| M | T | W | T | F | S | S |
|---|---|---|---|---|---|---|
| « Jun | ||||||
| 1 | 2 | 3 | 4 | 5 | 6 | |
| 7 | 8 | 9 | 10 | 11 | 12 | 13 |
| 14 | 15 | 16 | 17 | 18 | 19 | 20 |
| 21 | 22 | 23 | 24 | 25 | 26 | 27 |
| 28 | 29 | 30 | 31 | |||
Courtesy of the latest skinny on Google’s mammoth datacenters.
Regarding their infrastructure my favorite quote continues to be: “Server makers pride themselves on their high-end machines’ ability to withstand failures, but Google prefers to invest its money in fault-tolerant software.”
And in the long run, that probably will be their greatest competitive advantage.
In the past six weeks I’ve conducted at least 1802 searches (according to my history cache). And it’s not all for super sexy helicopter porn or aesthetically unappealing crocks.
I mention this because today Sergey Brin mentioned that one of the reasons Google has so-far weathered the economic downturn:
Let me highlight search improvements in last 90 days. have launched more than 100 improvements in search quality. New tailored home pages in international markets, such as Japan. Also better job in foreign countries selecting domestic results.
I can only imagine how many searches someone like Justin Raimondo does (his articles are, in the words of Gene Callahan: link portals).
Speaking of which, I haven’t written a single new article in that time frame (the 3 previous Mises articles this year were all written in the first week of February)… so my quarterly numbers are skewed!
Incidentally, I sat in a noisy internet cafe next to several middle-aged chain smokers (who were farming gold in WoW) for a solid week between teaching hours to whip the Pulitzer-prize winning pieces together. The green tea and steroids helped.
As an aside, unless it was Scarlet Johannson in a bikini, I don’t think I clicked on any web ad. Who clicks on ads?
Remember that cludgy old Newton? I had one for a year, back in 1995, and found it be of little use — aside from its inherent paperweight abilities and the obligatory game of Tetris during geometry class.
Now that Apple has released the anticipated iPod Touch, which is basically the iPhone sans phone-abilities, is this basically what a PDA should be like? Or is it just the opposite?
As far as the phone capabilities go, since it has WiFi built into it, all some enterprising individual(s) needs to do now is create a VoIP hack that allows you to do the unthinkable. Right?
Skype is already available in a duct-tape fashion for the iPhone. And some guys at Google (and Facebook) have helped spearhead the development of useful iPhone apps (despite rumors that Google itself developing an OS for the mobile segment).
I plan on grabbing the Touch within the next week or so and am crossing my fingers that a hack will be available shortly (so much cheaper and convenient for international calls…).
The only thing I wish that had been included was a digital camera, even a dinky one. [I already have a phone and don't care for virtual keyboards (I played around with an iPhone before I left the West, I'll stick with a meatspace QWERTY)].
In the past, I have helped maintain several pages on Wikipedia for various organizations. In addition to adding new content, most of the work was simply geared towards keeping vandals at bay.
Reinventing the information wheel
I believe the carte blanche policy of giving anonymous users the ability to edit every page should be changed. Some of these changes could be as simple as requiring that users sign on through a captcha, IP-stamped system. Or even easier, create a new function within the Wiki software, to where anyone can “suggest” changes, but these changes will not go into effect without the approval of a moderator.
In fact, a variation of this last example is currently being tested by the German edition of Wikipedia and is expected to be implemented in other editions within the coming months.
The online edition of the Encyclopedia Britannica has also adopted this “suggest a change” into their system (example), and I would suspect that this has helped them further refine their otherwise top-notch articles.
However, despite this change, based upon my own unscientific observations, Britannica is not even close to winning the all important traffic-based popularity contest. For instance, whenever I use a search engine, Wikipedia articles rank much higher than those of Encyclopedia Britannica (in fact, I honestly cannot recall ever seeing an EB result on the first SERP).
And while some may say each of these organizations is trying to cater to a different clientele, that would be nonsensical in this age of dying gatekeepers and open access. [Note: another article could discuss the differences between the structure of article layout between the two behemoths; accuracy is another issue altogether]
The Man vs. Fanboys
While Wikipedia currently enjoys holding the lofty status as the alternative to the Establishment - the Corporate world that Britannica supposedly represents - the capital resources expended in maintenance and upkeep come, by-and-large through donations by firms such as Google and Yahoo. So the “free” nomenclature is not entirely accurate (kind of similar to GNU/Linux vs Microsoft and the economic opportunity costs in development).
