content, calls to action, use of images or videos, ad format, hashtags, and target audience  - A/B testing is designed to improve online visitor experiences and increase the likelihood of purchasing by understanding which variations perform best **Social Media Advertising:** - Social media platforms are effective channels for delivering paid advertising campaigns and content, allowing marketers to reach a wide target audience \[5\]. - Facebook offers various marketing objectives for ads, including brand awareness, reach, and traffic, which help define what the ad aims to accomplish \[7\]. - Facebook\'s Ads Manager allows companies to plan their advertising campaigns and track ad efficacy and performance \[3\]. - Different types of ads available on Facebook include image ads, video ads, carousel ads, slideshow ads, collection ads, and Messenger ads \[8\]\[9\]. **Facebook and Instagram ads:** - Both Facebook and Instagram ads are part of the same company (Meta) - They appear throughout the app, including in users' feeds, Stories, Explore, and more - They look similar to normal posts but always contain a "sponsored" label to indicate that they are an ad - They also often have more features than a normal post, such as links, CTA buttons, and product catalogues - Both Facebook and Instagram ads can be managed through an integrated ad management platform, aka Facebook's Ads Manager: - Easy for marketers to select multiple audiences and placement - Easy to manage campaigns and download their results **Several factors contribute to the cost of the Ad Campaign:** - **Timing:** The month, day and even hour can affect ad cost - **Bidding strategy:** Whether you choose the lowest cost or a specific bid cap (only relevant for Facebook\*) - **Industry's competitiveness and ad placement:** Higher competition spots cost more - **Ad relevance:** Low scores for your ad's engagement ranking, quality ranking or conversion ranking can raise costs - **Target audience:** Higher competition audiences cost more - You need to have a Facebook business page - You need to use Facebook Ads Manager or Business Manager to create your Facebook ad campaign - Select campaigns then create - Facebook offers 11 marketing objectives based on what you want your ad to accomplish: - Brand awareness, reach, traffic, engagement, app installs, video views, lead generations, messages, etc **Step 2: Name the campaign** - You can also set an A/B Test -- run different versions - You decide how much money you want to spend- can choose daily or lifetime budget - Runnings ads on a schedule may be more efficient- choose to serve ad when target audience is active - Build target audience for ad -- target location, age, gender and language - Want to make audience as specific as possible **Step 5: Choose ad placement** - Simplest choice: automatic placement -- automatically place ads across Facebook, Instagram, messenger when likely to get best results. - Choose own options, based on: device type, platform, placements, specific mobile devices. - Brand Safety section to exclude any types of content that would be inappropriate to appear with your ad - Can optimise your ad bidding strategy and bidding type, and add an optional bid control - First choose ad format, then enter text and media components- formats will vary based on the campaign objective. - Can transform an image into a video or short animation **How to advertise on Instagram:** - **Two routes:** - Ads Manager (same as Facebook) - Promoting a post - Promoting a post is the easiest way to start advertising on Instagram - You need a business or creator account on Instagram and a Facebook Business Page connected to your Instagram account to do this - Click on Promote underneath the desired post - You'll be prompted to choose your preferred audience, destination, budget, and duration for your ad to run **Facebook Ads Manager Analytics:** 1. Customise metrics you want to see 2. Compare performance across different date ranges 3. Use charts to analyse results Lecture 8: ---------- **[Natural Language Processing: ]** - **The Background:** - Large amounts of data are now available---about 25% is structured and ready to use, while 75% is unstructured, coming from sources like social media posts, reviews, emails, customer service calls, sales calls, and chatbots. - **Examples**: Siri, ChatGPT, Alexa - **Uses**: Companies use NLP to analyse and interpret internal and external data. - **For Marketers**: NLP helps improve brand reputation, customer satisfaction, market intelligence, and product/service quality. - **Marketing Use**: - Marketers use NLP to understand the \"what,\" \"how,\" \"when,\" and \"why\" of customer and market behaviour. NLP helps extract underlying motivations, intentions, beliefs, and attitudes driving that behaviour. **[Text Analytics: ]** **Text Analytics Process** - **Purpose**: Convert text into numbers to gain insights. - **Steps**: 1. **Text Acquisition**: Collect text data from sources like reviews, social media, call transcripts. 2. **Text Preprocessing**: Clean and standardize the data by removing noise (e.g., punctuation, stop words) and applying techniques like: - **Tokenization**: Breaking text into smaller pieces (tokens). - **Stemming & Lemmatization**: Reducing words to their base form (e.g., \"loving\" becomes \"love\"). - **N-grams**: Identifying sequences of words that frequently appear together. - **Bag of Words**: Counting word occurrences, ignoring grammar. 