Is Brilliance Enough?
As long as we have these sharp observers of Google in a conversation, please allow me to ask some provocative questions in hope that y'all will enlighten me.
Why is Google so dominant in search? Why and how did Google increase its share of the search market 7 percent in less than two years? For the past five years or so it seems that MSN, Google, and Yahoo all provide similar search quality and many of the same ancillary services. Why would a person in 2009 stop using Yahoo and start using Google? Is it a network effect? Is it cultural power (there is no verb "to Yahoo")? Is it the corona of goodness that emanates from the company? Or is there a clear qualitative difference that I have never seen demonstrated and thus do not believe?
Randall started answering these questions in a previous post, as he did in his book. Hal Varian asserts that Google is winning because it is demonstrably better. Is there evidence of this claim?
We can all agree that Google is generally brilliant. The really cool thing I learned from Planet Google is that Google is usually brilliant. When it's not brilliant, it's lucky. When it's not lucky, it's rich enough to make up for failing to be brilliant or lucky (see YouTube). But is that enough?
As far as my chief theme or interest, Randall hit it right on with his question about the complicity of Google users. In an effort to avoid the anxiety that must have overwhelmed Randall (and David before him), I chose not to write about the company per se. I wrote The Googlization of Everything about our relationship with the company. My goal is to convince readers not to be blind, trusting users of Google. Instead, I want folks to ask hard questions about user data, quality judgements, search policies, copyright, and the general power the company exerts over our daily lives.
The second place I am moving with my book is global. I am taking into account the various reactions to Google's grown (and failures) in other markets such as India, China, and Russia. This, I am sure, will be Google's most interesting challenge in the next five years. The game is pretty much over in North America.
For that reason, I am troubled by David's suggestion that anti-trust scrutiny (which is, after all, the law) would necessarily break up Google and cost thousands of jobs. The idea behind anti-trust is to make the economy work better for the broadest slice of citizens. Ignoring or failing to enforce the law is never a good idea, no matter what the economic conditions. And the resolution of such an investigation would almost certainly not break up Google. The chances are good that Google would come out completely clean. After all, anti-trust has been shrunk by almost 30 years of poor enforcement to be all about prices. That's a shame. But it's not going to change soon. Plus, since when is breaking up a monopoly bad for the economy? How would it possible cost jobs? The breakups of Standard Oil, Edison's Patent Trust, and AT&T all unleashed creativity and investment that benefited all. I would argue that the anti-trust conviction of Microsoft made the company more careful and opened up opportunities for Linux, Firefox, and others to chip away at Microsoft's dominant market position. No one had to break up Microsoft. But the conviction made a difference in the world. And it cost no jobs.
Google's place in the anti-trust scheme is unique and unprecedented, as James has pointed out. While its increasing control over the relative value of information is troubling in its own right, it's not necessarily in restraint of trade. The real problem Google is going to have is in its price-fixing behaviors among de facto compulsory licensing non-markets like Google Book Search or its "auctions" of terms in AdWords. I have heard much rumbling about some suspect things happening with the basement pricing of AdWords auctions.
There might be other causes of action Google will have to worry about. Now that Google ranks are essentially real estate on the Web, and many small firms rely on Google Docs, AsSense, and Blogger to make money, they are all at Google's mercy. Many people in recent months have reported having their accounts frozen or cancelled with no clear explanation or opportunity to appeal. This is a growing problem and could point to a serious "restraint of trade" on Google's behalf.
That's not a major economic problem. But it's worth watching.
The most likely answer to your first question is that Google has made it into the browser interface, at least in Firefox and Safari, so that doing a Google search is only a tab key away in a new browser window. Also, it is often easier to surf to a site you already know via this method, than to actually type in the URL, if it's really long, confusing or somehow forgotten.
January 15, 2009 6:47 PM | Reply | Permalink
"Hal Varian asserts that Google is winning because it is demonstrably better. Is there evidence of this claim?"