However, arguably the most important change EB could implement for increasing traffic and thus increasing potential ad revenue is to open up their Walled Garden (they charge for “full access”). Answers.com makes a killing this way, Encarta has also adopted this model, and The New York Times is toying with this as well. They could even strike a monetary deal with a search engine to integrate and improve upon their lackluster offering.
If EB wanted to stay in the world of academia (which their area for “suggestions” leans heavily towards), they could set other restrictions in place, such as: limit contributors to .edu addresses, have several moderator/contributor levels based upon peer karma, longevity, etc.
See also: 1 2 3 and Murray Rothbard’s criticism of the theory linear history advocated by Whigs.
Note: Michael Arrington has some thoughtworthy (sic) comments regarding Nupedia, Citizendium and Digital Universe.
And as a last aside, structurally EB could branch out and perhaps set up a business akin to Techdirt’s information consulting service.
I have it.
Here is a brief description of what it is.
Google’s “advanced” search options give a user the ability to fine-tune health-related search terms. See also their useful co-op program.
I played around on Google’s Image Labeler a little bit more since it launched 24 hours ago.
Here are the results thus far:
You get 100 points for everytime you and your random partner successfully tag a photo. Thus, I’ve labeled 129 photos so far.
The overall leader board looks like this:
If you do the math, Google has nearly 6000 images tagged by the top 5 alone in addition to at least 125,000 other photos tagged by 1000 people.
So exactly how many people have labeled images now? I decided to change my identity by logging into a different GMail account as RickyBobbyNASCAR. As seen here:
After labeling 76 photos, my alter ego is still not ranked individually for one reason or another, so it is difficult to say how many people are currently using this (maybe the threshold contains only the Top 1000).
I do have to say that some of the people that label pictures have no clue about geography. A map of a tropical storm was shown with a path projected to hit the eastern coast. Kentucky was not on the map, yet it was listed as an “off-limit” word… which usually means alot of people have typed Kentucky as a label. Also, Wink was the best teammate by far (kept bumping into him).
Note: regarding numbers, Luis von Ahn (creator of this system) estimated that it would only take 2 months for Google to properly tag all of their hundreds of millions of photos with a paltry 5000 users. In 24 hours alone, they have done at least 131,000.
Not too shabby.
In my previous post I mentioned the automation and assembly of seemingly tedious tasks into a useful product.
Turns out that about a month ago, a guy by the name of Luis von Ahn visited Google and gave a talk to a group of its employees. Luis is the creator of a computer game called the ESP Game, which does exactly what Google Image Labeler does (though after playing both, I think his version has a better layout/design).
His talk is very effective, as he really makes you think outside the box in terms of spinning mundane tasks into productive creations (see for instance, his statistics on how much time is wasted playing solitaire).
Addendum: according to Danny Sullivan, Google recently licensed the technology from von Ahn for their own uses.
Do you recall the old timey Tom Sawyer character from the novel’s of Mark Twain?
Remember how he managed to get a group of his friends to paint a fence for him?
Somehow, Google has managed to do the same things with its ginormous database of images through a new program called Google Image Labeler.
The idea is simple, yet ingenious. Google has a system in which you gain irredeemable points for every image you tag.
You log into their system and label images with a few keywords. And you are also paired off with a random person as well. If you think about it, this helps diversify the pool of words that can be used, and it can help highlight users whom are merely pranksters.
The bottom line for Google is that they get to use the labor of its zealous fan base for free, whereas Amazon pays users to do the same thing.
Brilliant.
Business 2.0 has a really good write-up of all the blog-enterprises that are now funded through advertisements.
The main reason for their success: these blog cater to specific niches, thus the ads can be targeted to specific demographics much more effectively. It is the Long Tail of advertising, what Chris Anderson calls “nichification.”
Last week I mentioned that AOL accidentally released millions of search terms and the corresponding user ID to them.
Some nerds at SomethingAwful compiled numerous lists of very disturbing search requests from various users. And I mean disturbing (see the second to last one).
One wonders when someone will write some kind of existential book or movie about how search terms reflect some kind of Freudian self. Not it.
You can download the entire search log here.
Remember earlier this year, when the Department of Justice demanded that search engines turn over logs of over 1 million URLs?
Remember how Google resisted and countersued?
Remember how the Judge ruled that the DoJ could only have 50,000 nerfed results?
Well, someone at AOL haphazardly released the search results of over 650,000 quasi-anonymized people into the public domain a couple days ago.
Oops.
Markus Frind explains why this is not a good thing for Google, the same company that recently paid AOL $1 billion for exclusive rights at powering search across all of AOL’s properties.
