E-Commerce (II) PDF

E-Commerce (II) Artificial Intelligence (AI) in E-Commerce and Engaging Customers 1. The YOHO Case B2C E-Commerce Platform Electronics and home appliances retailer 2. First-Party Selling Model (1P Model) YOHO procures products directly from suppliers → Manages...

E-Commerce (II) Artificial Intelligence (AI) in E-Commerce and Engaging Customers 1. The YOHO Case B2C E-Commerce Platform Electronics and home appliances retailer 2. First-Party Selling Model (1P Model) YOHO procures products directly from suppliers → Manages inventory and handle customer orders → Distribute products and provide after-sales services High Control Asset-heavy Model 3. E-commerce Landscape in Hong Kong Limited Adoption: ○ about ⅓ individuals aged 15+ shop online ○ about ⅓ enterprises had a web presence Impact of Covid-19 Pandemic: ○ Accelerated shift ○ Increase in digital payments ○ Online sales grew by 27% 4. Strength and Weakness of E-Commerce Strength Weakness High accessibility Lack of Physical Experience Convenience Delivery Delays Cost efficiency Security Concerns Flexible Business Models Limited Personal Interaction 5. Online-Merge-Offline Model (OMO) Online + Offline stores 6. Porter’s Five Forces Model The Threat of New Entrants - Low barriers of entry facilitates competitors and increasing market competition The Bargaining Power of Buyers - customers have significant power due to numerous choices - compelling business to offer competitive pricing and quality The Bargaining Power of Suppliers - many suppliers exist, generally lower their bargaining power and allow E-commerce firms to negotiate better terms The Threat of Substitute Products - high availability of substitutes increases competition, as customers can easily switch Competitive Rivalry within Industry - increase competition among established players lead to price wars and innovation pressures 7. Generative (Gen) AI Foundation Models ○ Foundation models serve as a base for multiple applications, moving away from task-specific AI ○ They are trained on vast amounts of unstructured data, enabling them to perform diverse tasks Generative Capabilities ○ Generative Al models predict and generate content, such as predicting the next word in a sentence ○ With tuning and prompting, these models can adapt to specific natural language processing tasks with minimal labeled data Advantages of Foundation Models ○ Performance: Foundation models excel due to their extensive training on large datasets, outperforming traditional models ○ Productivity Gains: They require less labeled data for task- specific applications, thanks to their pre-training Disadvantages and Challenges ○ Compute Costs: Training and running these models can be expensive, posing challenges for smaller enterprises ○ Trustworthiness: Concerns arise from the unvetted, diverse data these models are trained on, leading to potential biases Applications Beyond Language ○ Vision Models: Technologies like DALL-E 2 generate images from text ○ Code Assistance: Tools like GitHub Copilot or ChatGPT help developers write code efficiently ○ Specialized Domains: IBM is applying foundation models in areas like chemistry and climate change research 8. Large Language Models LLMs are designed to generate human-like responses to text-based inputs and perform a wide range of natural language processing tasks, including language translation, text summarization, and sentiment analysis LLMs are distinguished by their massive size, with hundreds of millions to billions of parameters, and their ability to learn from vast amounts of data They utilize deep learning algorithms, such as transformers, to process and understand text data, enabling them to learn patterns, identify relationships, and generate accurate responses Performance Boost ○ Foundation models leverage vast datasets ○ Leading to superior performance in various applications compared to traditional models ○ Makes them attractive for businesses aiming for efficiency Generative Capabilities: ○ These models can generate new content, such as text or images ○ By predicting the next word or visual element ○ Opens up innovative possibilities in creative fields Adaptability: ○ Foundation models can be fine-tuned with minimal labeled data  ○ Making them versatile tools for specific natural language processing tasks ○ Saving time and resources Data Dependency: ○ Effectiveness of these models relies heavily on the quality and quantity of data they were trained on ○ Raising concerns about potential biases and reliability in outputs Cost Barrier: ○ High computational requirements for training and running large models can pose significant financial challenges for smaller companies ○ Limiting widespread adoption Trust Issues: ○ Uncurated nature of the training data can lead to ethical concerns ○ These models may inadvertently generate biased or harmful content Broad Application: ○ Beyond text, foundation models are being applied in multiple domains including visual content generation and scientific research ○ Showcasing their versatility and potential impact across industries 9. YOHO AI Assistant - Empowered By Generative AI Collaborated with OpenAI and Microsoft’s Azure OpenAI Functions: product search, recommendations, 24/7 customer services etc. Powerful business tool Insights from data patterns How could Gen AI