Revision Notes

COU 07408 Electronic Commerce · Semester II, 2023/2024 · 40 marks in 3 hours

Organised by the topics this paper actually examines. Each topic lists the questions it feeds, so you can revise against the paper rather than the whole syllabus.

E-commerce Fundamentals

Q1.iQ1.iiQ1.viiQ1.ixQ1.x

The one-mark definitions in Section A. Each needs a single precise sentence — examiners are looking for the defining idea, not a paragraph.

Core definitions

E-commerce — the buying and selling of goods and services over the Internet or other electronic networks, including the electronic transfer of funds and data that supports those transactions.
Shopping cart — the software component that holds the items a shopper has selected, keeps quantities, prices and a running total, allows add/remove/update, and hands the order to checkout.
Drop-shipping — a fulfilment model where the seller holds no stock; orders are passed to a third-party supplier who ships directly to the customer, and the seller keeps the margin.
E-commerce CMS — software for creating, managing and publishing store content (products, categories, pages, images, promotions) without deep coding, so non-technical staff can run the catalogue.
Conversion rate — the percentage of visitors who complete a desired action, usually a purchase. Conversion rate = (conversions ÷ total visitors) × 100.

Hosted vs self-hosted platforms

Hosted (SaaS): the vendor runs the servers, software, updates and security — Shopify, Wix, BigCommerce. The merchant pays a subscription and configures rather than codes.
Self-hosted: the merchant installs the platform on their own or rented server — WooCommerce, Magento, PrestaShop. The merchant owns the code, data, hosting and maintenance.
The trade-off is control against convenience: hosted is faster to launch and needs no sysadmin skill but limits customisation; self-hosted allows unlimited customisation but demands technical capacity.

Exam tip. Question 1 is ten one-mark answers. Two or three precise sentences each is enough — do not spend essay time here.

E-commerce Business Models

Question 2 asks for five models, each with an example, one advantage and one challenge. That is 1.5 marks per model — structure your answer so all four elements are visible for each.

The five standard models

B2C (Business to Consumer) — a business sells directly to end consumers. Example: Amazon, Jumia. Advantage: wide reach and a direct customer relationship. Challenge: intense competition and high customer-acquisition cost.
B2B (Business to Business) — a business sells to another business. Example: Alibaba. Advantage: large-volume, high-value, repeat orders. Challenge: long sales cycles and complex negotiation or systems integration.
C2C (Consumer to Consumer) — consumers trade with each other through a platform. Example: eBay, OLX. Advantage: very low entry cost and wide product variety. Challenge: trust, fraud and quality control.
C2B (Consumer to Business) — individuals offer products or services to businesses. Example: Upwork, Shutterstock contributors. Advantage: access to a flexible talent and content pool. Challenge: inconsistent quality and downward pricing pressure.
B2G / G2C (Business to Government, Government to Consumer) — e-procurement and public e-services. Example: TANePS, e-government portals. Advantage: transparency and efficiency in public procurement. Challenge: heavy regulation and slow bureaucratic processes.

Also creditable

B2B2C — a business reaches consumers through another business's platform.
D2C (Direct to Consumer) — a manufacturer bypasses retailers and sells straight to buyers.
Social commerce — selling directly inside social platforms such as Instagram or WhatsApp Business.

Exam tip. Answer in five labelled blocks. Each block: name the model, one sentence describing it, a named example, one advantage, one challenge. Missing the example or the challenge costs half a mark each time.

Revenue Models

Question 3 wants five revenue models, a named company for each, and an explanation of how the model drives that company's financial success. Note the third element — naming the company is not enough.

The five to know

Sales / transaction revenue — income from selling goods directly. Amazon retail, Jumia. Success: margin earned on every unit, scaled by volume and logistics efficiency.
Advertising revenue — selling ad space and impressions. Google, Facebook, YouTube. Success: monetises an enormous audience without charging users, so growth in users converts directly into revenue.
Subscription revenue — a recurring fee for continued access. Netflix, Spotify, Microsoft 365. Success: predictable recurring income and high customer lifetime value, which makes investment planning reliable.
Transaction fee / commission — a cut of each transaction the platform facilitates. eBay, PayPal, Airbnb, Uber. Success: revenue scales with marketplace volume without the platform owning inventory or assets.
Affiliate revenue — commission for referring buyers to another seller. Amazon Associates, review and comparison sites. Success: earns from traffic alone, with no stock, fulfilment or customer-service cost.

