E-commerce didn't arrive with a bang. It arrived through a handful of small, slightly awkward test transactions in the mid-1990s, each one proving out a piece of the infrastructure — payment, trust, delivery — that online shopping would eventually depend on entirely.
A pizza, paid for in cash on delivery (1994)
Pizza Hut's PizzaNet service let customers place an order online as early as 1994, making it one of the earliest examples of ordering a real-world product over the web. It came with an important caveat, though: there was no online payment involved. Customers still paid the driver in cash when the pizza arrived, meaning the internet's role was limited to placing the order, not actually completing a financial transaction.
A Sting CD, paid for with an encrypted credit card (1994)
On August 11, 1994, a small online storefront called NetMarket processed the sale of a CD copy of Sting's Ten Summoner's Tales for $12.48 plus shipping, with the buyer's credit card information encrypted during transmission. Because it involved an actual secured online payment rather than cash on arrival, this transaction is generally regarded as the first genuinely secure e-commerce purchase — a meaningfully different milestone than PizzaNet's order-ahead pizza, even though the pizza came first chronologically.
A broken laser pointer (1995)
Pierre Omidyar built an early auction site over Labor Day weekend in 1995, initially called AuctionWeb, and tested it by listing a laser pointer he owned that had broken shortly after he bought it. He assumed it wouldn't attract a real buyer. It sold for $14.83 to a Canadian electronics hobbyist who was, by his own account, specifically hunting for a broken laser pointer he could take apart and repair. That site became eBay.
A hard-to-find academic book (1995)
Amazon, which launched in July 1995 selling books exclusively, recorded its first customer order on April 3, 1995, during its testing phase before public launch — for a copy of Douglas Hofstadter's Fluid Concepts and Creative Analogies, an academic book about artificial intelligence and cognition. It was a fitting, if unglamorous, first sale for a company that would eventually sell almost everything.
Groceries, flowers, and the first wave of online retail (mid-to-late 1990s)
As secure payment processing became more standardized through the mid-1990s, a wave of category-specific online retailers followed the same basic model NetMarket had proven out: flowers, books, and eventually groceries all found early online sellers experimenting with the same core promise — browse a catalog, pay securely, wait for delivery — that had seemed genuinely uncertain just a year or two earlier.
Why these small purchases mattered
None of these transactions were large, remarkable, or planned as historic moments. They were mostly tests — a CD bought between friends, a broken gadget listed to see if anyone would bite, a book ordered during a pre-launch trial run. What made them significant in hindsight is that they worked: money changed hands over the internet, the goods showed up, and nothing went catastrophically wrong. That was enough to prove the basic premise of online shopping was viable, which is exactly the unglamorous kind of proof an entirely new industry needed before it could become the multi-trillion-dollar habit it is today.
The trust problem, not the technology problem
The biggest obstacle to early e-commerce was never really the technical difficulty of building a webpage that could take an order — it was convincing an ordinary person to type a real credit card number into a computer and trust that it would reach the right place safely. Every one of these early purchases functioned, in part, as a small public demonstration that the trust problem was solvable, which mattered just as much to the retailers watching from the sidelines as it did to the individual buyers involved. It's a reminder that plenty of technology adoption isn't really gated by whether something works, but by whether people believe it will.