help? ○ Boost visibility → Personalised Service → Increase sales → Stay competitive ○ Keep up with rapidly evolving AI technology ○ Enhance SEO with GAI to drive more traffic ○ Leverage GAI for tailored customer experiences from presales to aftersales ○ More traffic and better engagement lead to higher conversions 10. Customer Relationship Management Functions of CRM Systems i. Capture and integrate customer data from all over the organization ii. Consolidate and analyze customer data iii. Distribute customer information to various systems and customer touch points across enterprise iv. Provide single enterprise view of customers  CRM Software ○ From niche tools to large scale enterprise applications ○ More comprehensive packages have modules for: a. Partner relationship management (PRM) Integrating lead generation, pricing, promotions, order configurations, and availability Tools to assess partners’ performances b. Employee relationship management (ERM) Setting objectives, employee performance management, performance-based compensation, employee training CRM Software Capabilities ○ Sales Force Automation (SFA) Sales prospect and contact information Sales quote generation capabilities ○ Customer Service Assigning and managing customer service requests Web-based self-service capabilities ○ Marketing Capturing prospect and customer data, scheduling and tracking direct-marketing mailings or e-mail Cross-selling  Customer Loyalty Management Roadmap Operational and Analytical CRM Operational CRM Analytical CRM Customer-facing applications Based on data warehouses Sales force automation call center populated by operational CRM and customer service support systems and customer touch Marketing automation points Analyzes customer data (OLAP, data mining, etc.) Customer lifetime value (CLTV)  Analytical CRM Data Warehouse Business Value of CRM Systems Business value of CRM systems Churn rate Increased customer satisfaction Number of customers who stop Reduced direct-marketing costs using or purchasing products or More effective marketing services from a company Lower costs for customer Indicator of growth or decline of acquisition/retention firm’s customer base Increased sales revenue IS & Business Integration Big Data and Business Analytics 1. Big Data - Data is the new oil (most valuable resources in the world) Smartphones generate a large amount of data from activities like texts, calls, emails, photos, videos, searches and music. Data from billions of users is called “big data” Big data refers to extremely large and complex data sets that cannot be easily managed or analyzed with traditional data processing tools, particularly spreadsheets. Big data includes structured data, like an inventory database or list of financial transactions; unstructured data, such as social posts or videos; and mixed data sets, like those used to train large language models for AI. These data sets might include anything from the works of Shakespeare to a company’s budget spreadsheets for the last 10 years. Data is complex, enormous and cannot be dealt with by traditional processes 2. 5 V’s of Big Data Volume Scale of data/ the amount of data the more comprehensive your integrated view of the customer and the more historical data you have on them, the more insight you can extract from it → make better decisions Variety Different forms/ types of data - Structured data vs unstructured data - The more varied data, the more multifaceted view  - Enabling you to develop customer journey maps and personalisation to engage more with customer Velocity The speed at which companies receive, store and manage data - Rapid process is needed to maximise value - The more rapidly you can process information into your data and analytics platform, the more flexibility you get to find answers to your questions via queries, reports, dashboards - A rapid data ingestion and rapid analysis capability provides you with the timely and correct decisions Veracity Uncertainty of data/ how truthful is your data - Data on customers must remain consolidated, cleansed, consistent, and current to make the right decisions Value The value of data - comes from insight discovery and pattern recognition that lead to more effective operations, stronger customer relationships and other clear and quantifiable business benefits 3. Big Data and Decision Making Big data: massive amount of data is available to today’s manager Decision making is data-driven, fact-based and enabled by standardised corporate data and access to 3rd party datasets through cheap, fast computing and easier to use software 4. Big Data and Business Analytics Similarities Differences All dealing with data Different Focus BD: deals exclusively BD: technologies and processes with vast volume of involved in handling huge amount of high variety data data BA: more from the BA: focuses on operational and perspective of financial metrics that directly correlate business value with customer success or business Deriving insights profitability BD: last step after a series of complicated BD: requires a significant amount of data processing programming and creativity, the ideal BA: can be outcome is typically a set of jobs that straightforward run automatically to generate insights SQL Knowledge and metrics SQL: Structured Query BA: requires one to think from the