Online shopping did not become easy in one moment.
The Web made a store reachable. Search made products findable. Digital payments made remote transactions safer. Cloud software made stores easier to build. Smartphones made shopping continuous. Creator platforms connected entertainment with purchase. AI now aims to perform part of the shopping work itself.
Quick Answer
The history of e-commerce is a history of disappearing friction.
Each technology wave removed one difficult step from finding, comparing, paying for, delivering, or returning a product.
Each wave also changed who held power.
Every wave of e-commerce removed a different form of friction and shifted power toward the company controlling the new interface.
This principle is more useful than memorizing a timeline.
It gives us a way to judge what AI shopping agents, augmented reality, autonomous delivery, or another future technology may change next.
Wave 1: The Web Made the Store Reachable
Tim Berners-Lee invented the World Wide Web at CERN in 1989. CERN released the Web software in 1991, allowing the idea to spread beyond the laboratory.[1]
The early Web was not designed as a shopping mall. It created an open way to connect documents, addresses, and browsers across different computers.
Commerce followed because a seller could now publish a catalog that anyone with Web access could visit.
Amazon opened its online bookstore in July 1995. It argued that the Internet made a selection possible that no physical bookstore in one city could economically hold.[2]
eBay also began in 1995. Its marketplace connected individuals who might never have found each other offline.[3]
The first major friction disappeared:
The customer no longer had to visit the store to see what was available.
The new control point was access. The site that organized the catalog or marketplace decided which products could be found.
Wave 2: Payments, Search, and Advertising Connected Intent With Commerce
A product page was useful only if the buyer could trust the transaction and complete payment.
PayPal was founded in 1998 and developed digital payment services that allowed people to pay and receive money online.[4]
Search solved another problem.
A customer did not need to remember the address of every store. The customer could describe the need.
Google launched AdWords in 2000. The basic commercial insight was powerful: a search query could reveal intent, and an advertisement could appear near that intent.[5]
At the same time, marketplace systems expanded outside the United States.
Alibaba was established in China in 1999 with a mission to make it easier to do business. Taobao followed in 2003 and grew into a broad consumer-commerce ecosystem.[6]
The second friction disappeared:
The customer no longer had to know which seller carried the product or how to exchange payment in person.
Power moved toward search engines, payment identities, and marketplaces that could create trust.
Wave 3: Cloud Software Made Commerce Easier to Build
Early online stores were difficult to operate.
A merchant needed servers, product databases, shopping-cart software, security, payment connections, order management, and technical staff.
Cloud computing, software as a service, and application programming interfaces allowed companies to rent more of this capability.
AWS describes the 2000s as a period when cloud infrastructure, SaaS, and APIs helped digital commerce mature. Modern commerce architecture increasingly became modular: merchants could connect specialized services instead of building one closed system from the beginning.[7]
Shopify, co-founded in 2004, became one example of this shift. It developed a commerce platform that could support businesses without requiring each merchant to maintain the entire underlying software stack.[8]
The third friction disappeared:
A merchant no longer had to build every part of the commerce system alone.
Power moved toward cloud providers, commerce platforms, payment gateways, and software ecosystems.
Wave 4: Smartphones and Mobile Money Made Shopping Continuous
Desktop commerce still required a person to sit at a computer.
Apple introduced the iPhone in 2007 as a phone, media device, and Internet communicator in one product. The App Store opened in July 2008 with 500 apps.[9]
Shopping could now happen during travel, inside a store, after an alert, or while viewing social content.
Mobile commerce did not follow one global path.
In Kenya, Safaricom and Vodafone launched M-Pesa in 2007. It allowed mobile-phone users to send and receive money, including many people without traditional bank access.[10]
In Latin America, Mercado Libre developed an ecosystem connecting commerce with Mercado Pago, logistics, advertising, and merchant services.[11]
China developed another integrated model through marketplaces, mobile wallets, super-app behavior, livestreaming, and merchant ecosystems.