Other valid models

Freemium — a free basic tier converts a large user base into paying upgrades. Dropbox, LinkedIn, Zoom.
Licensing / usage-based — pay for what you consume. AWS, Adobe Creative Cloud.
Many large firms are hybrids: Amazon combines retail sales, commission (Marketplace), subscription (Prime) and advertising.

Exam tip. The marks split three ways per model: the model, the company, and the financial-success explanation. Write the third part explicitly — most lost marks here are for stopping after the example.

Payments and Security

Q1.iiiQ1.ivQ4.iQ5.i

SSL and payment gateways are one-mark definitions; mobile payments and cryptocurrency are five-mark discussions that need analysis, not lists.

SSL and payment gateways (1 mark each)

SSL encrypts data in transit between browser and server so intercepted traffic cannot be read, and authenticates the server's identity through a digital certificate. It protects card numbers, passwords and personal data, enables HTTPS, and signals trust to buyers.
A payment gateway is the service that authorises and processes online payments between merchant, customer and bank. It encrypts and forwards transaction details to the acquiring bank and returns an approval or decline, enabling secure real-time payment capture.

Mobile payment systems and e-marketplaces (Q4 i, 5 marks)

Financial inclusion — mobile money such as M-Pesa, Tigo Pesa and Airtel Money lets unbanked buyers transact, widening the marketplace's customer base substantially in East Africa.
Higher conversion and lower cart abandonment through fast one-tap checkout and stored credentials.
Trust and escrow — the platform holds funds until delivery is confirmed, cutting fraud risk for both buyer and seller.
Faster settlement improves seller cash flow, and automated reconciliation cuts manual accounting work.
Rich transaction data enables analytics, credit scoring, micro-lending and personalised offers.
Management challenges: transaction fees, API integration complexity, interoperability between operators, KYC and regulatory compliance, chargeback and fraud handling, and dependence on network availability.

Cryptocurrency in e-payments (Q5 i, 5 marks)

Role: a decentralised digital payment medium settled on a blockchain without a bank intermediary; smart contracts can automate escrow and release payment on delivery.
Advantages: low transaction and cross-border remittance fees; borderless access to global and unbanked buyers; fast near-24/7 settlement; no chargebacks for the merchant; pseudonymity with a transparent immutable ledger.
Risks: extreme price volatility between sale and conversion; regulatory uncertainty and outright bans in some jurisdictions, including restrictions in Tanzania; irreversibility, so no recourse for mistaken or fraudulent transfers; exchange and wallet security, key loss, hacking and money-laundering exposure; scalability, congestion, gas fees and energy consumption.

Exam tip. Question 5(i) says 'analyze ... and discuss its advantages and potential risks'. A one-sided answer covering only advantages cannot pass half marks. Balance both.

Digital Marketing and Customer Support

Q1.viiiQ4.iiQ5.ii

SEO is a one-mark definition; email marketing and live chat are 2.5-mark 'describe' questions where about five distinct roles earns full marks.

SEO impact (1 mark)

Improves ranking in search results, raising organic visibility and traffic to product pages, which converts into sales. It brings targeted low-cost traffic compared with paid advertising, improving ROI. Achieved through keywords, product descriptions, site speed, mobile friendliness and backlinks.

Email marketing (Q4 ii, 2.5 marks)

A direct, low-cost channel to reach existing and prospective customers.
Customer retention and relationship building through newsletters, loyalty and re-engagement campaigns.
Transactional email — order confirmation, shipping and delivery notifications.
Abandoned-cart recovery emails that bring shoppers back to complete checkout.
Personalisation and segmentation from purchase history, driving recommendations and upselling.
Measurable ROI through open, click-through and conversion tracking.

Live chat support (Q5 ii, 2.5 marks)

Real-time assistance while the shopper is still browsing, answering questions instantly.
Reduces cart abandonment by resolving doubts about price, sizing, shipping or payment at the moment of hesitation.
Increases conversions and average order value through guided selling and upselling.
Improves satisfaction, trust and loyalty; cheaper than phone support since one agent handles several chats.
Handles complaints, order tracking and after-sales support; chatbots give 24/7 coverage and escalate to humans.
Chat transcripts reveal common problems and product gaps.