What ignorant gas-baggery. Yes, as a matter of fact, there is such evidence. It's called "blind side by side comparison," and it's the standard method for evaluating multiple competing human information sources. The big three all do them against one another all the time. To see Microsoft's take on the issue, go to, say, http://blogs.technet.com/extreme/archive/2008/11/21/live-search-vs-google-search-vs-yahoo-search.aspx and follow the links he gives.
January 15, 2009 10:21 PM | Reply | Permalink
I'm shocked! You mean a blogger, who is a Microsoft employee, using the MS evangelical in-house blogging platform believes that Live Search is a better search engine the Google? Every link in that webpost, can be easily tied to Microsoft. The two which are not absolutely transparent are
mysearchoff.com
and
Searchvote.com
mysearchoff.com is regisistered to Brian Gorbett, who is an Architect in Microsoft's Developer & Platform Evangelism ( DPE) Division.
Searchvote.com is registered to Microsoft.
ROTFLMAO, and I even posted a response to this thread in which I stated that MSN and MS Live are comparable in quality to Google.
So is TPM user: "Anonymous TPM reference" a Microsoft sock-puppet?
January 16, 2009 1:59 AM | Reply | Permalink
That did seem a tad obvious, PCA.
January 16, 2009 2:20 AM | Reply | Permalink
One factor in Google's success was fortuitous roll-out when the net had become more than a curiosity, and place for computer geeks to play, had reached a critical-mass user-base about to take out exponentially, yet still geek predominated.
Long-time computer users felt very comfortable around Goggle's heavy textual, no frills user interface. No slow loading non-functional eye-candy, and it played fair across OS'es and browsers. It was a break-way from the limitations of pigeon-hole search categorisations, and holdover command switch esoterica from pre-WWW net adventuring via USENET, gopher wais and ftp.
The founders were geeks too. The internet had by then become big enough that information was becoming increasingly difficult to locate in a timely manner, unless, of course, one was looking for sexually explicit content, because that would pop-up in search results, using unexpected correlations. It was near impossible to filter out without also blocking significant avenues of potential data. Google's use of in-bound links for weighting search record placement immediately knocked the worst of the spamming web sties off the first several pages, and shuffled up the way search results were ordered. Suddenly, overt human bias was removed. Google set their algotrithms loose, and let them do it. WebWizardry suddenly became conceptually deistic. The watchmaker web gods who constructed it, and then let it go free.
Goggle was quickly embraced as default search engine by web aware geeks, who were at that time the friend and/or relative to call when confuse about the web, or net access had been screwed up. Google was the text based search engine easiest to explain to non-techies. Have you ever attempted to explain boolean logic switches and the inherent differences implied by punctuation delimiters to persons who had to take bone-head math and then algebra before being allowed to almost flunk out of 1st year Calculus? Have you ever been frustrated by an egghead's brevity answering a question, because s/he is incapable of understanding that the six or seven logic steps they just jumped over was transparently clear to you? Google eliminated major variables on both sides of that equation with their default AND between terms. It's pretty easy explaining what quotation mark bracketed text does. Those two functions alone go a long way towards effective web-searching, and Google's advance search interface had clearly labeled forms, which didn't take a textbook to figure out.
Geeks got downright evangelical about Google, and this was a significant reason why it surged ahead of preexisting engines. Other engines either evolved or they died.
I do not think Google's search results are presently significantly better than Microsoft's MSN and Windows Live. Yahoo would also be a contender if they didn't load up their home page with potentially seizure causing blinkity blinking widgetry as distraction. A few years ago Google algorithms got seriously jacked by blog-rolls and link exchanges, which injected far too much noise into the search records. They were very slow inplmenting a fix for it, and when they finally got around to it, there was loud lametations, hair pulling and wails of unjustice wafting up from the blogs that had been the miost sucessful cheating PageRank JuJu. There is now a constant battle between Google developers and seekers of SEO fame. As Google tweaks their algorithms in response, new circumventions are conceived, and many of the newer web-portal (TPM included) drive PageRank and page views with never-ending streaming of new content, with low persistence. Most of these same websites (TPM not included) publish low-end rehased trash that has a temporary lustre of cheap and tawdry pretties. I really do not care to know how many times a dim-witted diva has been photographed sans-a-panties getting out of a low slung limo. Website profitability now has little to do with well-organised quality data, but instead with content fresh and constantly refreshed. There are even blog and CMS tools that can be configured to automatically scrape data from other web sites, using pre-chosen topic parameters, that then do nothing more than republish the same data. It's reformatted aggregation of other people's original content.