Update: the New York Times actually managed to track down one of the users and interviewed her.
I’ll be honest, I don’t care much for the Google Reader. Here’s my brief review of it.
After a solid year, I still have stuck with Bloglines (despite trying others such as Rojo).
However, some people do like it and the gReader team has recently provided some interesting aggregated numbers for our consumption.
Note: these are extensions to the namespaces only — they only looked at things like how often a Creative Commons license is embedded into a feed (and not if it was RSS/Atom).
Via Niall Kennedy.
Exhibit A: This video of some people that really should not be wearing skin-tight clothing, singing about some kind of granola crunching nirvana that has never existed.
Look, “the Internet” is just a gigantic collection of independently run networks. Some are privately owned, some are owned by NGO’s and others are financed via taxes.
There are over 150,000 routers that relay packets back and forth, many of these are owned by large companies.
Some of these companies want to charge variable rates — and already are charging variable rates. It is a concept that is neither new, nor foreign to the service world. For instance:
The problem with the telecom industry however is that it is a highly regulated industry. It is not the product of a free-market distribution of resources. And none of the “net neutrality” legislation will fix it.
Rather than deregulate the industry, “net neutrality” will only add more costly and inefficient layers of regulation and oversight to it. After all, oversight and execution of the policies have to be financed somehow, guess who gets to pay for that?
What costs money? You see, in order to make sure every packet is living in some kind of Egalitarian-world, equipment will need to be installed to monitor and analyze the packets. And don’t forget the new civil servant positions that must be erected to analyze the data reports. Guess who finances that now? And so on.
Having a hard-on for Google and other web services is not going to nullify the laws of economics, namely scarcity. And any kind of nationalization of the network will result in a terribly ineffective method at distributing scarce resources.
If you are frustrated with dropped or static-filled calls today, imagine having a dropped call in the future because network operators can no longer discriminate (oooo, an evil word!!!) and reallocate bandwidth based upon peak usage.
Chuck Norris really should round-house kick this fallacious argument into the recycle bin. Bring back the freedom of contract.
Several months ago Paul Kedrosky lamented the fact that Google was merely a one-hit wonder, with all of its revenue eggs in one proverbial basket based on search-related queries. And the data still continues to be on his side.
He also mentioned that they should move towards releasing polished products instead of merely launching them as soon as they are created (or soon thereafter).
When Seth Godin visited the firm earlier this year, he specifically addressed this point (see the Q&A session). Godin suggested that rather than trying to be the “first mover” in every product category, that they should instead work on creating a stable, robust product that would be more than a flashy, whiz-bang ornament.
At this point I do not think the blame rests on one sole individual. However, based upon the way the firm is currently organized, I think that formally institutionalizing gatekeeper positions like Marissa Mayer actually works against them in the long-run.
The centralized nature of their product manager suffers from, among other things, asymmetrical information — a knowledge problem that curtails effective decision-making and stymies prompt execution.
To that point, last year, News.com put together a good overview contrasting the culture differences between Yahoo and Google. They noted that because Google was leaner, flatter and a more decentralized organization, they were more effective at rolling out creativer (sic) solutions.
However, based upon their recent history, might they become the very Yahoo they supplanted? Today, Steve Rubel posits the same question.
One solution could be a reorganization along the lines of several smaller, confederated start-ups. Example: delegating task-management and dispersing this vaunted focal authority into small cells of semi-autonomous, entrepreneurial units each responsible for their own P/L (see this facsimile of “Unbundling the Corporation”).
Deep Thoughts with Jack Handey
How big is too big? Do the leveraging abilities of traditional public firms (financial resources, clout, etc.) outweigh the long-term costs (e.g. salaries, pensions)?
In an age where you can reliably and securely contract any task to the ends of the earth at a fraction of the domestic cost, do web companies need to go public anymore… let alone receive venture capital? And do you really need to have people from Mensa working for you (see also the seminal works of Jim Collins; in BtL he suggests no)?
If you’ve grown tired of trying to organize your bookmarks at each computer you use, perhaps you might be interested in two handy services.
One is Deli.cio.us which is now operated by Yahoo. The way it works is fairly simple. You create an account and install a little utility that integrates with your browser. Whenever you come across a site you that you want to bookmark, you simply click the “tag this” button and it instantly becomes immortalized in your ever growing cornucopia of links. You can also look at what other people are bookmarking and build RSS feeds off of specific tags (e.g. technology, golf, base jumping). See mine.