Exam tip. For the 2.5-mark parts, five short well-labelled points is the efficient shape — roughly half a mark each.

Machine Learning in Product Search

A full 7.5-mark question on one topic. Name the technique and explain how it improves catalogue search — both halves are needed for each point.

Techniques and applications

Learning to Rank — supervised models order results using click-through, purchase and relevance signals, so the products most likely to be bought surface first.
Query understanding — intent classification, named-entity recognition, spelling correction and query expansion turn a messy query into a searchable one.
Semantic / vector search — embeddings and NLP match meaning rather than exact keywords, so synonyms and descriptive queries still return the right products.
Autocomplete and query suggestion learned from historical search logs, shortening the path to a product.
Personalisation — collaborative and content-based recommender systems reorder results for each user's history and segment.
Visual search — CNNs and image embeddings let shoppers search by photograph instead of words.
Voice search — speech recognition plus NLP for conversational product queries on mobile and smart speakers.
Automatic categorisation and attribute extraction from titles, descriptions and images, enriching a catalogue too large to tag by hand.
Zero-result and long-tail handling through query relaxation and similar-item retrieval.
Fraud, spam and duplicate-listing detection keeping the catalogue clean and trustworthy.
A/B testing and reinforcement learning to continuously optimise ranking against live behaviour.

Benefits and challenges

Benefits: better relevance, faster product discovery, higher conversion, lower search-abandonment, and scalable handling of catalogues with millions of items.
Challenges: data quality and labelling, cold start for new products and users, training cost, inference latency at scale, and bias in learned rankings.

Exam tip. Five well-explained techniques at roughly 1.5 marks each fills the question. Adding a short benefits-and-challenges paragraph at the end lifts a listing answer into a discussion.

Development: Scripting, Frameworks and M-commerce

Q1.vQ7.iQ7.ii

The technical question. Question 7(i) is worth 5 marks and rewards a clear split between client-side and server-side scripting.

Mobile commerce (1 mark)

M-commerce is the conduct of commercial transactions using mobile devices over wireless networks — a subset of e-commerce covering mobile shopping, mobile banking, mobile money and mobile payments, enabled by apps, responsive sites, location awareness and mobile wallets.

Scripting languages in shopping carts (Q7 i, 5 marks)

Client-side (JavaScript): add/remove items, update quantities and recalculate totals without reloading the page.
Client-side form validation before submission, improving responsiveness and cutting invalid requests.
AJAX/fetch calls for asynchronous cart updates and live stock or price checks.
Server-side (PHP, Python, Node.js, Ruby): processes cart logic and applies pricing, tax, discount and shipping rules.
Session and cookie management to persist the cart across pages and visits, for both guests and logged-in users.
Database interaction — reading product data and writing order records.
Integration with payment gateways and inventory systems during checkout.
Security handling — input sanitisation, CSRF protection, and preventing client-side price tampering.

Django for an e-commerce cart (Q7 ii, 2.5 marks)

Batteries-included: authentication, sessions, forms and admin ship with the framework, so development is fast.
The ORM abstracts the database and simplifies product, cart and order models with managed migrations.
The built-in admin interface manages the catalogue and orders with almost no extra code.
Strong security defaults — CSRF, XSS, SQL-injection and clickjacking protection, plus secure password hashing.
MVT architecture gives clean separation of concerns and maintainable code.
Proven scalability, with caching and middleware support for high-traffic stores.
A rich ecosystem — django-oscar, Saleor, Django REST framework and ready payment integrations.

Exam tip. In 7(i), explicitly label which languages are client-side and which are server-side. That distinction is what separates a 5/5 answer from a 3/5 one.

Exam Strategy

1Three hours for 40 marks. Section A is 10 marks — give it about 30 minutes, roughly 3 minutes per one-mark answer.
2That leaves about 30 minutes for each of the four Section B questions, plus time to read the paper and check your work.
3Read all six Section B questions before choosing. Pick the four you can support with examples, not just the four you recognise.
4Watch the command word: 'define' and 'describe' want content; 'discuss' and 'analyze' want reasoning, trade-offs and both sides.
5When a question asks for a fixed number — five models, five revenue models — give exactly that number and label them clearly. Marks are allocated per item.
6Named real-world examples earn marks. Amazon, Alibaba, eBay, Jumia, M-Pesa, Netflix, Shopify and Django are all safe, specific choices.

Sit the exam →