The fourth friction disappeared:
Shopping and payment no longer required a desktop computer or even a nearby bank branch.
The new control points were mobile identity, app distribution, payment, logistics, and local ecosystem data.
Wave 5: Content Became Part of the Store
Search begins with a known need.
Recommendation feeds can create the need.
Short video, livestreaming, creators, product tags, and in-app checkout reduced the distance between entertainment and purchase.
TikTok calls this discovery e-commerce. In supported markets, users can discover products through shoppable videos, livestreams, and a marketplace, then complete purchases without leaving the app.[12]
YouTube, Meta, Naver, Pinterest, Amazon, and other platforms are building different versions of the same commerce stack.
The fifth friction disappeared:
The customer no longer had to stop watching, search again, and rebuild the buying decision.
Power moved toward recommendation systems, creators with trust, attribution tools, and platforms that could connect content with transactions.
Figure 1. E-commerce evolved through different regional paths, but each wave removed friction and created a new control point.
Wave 6: AI May Remove the Shopping Work Itself
Until now, the customer usually performed the workflow.
The customer searched, opened several pages, compared specifications, read reviews, checked availability, selected an item, and completed payment.
AI shopping agents aim to compress those steps.
In January 2026, Google introduced the Universal Commerce Protocol, an open standard designed to let agents, retailers, and payment providers communicate across discovery, purchase, and post-purchase support.[13]
In May 2026, Google added Universal Cart features that can gather products from participating merchants and connect the shopper with Google Pay or a merchant checkout.[14]
The emerging workflow looks different:
Human states a goal
→ AI searches and compares
→ AI recommends
→ Human approves
→ Agent and merchant complete the transaction
The sixth friction may be the shopping workflow itself.
The new control points are likely to include:
- the AI interface where the customer begins
- structured and current product data
- identity and payment permission
- the rules used to compare products
- merchant and logistics connections
- the customer’s long-term preferences
Why Technology Adoption Changes Commerce
A commerce technology matters only when enough people and businesses can use it.
ITU estimated that about 6 billion people, or 74% of the world’s population, were using the Internet in 2025.[15]
Faster networks, cheaper smartphones, cloud infrastructure, payment systems, logistics, and software standards created the conditions for today’s commerce.
AI is not replacing that foundation. It is being built on top of it.
A new interface becomes commercially powerful when infrastructure, software, trust, payment, and user behavior become ready at the same time.
How to Read the Next Commerce Technology
History gives us a reusable framework.
When a new technology appears, ask five questions.
1. What Friction Becomes Cheaper?
Does the technology reduce search time, comparison cost, payment steps, delivery time, uncertainty, or development cost?
2. Which Step Disappears From the Interface?
Does the customer still need a search box, a website visit, a shopping cart, or a separate checkout screen?
3. Who Receives the New Data?
Does the data go to the seller, search engine, content platform, AI company, payment provider, or logistics company?
4. Who Becomes the Default Decision Interface?
Where will the customer begin: a marketplace, creator, search engine, specialist publisher, or AI agent?
5. What Becomes More Valuable?
Automation can increase the value of trust, original evidence, structured product data, direct experience, accurate availability, payment access, and customer relationships.
Figure 2. Use the Future Commerce Lens to evaluate new technology, then score the current strength of your own position.
Action: Run the Commerce Power Score
Score each area from 0 to 2.
- 0: absent or highly exposed
- 1: partially developed
- 2: controlled, measurable, and reusable
| Dimension | Question |
|---|---|
| Discovery diversity | Can people find you through more than one platform? |
| Trust and evidence | Do you provide evidence that cannot be replaced by a generic summary? |
| Structured data | Can people and machines understand your products or knowledge accurately? |
| Attribution | Can you connect content with a useful result? |
| Checkout access | Is there a clear path from decision to action? |
| Customer relationship | Can you reach the reader or customer again? |
| Revenue diversity | Can the business survive weakness in one revenue source? |
| AI readiness | Are data, workflows, and rules ready to connect with AI? |
Add the eight scores.
