Yahoo: An Important Historical Milestone
August 5, 2025
Sorry, no smart software involved. A dinobaby’s own emergent thoughts.
I read “What Went Wrong for Yahoo.” At one time, my team and I followed Yahoo. We created The Point (Top 5% of the Internet) in the early 1990s. Some perceived The Point as a variant. I suppose it was, but we sold the property after a few years. The new owners, something called CMGI, folded The Point into Lycos, and — poof — The Point was gone.
But Yahoo chugged along. The company became the poster child for the Web 1 era. Web search was not comprehensive, and most of the “search engines” struggled to deal with several thorny issues:
- New sites were flooding the Web One Internet. Indexing was a bottleneck. In the good old days, one did not spin up a virtual machine on a low cost vendor in Romania. Machines and gizmos were expensive, and often there was a delay of six months or more for a Sun Microsystems Sparc. Did I mention expensive? Everyone in search was chasing low cost computer and network access.
- The search-and-retrieval tools were in “to be” mode. If one were familiar with IBM Almaden, a research group was working on a system called Clever. There were interesting techniques in many companies. Some popped up and faded. I am not sure of the dates but there was Lycos, which I mentioned, Excite, and one developed by the person who created Framemaker, among others. (I am insufficiently motivated too chase down my historical files, and I sure don’t want to fool around trying to get historical information from Bing, Google, Yandex, and (heaven help me! Qwant). The ideas were around, but it took time for the digital DNA to create a system that mostly worked. I wish I could remember the system that emerged from Cambridge University, but I cannot.
- Old-fashioned search methods like those used by NASA Recon, SDC Orbit, Dialog, and STAIRS were developed to work on bounded content, precisely structured, indexed or “tagged” in today’s jargon, and coded for mainframes. Figuring out how to use smaller machines was not possible. In my lectures from that era, I pointed out that once something is coded, sort of works, and seems to be making money — changes is not conceivable. Therefore, the systems with something that worked sailed along like aircraft carriers until they rusted and sank.
What’s this got to do with Yahoo?
Yahoo was a directory. Directories are good because the content is bounded. Yahoo did not exercise significant editorial control. The Point, on the other hand, was curated like the commercial databases with which I was associated: ABI/INFORM, Business Dateline (the first online information service which corrected erroneous information after a content object went live), Pharmaceutical News Index, and some others we sold to Cambridge Scientific Abstracts.
Indexing the Web is not bounded. Yahoo tried to come up with a way to index what was a large amount of digital content. Adding to Yahoo’s woes was the need to indexed changed content or the “deltas” as we called them in our attempt at The Point to sound techno-literate.
Because of the cost and revenue problems, decisions at Yahoo — according to the people whom we knew and with whom we spoke — went like this:
- Assemble a group with different expertise
- State the question, “What can we do now to make money?”
- Gather ideas
- Hold a meeting to select one or two
- Act on the “best ideas”
The flaw in this method is that a couple of smart fellows in a Stanford dorm were fooling around with Backrub. It incorporated ideas from their lectures, what they picked up about new ideas from students, and what they read (no ChatGPT then, sorry).
I am not going to explain what Backrub at first did not do (work reliably despite the weird assemblage of computers and gear the students used) and focus on the three ideas that did work for what became Google, a pun on a big number’s name:
- Hook mongrel computers to indexing when those computers were available and use anything that remotely seemed to solve a problem. Is that an old router? Cool, let’s use that. This was a very big idea because fooling around with computer systems could kill those puppies with an errant instruction.
- Find inspiration in the IBM Clever system; that is, determine relevance by looking at links to a source document. This was a variation on Gene Garfield’s approach to citation analysis
- Index new Web pages when the appeared. If the crawler / indexer crashed, skip the page and go to the next url. The dorm boys looked at the sites that killed the crawler and figured out how to accommodate those changes; thus, the crawler / indexer became “smart”. This was a very good idea because commercial content indexing systems forced content to be structured a certain way. However, in the Web 1 days, rules were either non existent, ignored, or created problems that creators of Web pages wrote around.
Yahoo did none of these things.
Now let me point out Yahoo’s biggest mistake, and, believe me, the company is an ideal source of insight about what not to do.
Yahoo acquired GoTo.com. The company and software emerged from IdeaLab, I think. What GoTo.com created was an online version of a pay-to-play method. The idea was a great one and obvious to those better suited to be the love child of Cardinal Richelieu and Cosimo de’ Medici. To keep the timeline straight, Sergey Brin and Larry Page did the deed and birthed Google with the GoTo.com (Overture) model to create Google’s ad foundation. Why did Google need money? The person who wrote a check to the Backrub boys said, “You need to earn money.” The Backrub boys looked around and spotted the GoTo method, now owned by Yahoo. The Backrub boys emulated it.
Yahoo, poor old confused Yahoo, took legal action against the Backrub boys, settled for $1 billion, and became increasingly irrelevant. Therefore, Yahoo’s biggest opportunity was to buy the Backrub boys and their Google search system, but they did not. Then Yahoo allowed their GoTo to inspire Google advertising.
From my point of view, Cardinal Richelieu and Cosimo were quite proud that the two computer science students, some of the dorm crowd, and bits and pieces glued together to create Google search emerged as a very big winner.
Yahoo’s problem is that committee think in a fast changing, high technology context is likely to be laughably wrong. Now Google is Yahoo-like. The cited article nails it:
Buying everything in sight clearly isn’t the best business strategy. But if indiscriminately buying everything in sight would have meant acquiring Google and Facebook, Yahoo might have been better off doing that rather than what it did.
Can Google think like the Backrub boys? I don’t think so. The company is spinning money, but the cash that burnishes Google leadership’s image comes from the Yahoo, GoTo.com, and Overture model. Yahoo had so many properties, the Yahooligans had zero idea how to identify a property with value and drive that idea forward. How many “things” does Google operate now? How many things does Facebook operate now? How many things does Telegram operate now? I think that “too many” may hold a clue to the future of these companies. And Yahooooo? An echo, not the yodel.
Stephen E Arnold, August 5, 2025
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