The other popular one is Notebook from Google. It operates in a similar manner, yet it is not only more substantive, but also influences the underlying ranking system of each site that is “noted.”
Watch the video. Remember, that guy is currently the CEO of Google.
Moral of the story: one giraffe is suspicious, but two is down right incriminating.
Via Valleywag
SFGate scored some pictures of Google’s new facilities being constructed in Oregon off of Interstate 84.
The first comment really tickled my fancy:
They are obviously planning the next version of Oregon Trail.
One of the popular memes being discussed today in the blogosphere is the seemingly contrarian statement from Eric Schmidt of Google regarding click fraud. And like usual, The Stalwart uses an interesting analogy to better understand Schmidt suggesting that click-fraud is by-in-large a self-correcting process.
My question is, what is an objective definition for click fraud? And can this be tied into time-shifting technologies used by TiVo and MythTV users to by-pass advertisements altogether?
The basics
The way AdWords and Overture are currently setup is through an auction system on keywords. An advertiser will bid for keywords that are then displayed on participating websites. The advertiser is then charged every time their ad is clicked based upon the auctioned off price point.
Individuals like Mark Cuban have suggested that the current PPC model is flawed due to massively underreported click fraud. While automated bots and unscrupulous competitors may be the easiest to identify (e.g. tracing IPs and click patterns), how do you measure the intent of a John Doe web surfer? Is it quantifiable with existing technology?
Measuring intentions
The biggest problem in assessing the size of this problem is measuring it in some objective way. As of right now Google and others have fraud departments and use a series of algorithms to identify and measure questionable click patterns from suspect IP addresses.
But how do you really measure the intent of any web surfer? What is a legitimate click versus an illegitimate one? As of this writing, I am unfamiliar with any technology that can scan the mind of web surfers to discover their preferences and intentions. Assuming none exists and ignoring the identifiable fraudsters mentioned above, how can anyone claim that there is massive fraud being perpetrated?
Rumor or not, one theory proposed by John Battelle for why Google did not purchase MySpace when they had the chance was because of really poor conversion rates. The thought is, while there may be millions of users accessing the site each day, the quality of clicks is deplorable (i.e. they click on everything whether they care about it or not and thus cost advertisers millions, see Dan’s comments).
Caveat emptor
Chris Anderson, editor-in-chief of Wired magazine, popularized the term “The Long-Tail” in an effort to describe how companies such as Google and Overture effectively paired advertisers with potential customers through targeted keywords.
The holy grail of targeted marketing is catering each ad to the preferences of each individual (Vistas Unlimited in Plano has made some great advances in this area). These web firms essentially paired their large databases of keywords with an auction system which allowed advertisers to target their ads more effectively and with a potentially higher RoI.
This was in stark contrast to the shotgun approach used by incumbent search companies like Lycos and Alta Vista whom simply plastered ads comprising the most commonly used keywords (e.g. sex, mp3, britney spears). And when the dust settled, the keyword auction method became the de facto standard.
Interruption Advertising
Seth Godin, among others, has done a good job explaining the need for a change in how advertising is done. He coined the term “permission marketing” which is essentially an opt-in approach, one in which the marketer asks permission to engage potential clients.
Over the past several decades the traditional advertising approach used in media such as radio and television, was typically done by interrupting a program (or pop-ups in the case of the internet). Due to a shift in viewing habits (e.g. DVRs) and demographic groups, advertisers have now been forced to change (e.g. subtle product placement) and have used technology to further fine tune where ads are run in the first place (e.g. geo-location).
I mention this because Google and Yahoo have successfully used a toned down method of displaying advertisements that minimize disruptions while you search and visit websites. Their use of clearly defined advertisements also sets the tone for consumers, letting the customer ultimately initiate the sale-cycle (see also Pay-Per-Call).
Evolution schmevolution
How you target and who you target ads to — these are a yin-yang interconnected process that is continually being refined. And I suspect that much of what is being blamed on “click fraud” can be attributed to an inefficiency in the execution by one of these and not by conspiratorial underpant gnomes trying to destroy the system (1. Click Ads; 2. ???; 3. Profit!!!).
Be sure to also check out the video on Etre to see how the layout of your page can change viewing and click habits.
Here are rooms some of the creative geniuses at Pixar call home.
These are some of the walled-off areas used by their counterparts at Google.
And for comparison sake, here is a video of the offices at Apple.
Google’s researcher’s created this new technology that does the following: using the ambient noise heard from a computer microphone, it will analyze it and tell you something about it (usually linking it to an ad of some sort).
Cue omnipresent Hal 9000 voice.