If the web is going to continue evolving towards better and bigger data accessibility, it needs to be yanked away from the short-attention span transactions. It will devolve into nothing more than a tsunami of twittering tweets and circle-jerked content aggregation with no-value added in the middle-man reprocessing. Spam is preferable to Velvetta and cheeze-whiz. This is where Google is failing, and if they do not evolve, they will also die.
January 16, 2009 1:06 AM | Reply | Permalink
That's a wonderful summary and analysis. Thank you very much.
January 16, 2009 4:11 PM | Reply | Permalink
welcome, thank-you for complementing a ravening streamed-riff. been poking around your web pointers briefly over the last day. "The future of The Book" is a fascinating topic, having many intersections with subjects i often muse about. here's a bit of a heads-up: if you receive any mail sent from an account that is an obvious play on my TPM pseudonym, it's almost certain that i sent it. as far as i know, i am the only person who has used this pun as a pointer, but have used several variations since conceiving it. some of them begin with 'suedoe', because the 1st time, i discovered that many people made gender assumptions based solely upon it, and this, aside from being amusing, can offer valuable insight, as well as providing exploitable opportunities for acts of creative trolling. we all have embedded gender biases, no matter how egalitarian we have struggled to make our perspectives. i still do not understand why 'suedoecyants' would come laden with a presumption of femininity; it's only a cheap pun.
January 16, 2009 9:59 PM | Reply | Permalink
This is very funny, and one of my favorite things to observe on sites like this.
January 16, 2009 10:39 PM | Reply | Permalink
Thank you for putting words to some things that are really depressing me recently about changes in internet culture. I know I am not imagining it--it was always an inherent threat lurking in the background, but it seems to me that actualization has really escalated a great in the last year or so. Doesn't it partly have something to do with how google makes money--the whole ranking thing--and could it be that the lowest common denominator is finally kicking in now because of that combined with growth of use? (I'm old enough to know that this is sort of similar to what happened with television. :-))
January 17, 2009 12:13 AM | Reply | Permalink
careful artappraiser, or you're liable to get me drifting into post-digital visions. some of it leads to very dark futures. the digital-age came with wonderful technology, and has greatly expanded data exchanges. it should never be perceived as an end though. it has downsides, and at its foundation is irreparably flawed. all cannot be condensed down to collections of 2-position switches. that is boolean deuce dupery.
you are surely aware of what is lost when digitalising images of paintings. human senses are inherently analog peripheral devices, which provide I/O for our minds. sensual dimensionality is lost in digitalised transformations. images from a film negative can be enlarged beyond recognition, but it never becomes recognisably pixilated. sound is four dimensional, spatial waves travelling at a fixed speed down timelines. our ears are also multi-dimensional receivers. no matter how good the filters get, digital sound is still flat in comparison. at some level, transforming analog media to digital can never be lossless, and we should care about what has been lost.
January 17, 2009 1:56 AM | Reply | Permalink
Lowest common denominator is somewhat apt, but not exactly. Google's pageranking is more of a incredibly complicated popularity contest, where the contest judges' vote value is itself a function of their popularity. Initially, the concept was outstanding, and its underlying axioms strong. The idea was to remove human bias from internet searches as much as possible, yet still return high-value records at the top of the list.
