Where technology is concerned, we’ve seen hype cycles build and crash for the past thirty years. Retail has been at the sharp end throughout – as both playground and proving ground.
From initial dotcom euphoria, through Web 2.0, to the metaverse and blockchain, we’ve seen wave after wave of innovations that looked great on paper, but which have often proved underwhelming once real-world customers and business basics have entered the equation.
Channeling Maya Angelou’s words, “You can’t really know where you are going until you know where you have been,” I thought it would be instructive to look at the tech cycles that have brought us to where we are today – and explore what they can teach retail companies about achieving success in the AI era.
In this first of two blogs, I’ll be looking at how we got here. For our follow up blog, I'll focus on where AI can truly deliver value and how retail companies can harness it.
Web1.0 – hypertext, Mosaic and the beginnings of commerce
While Web1.0 started around 1990, it didn’t begin to flourish until 1993 – a labyrinth of static pages and eventually, clunky shopping baskets.
Indexes like Yahoo! and search engines from the likes of Excite, Infoseek and (my favorite of the time) Alta Vista, all preceded Google and allowed users to seek out content which wasn’t easy to find at that time. At the start of 1995, there were about 10,000 sites and by the end of the year, this had increased tenfold.
Even with its simplicity by today’s standards, it triggered one of the wildest speculative manias in modern business. In the internet’s early days, around 1995, I was in the thick of it, working with a team to develop what was arguably Australia’s first e-commerce platform; as we knew it then, a shopping-cart development tool – a Perl-based beast that showed just how much elbow grease went into making online retail work.
The dotcom bubble
Soon after that the hype train gained momentum and things went crazy, with the NASDAQ more than doubling from the late 1990s through to March 2000. Retail was squarely in the crosshairs, both on the way up and the way down.
When the bubble burst around March 2000, the NASDAQ plummeted three-quarters from its peak – and many dotcom retailers (remember Boo.com and Pets.com?) disappeared, having burnt cash with no route to profit.
Emerging from the wreckage, players like Amazon and eBay dominated because they did the unglamorous basics: marrying a new channel to old-fashioned unit economics. But ultimately, lessons that should have been learned, weren’t.
Web2.0, platform power and “social commerce”
Despite the setback, retail companies continued to rush towards the latest digital frontier for fear of missing out. Next, capital gushed in, before valuations and value parted company. And then, inevitably, reality reasserted itself, the tourists vanished and builders quietly became part of the infrastructure.
It was the same story with Web2.0. User-generated content, social networks and app stores transformed the internet from read-only brochure into a read-write environment. For retail, three things really mattered: platforms controlling discovery; reviews as a new trust-building layer; and social media as both megaphone and complaints desk.
The upside was clear. Retailers could reach global audiences cheaply, experiment with marketplaces, and mine behavioral data at unprecedented scale. The downside? Dependency. Traffic, attribution and even brand perception became hostage to search algorithms, social feeds and app-store policies that were beyond retailers’ control.
As ever, snake oil showed up, with companies being told to “build a Facebook app, go viral, job done.” Agencies sold vanity metrics, likes and followers, long before they could point to incremental margin or lifetime value. That didn’t start to materialize until retailers learned to use Web2.0 data to tune assortment, pricing and operations, rather than just chasing the latest social trend.
See you in the metaverse?
As Web2.0 played out, Second Life – launched in 2003 – emerged as a clunky premonition of the metaverse: a user-created virtual world with its own economy and digital property. By the mid-2000s it had accrued millions of residents, a sizeable virtual goods economy and many breathless articles about people earning six-figure incomes inside a synthetic world.
Between around 2005 and 2007, major retailers piled in. The assumption was that this was the next internet and they had to plant a flag by building virtual campuses and “flagship” stores. Many of these experiments were quietly killed off because they misread the size and depth of the audience and badly overestimated customer's willingness to bother with plug-ins, avatars and unfamiliar online controls.
The fakery here was mostly theatrical: screenshots of virtual branches, overblown press releases hailing “our presence in the metaverse,” and internal decks that used exposure and novelty as proxies for customer value.
The bottom line? While Second Life did pioneer ideas around digital land, creator economies and virtual identity that resurfaced later in metaverse and NFT conversations, the timing, tooling and user base were nowhere near ready to carry a mainstream retail brand.
Blockchain, NFTs and the quest for use-cases
Then blockchain arrived, promising trustless, decentralized systems that did away with expensive intermediaries. In retail, the sensible end of the conversation focused on supply chain transparency, anti-counterfeiting, provenance for high-value goods and programmable loyalty – all areas where an immutable ledger can and does reduce fraud and support audit.
NFTs were the loud, speculative tip of the spear. On paper, they offered digital ownership for collectibles, brand experiences and loyalty assets. While larger UK brands dipped in with NFT experiments, many initiatives fizzled out with little lasting impact on sales or loyalty. Let’s not forget that Justin Bieber bought a Bored Ape NFT for $1.3m in 2022. Today, it’s worth $12,000 – that ape is more glum than bored these days.
This is an area where the term “snake oil” feels especially fair. A fog of jargon, non-fungible tokens, decentralized autonomous organizations and so on, was used to hide the simple fact that, in most cases, there was no customer problem worth solving. The elements most likely to endure now sit quietly in the background: traceability, authentication and programmable loyalty – not speculative JPEGs on a trading site.
Next stop, AI, Gen AI and agentification. In my next blog I’ll investigate how lessons learned from the past thirty years can pay dividends from now on. Stay tuned.