Language is used to perspective of the customer and forn communicate with a database ideas based on how to evolve the Connects both BD and business in the best possible way  BA Writing SQL queries to BD: deals with finding ways to improve explore excel files or operational and financial performance databases of the organisation BA: finding creative ways to represent data in a form that directly translates to business outcomes Difference in data sources BD: can be structured or unstructured data and can come from any source BA: structured plus unstructured data (usually some aggregation performed) Key skills BD: core engineering and computer science skills BA: knowledge of the business and domain itself, predictive modelling 5. Types of Business Analytics Methods 6. Turning data into insights  Analytics has become an important tool for business to gain insights from data and customer information to stay competitive Different industries have specific analytics needs Companies need to understand what analytics can do for their business, how it can be applied and build a solid foundation of data to analyse 7. Applications of Business Analytics Banking and Finance ○ Locate fraudulence via identifying atypical patterns ○ Enhance cross and up selling Auditing ○ Data driven auditing Securities Trading ○ Anticipate stock fluctuations ○ Sensitivity analysis of market movements towards certain events ○ Forecast bond prices by taking into account numerous facts ○ Reveal fraudulence in trading activities Customer Lifetime Value (CLV) ○ CLV refers to the total value that a customer brings to a company over the duration of their relationship with the company ○ CLV calculation: based on your business model ○ Understanding CLV helps companies to make informed decisions regarding customer acquisition, retention strategies and investment in content creation ○ Personalisation and Recommendations: use CLV data to personalise user experience and provide tailored recommendations, increasing customer engagement and retention ○ Long-term customer relationships: leading to sustainable growth and profitability 8. Business Analytics-enabled Business Tracking online user behaviour through various stages of the purchase funnels like awareness, interest, engagement, and conversion Business can collect and analyse data to better understand customer purchase journeys Optimise shopping cart experiences, measure campaign effectiveness Eg. Google Analytics: compile data into reports that business can use to perform deep analysis and test solutions to improve business performance 9. Data Aggregation Data Aggregation: aggregate Big Data and Business Analytics Hadoop - an open source software to store unlimited amounts of diverse data in “data lake” using low cost hardware and distributed processing Unlock the value from data lake requires advanced analytics SAS and Hadoop are natural complements - SAS helps data scientists perform deeper analysis and quickly turn ideas into actions when used with Hadoop With SAS, data scientists can access and integrate with Hadoop to simplify  data preparation, freely explore and visualize data, develop predictive models using advanced analytics, and quickly deploy models inside Hadoop 10. Visual Analytics Visual analytics is the use of sophisticated tools and processes to analyze datasets using visual representations of the data Visualizing the data in graphs, charts, and maps helps users identify patterns and thereby develop actionable insights These insights help organizations make better, data-driven decisions. 11. (Digital) Dashboards Refer to the display of complex data relationships using a variety of graphical methods Data usually highly aggregated Used to visually display key performance indicators (KPI) used by management Used a variety of design elements to present data in a user-friendly way 12. Online Analytical Processing (OLAP) 13. Tableu  To conduct business analytics to make insights (similar to power BI) Import data sources from IG, X, FB to unleash power of data 14. Cookies Cookies is ○ not programs ○ is a piece of text ○ A cookie is: able to gather information from your computer ○ Allow a website to store information on user’s machine and later retrieve it ○ Potentially dangerous, because it can be a virus or spyware in disguise Cookies can ○ Provide capabilities that improve web surfing experience ○ Gather accurate information about the site’s visitors Cookies are small data files collected from websites that store information about users, like age, location, interests ○ First-party cookies: to remember things like items in a shopping cart between pages, and make the site functional ○ Third-party cookies: used by advertisers and companies to collect user data and track browsing activity across different sites How does tracking cookies work? - aggregation of data ○ allow target ads ○ there are privacy tools that can show which trackers are using cookies and allow blocking them → can limit how much data is collected 15. Data Governance Both internal and external data are needed Need to decide…. ○ Relevance (what is needed?) ○ Sourcing (where to get?) ○ Quantity and quality (how much and can be trusted?) ○ Hosting (Where to host the system, inhouse or cloud?) ○ Governance (How to properly manage data? Backup, legal, privacy, protection, etc.?)