This was implemented by analysing the links pointing to content, and collating the data into a format easy referenced in searches. If some asked about "A", then the links on web pages which discussed "A", pushed the linked content up in the rankings. The more external links that point to content on a website, the higher the site's credibility, and the higher a site's credibility, the higher the value given to the content that it links to. This is still a big gloss, and there are other factors. Google doesn't use open-source algorithms, and the factors are reversed engineered best-guesses. It does seem that a webpage with just one hyperlink on it confers greater pagerank to its recipient, than each individual recipient on a page with more than one hyperlink on it. Some have postulated that the calculation is made with straight-forward simple division, but I am inclined to think that the calculation is much more complex, and that high pagerank pages with many hyperlinks on it end up with a sort of synergistic bump in the sun-totalof pagerank they confer. This is why I often use magik terms when talking about search engine placement. Maybe Aleister Crowley would be a master at SEO, were he still alive today.
Initially, this pagerank process worked very well, and Google searches were by far the best on the web. Some people naturally try to cheat this process. At first is was link farm pages or sites. Google factored this into their algorithms, and their bots learned how to recognise and ignore them. Some bright individuals realised that strategic use of anchor text could affect search results. This was how the search term "miserable failure" ended up pointing to GW Bush's official White House bio, as well as a few other well-known jabs, and many lesser ones. For a while, "French military victory" pointed to a fake Google parody page that said there were no pages returned on that search, and queried if they might have instead meant, "French military defeat". Google resisted the calls to change their alogrithms, so that this could not be exploited, but at first Google said that it did indeed represent the will of the net citizenry, even if expressed sarcastically. It was only effective if the search term was not one likely to be used frequently and/or the number of people using that exact term in anchors pointing to the desired target were numerous. At one point, Tony Blair's government bio page was sitting about fifth in a simple search for "poodle". That was incredible, considering how many internet pages are about poodles. As this hack became more and more exploited, sometimes for commercial or non-political slanderous purposes, Google finally implemented a patch to fix it.
At the onset of the blogging craze, people figured out that if bloggers linked to their friends, and those friends linked to them, they could elevate each other up in the search returns, and the blogroll was born. The quality of Google searches degraded quickly, once this became well-known, but at first, Google simply ignored the criticism. A fairly well-known Brit tech journo writing for The Register and living in San Fransisco, Andrew Orlowski, was a primary force in getting Google to adjust for it. Orlowski believes himself to be some sort of dragon-slayer, and his work is often saturated with strong influences of Brit sarcasm. He is still vilified by some old established bloggers who awoke one morning to find their pagerank had dissipated into the thin air after Google had changed it algorithms. Orlowski remains one of the best tech journos who publishes Goggle critiques. He certainly does not suffer from revisionary deification.
Google has taken a more proactive stance against their searches being jacked, and every time I've taken the time and made the effort to research rumours of Google unfairly stripping a site of its pagerank, it turns out to be that the site was attempting to cheat, even though in several instances, the sites' owner(s) were too stupid to realise what they were doing was cheating.
Pageranking has value for use in internet search placement, but it can no longer be expected to return relevant results by itself. There are just too many ways for exploiting it unfairly. Google may not let human bias from their end interfere with the search results, but they need to implement some major changes in their ranking process.
January 17, 2009 3:06 AM | Reply | Permalink
Why is Google so dominant in search?
1. Google, when launced, immediately got every Linux/Open Source-minded person IN THE WORLD bheind it.
2. Many journalists simply loved Google from the start as well; Google was always getting great press.
3. Wall Street loved Google too, always pumping it up.
4. Google gave good search without tons of horrible grapics and javascript crap that search engines like Yahoo had/have.
Etc.
It is so obvious why.
Google might not always be a leader (as it keeps on overreaching) but it will never die.
January 16, 2009 9:16 AM | Reply | Permalink
Why would a person in 2009 stop using Yahoo and start using Google?
Well, I stopped using Yahoo and started using Google in about 2007/2008. Originally, I preferred Yahoo to Google.
I started switching for multiple reasons.
Google Books and Scholar. One of the biggest reasons. For awhile, I'd still use Yahoo and Google.
