Thirty‑five degrees of heat in West Oxfordshire is enough to make even the most stoic of us look for any way at all to keep cool. So I ended up in the garden studio, hunting for an old electric fan I knew was tucked away somewhere.
I found it quickly enough. And as soon as I picked it up, I remembered where it had come from, twenty‑odd years ago. On the centre badge was a logo I hadn’t thought about for years: Overture.
 |
| My Overture Sponsored Fan |
The name once meant something. Overture powered a huge share of the sponsored links on the early web — a company right at the centre of the first wave of Search Engine Marketing, and one that had sent me, and others in the field, this fan as a promotional gift.
Now, with this relic blowing warm air at me, the past felt unexpectedly close. The fan still worked. The company did not. And somewhere in the back of my mind, in heat‑induced nostalgia, a memory surfaced of a diagram — a map of how search engines once related to one another, before the web settled into the shape we now take for granted.
That rediscovery sent me back into my cloud archive. And there it was: an image file from 2004, a tangle of arrows and logos showing the hidden plumbing of early search. A reminder that the web search was once built around a federation, not an empire.
This is it:
 |
| Search Engine Relationships 2024 - Click/Tap to Enlarge |
Why the diagram looked like this
Today we’re used to the idea that a search engine is a single, unified thing: one company, one index, one set of results. But in the early 2000s, it simply didn’t work like that.
Most companies calling themselves “search engines” didn’t actually crawl the web. Building a full index was expensive and technically demanding. So many licensed results from someone else. Some mixed those licensed results with their own small index. Others layered paid listings on top. And almost all of them relied on partnerships to fill the gaps.
It was a patchwork — but a functional one. A kind of cooperative competition. Everyone needed everyone else, and the boundaries between “search engine,” “portal,” “directory,” and “advertising network” were far more porous than they are now.
That’s the world the diagram captures.
How the diagram works
The diagram itself is straightforward once you know what you’re looking at.
At the top are the big names of the time: Google, Yahoo, MSN, Ask Jeeves, Lycos, AltaVista, AOL. Each sits in its own box, but none of them stands alone. Arrows run between them, showing who supplied results to whom:
Google powering AOL’s search results
Overture supplying paid listings to Yahoo
Teoma feeding Ask Jeeves
MSN still relying on partners while building its own crawler
The arrows show the actual flow of data — who depended on whose index, whose ads, whose directory listings.
Below that, the diagram widens into the “verticals”: travel, cars, shopping, insurance, property, finance, business, technology. These were early attempts to organise web search into categories. Each vertical lists the major UK sites of the time — lastminute.com, Auto Trader, Ciao!, Norwich Union, PC Pro — many of which licensed search or directory data from the engines above.
Taken together, the whole thing looks like a wiring diagram. And in a sense, that’s exactly what it is: the wiring of early web search, before consolidation simplified the picture into something far more monolithic.
Looking back, the most striking thing about this 2004 diagram is how provisional everything feels. None of the relationships were fixed. Most of the companies were still experimenting. And the whole structure depended on assumptions that didn’t survive the decade.
Three big developments reshaped the landscape.
1. Crawling the web became cheaper
More companies could build their own indexes. Partnerships became less necessary.
2. Google’s model proved overwhelmingly effective
Cleaner interface, faster index, scalable business model. Portals stopped trying to compete.
3. Advertising consolidated
Overture’s pay‑per‑click model was clever, but Google’s integrated approach was simpler and more profitable. The old PPC networks faded.
Put those shifts together, and the old ecosystem simply couldn’t hold. The arrows on the diagram stopped making sense. One by one, the portals either adopted Google wholesale or faded from relevance.
Finding that old fan — and then this old diagram — was a reminder of how different search on the web once was.
The modern landscape is simpler, and in many ways more efficient. But simplicity has its own cost. Today, Google dominates both the search index and the paid advertising that sits alongside it. And of course AI driven results, a whole other story. Bing is there, and it matters, but the balance is nothing like the distributed ecosystem shown in that 2004 diagram.
Ah well - perhaps that’s why the rediscovery of a dusty fan on a hot afternoon felt oddly significant. It wasn’t just a piece of early‑2000s junk. It was a small, working fragment of a world that has since disappeared.