3. **Text Exploration**: Use visual tools like word clouds or frequency charts to see patterns in the text. 4. **Text Modelling**: Build models to uncover insights, such as: - **Topic Modelling**: Discover themes (e.g., using LDA). - **Sentiment Analysis**: Detect emotions or opinions in text. **[How to interpret Text Mining with Leximancer]** - **Leximancer**: Software used to analyze text and visually represent its conceptual structure. - **Key Features**: - **Conceptual Analysis**: Identifies important words and groups them into concepts. - **Relational Analysis**: Measures the relationships between concepts and displays them visually. - **Heatmap**: Highlights themes by color, with red as the most prominent and purple as the least. - **Concept Clusters**: Words that frequently appear together are displayed close to each other, showing their connection. - **How to Interpret the Map**: - The size of the grey dots represents the importance of concepts. - Themes are named after the most dominant concept. - Understanding the context of the source text is crucial for interpreting the map correctly. **[How to Interpret Sentiment Analysis]** - **Purpose**: Analyse text to determine the sentiment (positive, negative, neutral) and emotions behind it. - **Types of Sentiment Analysis**: - **Polarity Analysis**: Determines whether the communication is positive or negative. - **Entity & Emotion Tracking**: Identifies the subject of communication and the associated emotions. - **Context Analysis**: Looks at the context of the text to understand its meaning more accurately. - **Challenges**: - Confusion over the number of emotions to use (some systems use 6 emotions, others 8 or more). - Contradictory feedback can make it hard to extract a clear opinion. - **Solutions**: - **Aspect-based Analysis**: Focus on specific aspects of the text to improve accuracy. - **Geo-location Analysis**: Link sentiments to specific geographical locations. Lecture 9: ---------- **[Geo-location Analytics]** **Definition**: - Geo-location analytics involves mapping, visualizing, and mining location data to reveal patterns, trends, and relationships hidden in data. It is used for business intelligence and can include techniques like clustering, heat mapping, and data aggregation - **Business Data-driven Location Analytics**: Focuses on mapping and visualizing business data to gain insights  - **Social Media Data-driven Location Analytics**: Utilizes social media location data to analyse user behaviour and preferences **Applications**: - Common applications include recommendations, customer segmentation, advertising, navigation, and search and rescue **Methods**: - Techniques such as movement and visitation analysis help identify trends in foot traffic and audience behaviour ### ### ### **[Mobile Analytics]** **Definition**: - Mobile analytics refers to the analysis and tracking of data from mobile applications to understand user behaviour and improve user experience \[5\]. **Tools**: - Notable mobile analytics tools include Google Mobile Analytics, Count.ly, Mixpanel, and Flurry, each serving different analytical purposes \[5\]. **Importance**: - Mobile analytics is crucial for marketers to optimize mobile app experiences and enhance advertising effectiveness Lecture 10: ----------- **[Network and Influencers]** **What Are Influencers:** - Individuals who actively engage others in conversation and are well-connected within a social network. They function similarly to opinion leaders and are crucial in marketing for creating awareness and engagement. **Importance in Marketing**: - Influencers are especially significant at the top of the digital marketing funnel, where the focus is on creating **awareness** and driving **consideration**. **Metrics to Track**: - **Top of the funnel**: Reach, impressions, clicks. - **Middle of the funnel**: Interactions, shares, engagement. - **Bottom of the funnel**: Sales, conversions, referrals, sentiment. **[Social Network Analysis (SNA)]** **Definition**: - SNA identifies relationships, influencers, information dissemination patterns, and behaviors within a network. This analysis visualizes and helps understand how social networks operate and influence marketing. **Applications**: - Helps businesses understand customer conversations, relationships, and brand perceptions on social media platforms. **Structure**: - **Nodes**: Represent entities (e.g., people, products, organizations). - **Links/Edges**: Represent the relationships between nodes. - **Egocentric Networks**: Networks focused on a single node and its direct connections. - **Real-World vs. Online Networks**: Real-world networks (e.g., friends) differ from online networks (e.g., Twitter followers). **[Social Network Properties and Metrics]** **Node-Level Metrics**: - **Degree Centrality**: Measures a node's importance based on the number of connections (e.g., followers on Twitter). - **Betweenness Centrality**: Reflects a node\'s central position and its ability to control the flow of information between other nodes. - **Eigenvector Centrality**: Measures the importance of a node based on its connections with other important nodes (e.g., Google search rankings). - **Closeness Centrality**: How close a node is to all other nodes, affecting how quickly information spreads. - **Structural Holes**: Refers to gaps in a network where nodes can have advantages or disadvantages based on their location. **Network-Level Metrics**: - **Density**: Indicates how fast information travels based on the number of connections. - **Clustering Coefficient**: Measures the extent to which nodes in a network form tightly-knit groups. - **Diameter**: The longest path between any two nodes in the network. **[Link Prediction]** **What Is Link Prediction?**: - Predicts future links between unconnected nodes in a network by analysing their proximity or similarity. Common in platforms like Tinder (suggesting matches) and Facebook (suggesting \"People You May Know\"). **Applications**: - Helps businesses predict new customer connections or suggest partnerships, enhancing the customer engagement strategy. **[Hyperlinks and Hyperlink Analytics]** **Hyperlinks**: - References to other web resources, critical for generating web traffic and validating relationships between organizations. **Types of Hyperlinks**: - **In-Links**: Links coming into a website from other sites (important for SEO and traffic). - **Out-Links**: Links going out from a website to others. - **Co-Links**: When two websites receive or provide links to/from a third website. **Importance of Hyperlink Analytics**: - Tracks website traffic and interactions. - Improves **SEO** by increasing page visits and reducing bounce rates. - Measures relationships (trust, validation) between websites.