Then started using Firefox. Home page. Started being easier just to use Google for everything.
They added Blog Search, which I like.
Finally switched emails from Yahoo to Gmail. The Yahoo email was my original email, almost 15 years old. It was a mountain of spam every day. Switched to Gmail. Also like the IM feature on Gmail.
Did creep me out quite a bit when I realized the ads on the sides of my email were about the text of my email.
Do agree with Pseudo on this: when searching for something to find half the results are blog posts consisting of a link to the same article is annoying.
Also, being able to text Google for phone numbers/directions/weather/Steelers scores is brilliant.
January 16, 2009 9:48 AM | Reply | Permalink
Hilarym, don't take this personally. I am wondering why you use Google Search as a home page. I've seen others mentioning it in this week's Google book topic, and it actually surprised me. I suspect it one of those things my inherent geek factor perceives as dissonance.
When you first connect to the internet, is your primary objective to initiate a web search? Do you usually visit a few favorite sites? Do you scan current news, and if so, is Google News your first stop doing it? Wouldn't Google News be a better home page then? I don't understand the appeal of Google search as a home page, other than it provides quick content-neutral jump using the home button. One of the reasons this is odd to me is that my home page is an html on my computer that has links to sites I frequently visit. those links include several different search engines. This is part of the geek quotient though, because I am as comfortable writing simple html as I am writing plain text. I still have a hard time understanding why other people seldom do this. It's relatively easy to do with free-software WYSIWYG html editors.
Again, I am not criticising you; instead trying to understand why people remain complacent about their browser's home page. I hadn't realised that a browser's install default home-page significantly increased visits to the website in the long-term. Changing the home page default is one of the first things I do when installing a new browser.
January 16, 2009 9:28 PM | Reply | Permalink
LOL! I definitely don't take offense. I had to think about the answer though.
It would probably make more sense for me to make TPM my home page, it's usually the first page I go to. Or sometimes email. Which I can get to easily from the Google home page. But I always used to change my home page first thing also, and wonder why people left it at the so-simple Google home page. I hadn't thought about it since then. I realize why now though - I think it's mostly due to tabbed browsing. I really don't spend any time at all on the home page, and never go back to it when I am deep into the Internet. I usually have many, many tabs open. When I switched to Firefox, it was the first tabbed browser I used. So it was just as easy for me to hit Control/T and then type search terms into the address bar, which automatically used Google. One of the things I really liked about Firefox was that you could just type TPM (or whatever) and it would take you straight to the site, rather than through search results first.
I switched to Apple/Safari a few weeks back and still use it the same way, only now I search in the Google toolbar. I don't know how/if you can use the address bar in Safari the same way as Firefox. If the automatic search bar in Firefox or Safari was Yahoo or another search engine, I'd probably use that one most of the time. Though I have started to take a preference to Google searches, though I can't precisely explain why.
This part is hilarious to me though:
That sounds like days of work to me!! Did you have to register a website to do that?
With the address bar autofill, it's just as easy for me to open a new tab and type the first letter(s) of wherever I'm going and it pops up.
January 16, 2009 10:36 PM | Reply | Permalink
The only real work is deciding which WYSIWYG editor to use, and learning its ropes. No, my home page is literally Home, resides on my computers main hard drive, in a folder appropriately titled, "local".
Once you are comfortable with an editor, it becomes as easy as a TPM blog post to create and edit. Just make an html page with a list of your usual suspects web sites, save it to a local file, open that file up in your browser, and set it to be your home page. No need to worry about web standards or other such nonsense that comes with publicly served pages. all that matters is that it display nicely for you in you own browser, and has links to the sites you often frequent. After that, whenever you want to edit it, open it up in the html editor, and copy/paste away.
My bookmark file is an almost 2mb monster, btw. It's overused as it is, but converting it to several local pages of categorised links, is always an intended task on the back-burner.
January 16, 2009 11:33 PM | Reply | Permalink