- 0–5: Platform-exposed
- 6–10: Transitioning
- 11–16: Resilient and adaptable
The total is not the main result.
The two lowest dimensions become the next actions.
Examples:
- Low trust: publish one original test, calculation, or dataset.
- Low customer relationship: create a newsletter, account, or direct-return reason.
- Low structured data: organize products or research into clear tables and metadata.
- Low attribution: track the next useful action, not only pageviews.
- Low revenue diversity: test one new model without weakening editorial trust.
What the Next 30 Years May Change
AI shopping agents are not the end of commerce history.
Future waves may combine:
- augmented-reality product testing
- autonomous vehicles and delivery robots
- personal agents with long-term memory
- machine-to-machine purchasing
- dynamic manufacturing and local production
- new identity, payment, and privacy systems
The exact winner is difficult to predict.
The analytical method is more stable.
Find the friction that disappears. Find the data that appears. Find the interface that becomes the default. Then ask who keeps the customer relationship.
Conclusion
The Web made stores reachable. Search made products findable. Payments made remote exchange practical. Cloud software made commerce easier to build. Smartphones made shopping continuous. Creator platforms placed products inside content. AI agents now aim to perform the shopping workflow.
None of these waves erased the earlier layers.
Each one was built on top of them.
That is why the next commerce system will not be created by AI alone.
It will depend on networks, software, accurate data, payments, logistics, merchant trust, and customer permission.
The companies and creators that understand this pattern can do more than react to the next technology.
They can prepare for the control point that technology is about to create.
Key Vocabulary & Phrases
friction
A cost, delay, difficulty, or uncertainty that makes an action harder.
Digital payments removed friction from remote transactions.
control point
A stage where ownership of data, access, or infrastructure creates power.
Search intent became an important control point.
default interface
The place where users normally begin a task.
An AI assistant may become the default interface for shopping.
structured data
Information organized in a consistent form that software can understand.
AI agents need accurate structured product data.
leapfrog
To move directly to a newer system without passing through every earlier stage.
Mobile money allowed some markets to leapfrog branch-based banking.
built on top of
Added as a new layer while the earlier foundation remains.
Agentic commerce is built on top of mobile networks, cloud software, and digital payments.
Read the Series
- From Attention to Transactions: Why the Internet’s Business Model Is Changing
- The Race to Make Every Video Shoppable: YouTube, TikTok, Meta, and Naver
- Affiliate Marketing vs. Advertising: Which Model Wins the Creator Economy?
- Is Google Still an Advertising Company? Search, YouTube, AI, and Commerce
- Who Wins When Content Becomes Commerce? Platforms, Creators, and Publishers
Sources
- A short history of the Web — CERN.
- World’s Largest Bookseller Opens on the Web — Amazon, October 4, 1995.
- Our History — eBay.
- PayPal corporate history — PayPal.
- Introducing simpler brands and solutions for advertisers and publishers — Google, June 27, 2018.
- Introduction to Alibaba Group — Alibaba Group.
- A Short History of Digital Commerce — AWS.
- Shopify board and founder information — Shopify.
- The App Store turns 10 — Apple, July 5, 2018.
- M-Pesa marks 15 years — Vodafone.
- About MELI — Mercado Libre.
- TikTok Shop: The Future of Shopping Fueled by Discovery E-Commerce — TikTok.
- New tech and tools for retailers to succeed in an agentic era — Google, January 11, 2026.
- Introducing the Universal Cart and more ways to help you shop — Google, May 19, 2026.
- ITU Statistics — International Telecommunication Union.
This article presents a historical and strategic framework. Dates and product features are based on the cited sources. Company and platform development followed different paths across countries. Sources checked through July 25, 2026.