Google and Search: A Fix or a Pipe Dream?
September 6, 2024
This essay is the work of a dumb dinobaby. No smart software required.
I read “Dawn of a New Era in Search: Balancing Innovation, Competition, and Public Good.”
Don’t get me wrong. I think multiple search systems are a good thing. The problem is that search (both enterprise and Web) are difficult problems, and these problems are expensive to solve. After working more than 50 years in electronic information, I have seen search systems come and go. I have watched systems morph from search into weird products that hide the search plumbing beneath fancy words like business intelligence and OSINT tools, among others. In 2006 or 2007, one of my financial clients published some of our research. The bank received an email from an “expert” (formerly and Verity) that his firm had better technology than Google. In that conversation, that “expert” said, “I can duplicate Google search for $300 million.” The person who said these incredibly uninformed words is now head of search at Google. Ed Zitron has characterized the individual as the person who killed Google search. Well, that fellow and Google search are still around. This suggests that baloney and high school reunions provide a career path for some people. But search is not understood particularly well at Google at this time. It is, therefore, that awareness of the problems of search is still unknown to judges, search engine marketing experts, developers of metasearch systems which recycle Bing results, and most of the poohbahs writing about search in blogs like Beyond Search.
The poor search kids see the rich guy with lots of money. The kids want it. The situation is not fair to those with little or nothing. Will the rich guy share the money? Thanks, Microsoft Copilot. Good enough. Aren’t you one of the poor Web search vendors?
After five decades of arm wrestling with finding on point information for myself, my clients, and for the search-related start ups with whom I have worked, I have an awareness of how much complexity the word “search” obfuscates. There is a general perception that Google indexes the Web. It doesn’t. No one indexes the Web. What’s indexed are publicly exposed Web pages which a crawler can access. If the response is slow (like many government and underfunded personal / commercial sites), spiders time out. The pages are not indexed. The crawlers have to deal in a successful way with the changes on how Web pages are presented. Upon encountering something for which the crawler is not configured, the Web page is skipped. Certain Web sites are dynamic. The crawler has to cope with these. Then there are Web pages which are not composed of text. The problems are compounded by the vagaries of intermediaries’ actions; for example, what’s being blocked or filtered today? The answer is the crawler skips them.
Without revealing information I am not permitted to share, I want to point out that crawlers have a list which contains bluebirds, canaries, and dead ducks. The bluebirds are indexed by crawlers on an aggressive schedule, maybe multiple times every hour. The canaries are the index-on-a-normal-cycle, maybe once every day or two. The dead ducks are crawled when time permits. Some US government Web sites may not be updated in six or nine months. The crawler visits the site once every six months or even less frequently. Then there are forbidden sites which the crawler won’t touch. These are on the open Web but urls are passed around via private messages. In terms of a Web search, these sites don’t exist.
How much does this cost? The answer is, “At scale, a lot. Indexing a small number of sites is really cheap.” The problem is that in order to pull lots of clicks, one has to have the money to scale or a niche no one else is occupying. Those are hard to find, and when one does, it makes sense to slap a subscription fee on them; for example, POISINDEX.
Why am I running though what strikes me as basic information about searching the Web? “Dawn of a New Era in Search: Balancing Innovation, Competition, and Public Good” is interesting and does a good job of expressing a specific view of Web search and Google’s content and information assets. I want to highlight the section of the write up titled “The Essential Facilities Doctrine.” The idea is that Google’s search index should be made available to everyone. The idea is interesting, and it might work after legal processes in the US were exhausted. The gating factor will be money and the political climate.
From a competitor’s point of view, the index blended with new ideas about how to answer a user’s query would level the playing field. From Google’s point of view it would loss of intellectual property.
Several observations:
- The hunger to punish Big Tech seems to demand being satisfied. Something will come from the judicial decision that Google is a monopoly. It took a couple of decades to arrive at what was obvious to some after the Yahoo ad technology settlement prior to the IPO, but most people didn’t and still don’t get “it.” So something will happen. What is not yet known.
- Wide access to the complete Google index could threaten the national security of the US. Please, think about this statement. I can’t provide any color, but it is a consideration among some professionals.
- An appeal could neutralize some of the “harms,” yet allow the indexing business to continue. Specific provisions might be applied to the decision of Judge Mehta. A modified landscape for search could be created, but online services tend to coalesce into efficient structures. Like the break up of AT&T, the seven Baby Bells and Bell Labs have become AT&T and Verizon. This could happen if “ads” were severed from Web search. But after a period of time, the break up is fighting one of the Arnold Laws of Online: A single monopoly is more efficient and emergent.
To sum up, the time for action came and like a train in Switzerland, left on time. Undoing Google is going to be more difficult than fiddling with Standard Oil or the railroad magnates.
Stephen E Arnold, September 6, 2024