The Addled Goose’s KMWorld Lecture
October 6, 2008
After stumbling off the airplane from Schipol, I found myself facing an empty, dark lecture hall in the San Jose Convention Center. The building frigthtens me. The weird parking set up, the mysterious passages that connect the lecture halls to two hotels, and the absence of people who know where things are combine to make my goose feathers fall out.
Nevertheless, I was able to face the empty lecture hall. I was about 30 minutes early. I plugged my USB drive into the conference organizer’s laptop, got my flapping duck logo on the screen, and sat back to see if I would have more than six listeners or fewer than six. As it turned out, a brave group of 60 people filed in for my Google talk.
I write a column about Google for KMWorld, and the publisher takes pity on me. No one reads my column, although one high school teacher wrote me and said he found inspiration in my write up about Google Maps. Apparently I was a sufficiently magnetic oddity that five dozen folks trundled in to hear me provide a glimpse of what I was pondering for my November and December columns for KMWorld. (I don’t want to keep you in suspense. My November column will be about places to find informatoin about Google on Google.com. The December column will tackle Google’s most important announcements in 2008. I think you can figure out that I will talk about what Google will do in 2009 for my January 2009 essay.)
So what did I tell the folks who wrestled with the unfathomable floor plan for the San Jose Convention Center. Here’s a summary of the main points of my talk, Google G Cubed.
First, I ran over the now familiar theme from my 2005 monograph The Google Legacy. A happy 20 year old from an outfit called Oliver Wyman, a consulting firm with about which I was and remain deeply ignorant, told me last week that she “really read” my book. Way to go, Laura! The basic premise is that most people see Google as an ad and Web search company. My take is that Google is an application platform and the 21st centruy version of the original Ma Bell in her pre-break up days. The distinction is important because those who see Google as selling ads and indexing the Web miscalculate the risk of Google’s poking its snoot into other business sectors. Get this perception wrong, and the GOOG and its business models will test the market. Clicks, not top down mandates, dictate what the GOOG’s next moves will be. I use this diagram to explain the perception issue. What do you see?
Second, I mentioned Google’s interest in telephony in general and mobile services in particular. I mentioned that one of Google’s earlier telephony related inventions was for Quality of Service. QoS is the girdle for SLAs or service level agreements. These are guarantees that a particular service–now called a cloud service–will be available to a calculated level or some similar provision of the contract for services, appliatons, or software. But the key point I made was that mobile was important because the inventor of Google’s voice search system was not some 18 year old from Vladivostok. The inventor was Sergey Brin, soon to be a spaceman courtesy of Vladimir Putin’s space program. You can verify this by reading the patent document, US 7027987. As I pointed out in my second Google monograph, Googel Version 2.0, apparently of little interest to the hot MBAs at Oliver Wyman, certain big names on a Google patent signal a keen interest in a particular technical field. Sergey Brin’s name qualifies as a big name at Google. I suggested that even the Bell heads at AT&T and Verizon now recognize that Google is a player in their little monopolistic polo fields now. This time last year, these companies weren’t sure Google was “serious” based on my personal experiences with both firms’ executives. Thefigure below shows the flow of query processing in Mr. Brin’s invention:
(Oh, Cyrus, dear Googler, this illustration comes from the Google patent document, not Photoshop. Cyrus has a “ready, fire, aim” approach to figuring out where my Google illustrations come from.)
Third, I talked a bit about the Four complementary technologies and services that are discussed in hundreds of Web log postings, consultant studies, and trade journals. These are: Android, Google Gears, Google Docs, and Google Chrome. Instead of doing a tutorial on how each of these can be used by a programmer, I turned to some screen shots to make the ideas clearer. Most of the audience was snoozing at this point in my lecture. I wouldn’t listen to me either.
I pointed out that these “views” contain virtual machines, perform regular desktop application functions regardless of the user’s device or operating system, and provided Google with a mechanism for personalizing the information, the services, and what Microsoft calls “the experience”. I said, “Microsoft makes many experiences but they are not based on a homogeneous platform. Google delivers increasingly personalized experiences from a homogeneous platform.” No one in the audience knew what I was talking about.
To reinforce this point, I showed some “alleged” screen shots from Google and TMobile Android based device:
I was greeted with a couple of “that’s an iPhone” comments. Those sleeping continued to nap.
I then highlighted several tunes known to the Android-Gears-Docs-Chrome quartet. I mentioned and illustrated:
- A Google based automatic teller machine. I reminded the audience Google has money plumbing. The company uses it to collect its billions in ad revenue and to pay me pennies for AdSense clicks.
- A Google enabled BMW. After fiddling with the computer controlling my neighbor’s BMW, Google’s dashboard is a lot better than BMW’s own engineering clunker.
- A Google refrigerator. The idea is that Googlers who love recipes can log on to their refrigerator, find out what’s inside, and then get recipes for what to make. If the suggestions are not appealing, the system can provide a shopping list and generate more recipe options. This drew a big laugh from those awake. A segment of Google’s recipe interface appears below.
- I reprised my “I’m feeling doubly lucky” information. I included this innovation in a talk in early 2007. No one cared then that Google couold do remote medical monitoring from an Android based device and no cared in San Jose in 2008. At least I am consistent in my crowd appeal.
To conclude the talk, I showed one of my math-related diagrams. I said, “Google can grow quickly with little ramp time. The competitors may be stuck in the never never land of i (the square root of minus one). No laughs, of course.
I pay a $1.00 to anyone who will ask me a question after one of my lectures. One fellow raised his hand and asked, “Are you through?” I paid him a dollar and went back to my room.
Stephen Arnold, October 6, 2008
Hitwise: Microsoft’s Pay for Traffic Working
October 6, 2008
Hitwise, an Experian company, reported on October 3, 2008, that “MSN Cashback [is] successfully attracting visitors. You can read this remarkable article and see the graph “proving” the success here. Hitwise wrote:
In looking specifically at MSN Cashback, we see an interesting trend where the share of visits to the Cashback section of MSN Live is increasing. Eleven weeks ago, MSN Cashback represented 3.75% of the traffic to Live.com and grew to 6.29% last week. This rise in Cashback’s traffic underscores the interest in the program, which is likely to be getting a boost from shoppers looking to save money and stretch their budgets given the current economic climate.
The only challenge in “pay for users” is that users form a habit to get paid, not to search in my opinion. The for fee commercial services have used a variant of this with college students. Commercial online seems free. When the lads and lasses break into the real world, it’s free online services first and foremost. Maybe Microsoft has cracked the code? I doubt it, though.
The key question is, “Why does a vendor have to pay a person to use a service?” Answer: the other vendor–in this case the GOOG–has about 70 percent market share of US search with its share clicking up 0.01 or 0.02 each month. Users choose the GOOG and some of them go for the pay deal as well. The big number is the 70 percent share obtained without paying, answering email, or being particular helpful.
Hitwise is excited, but I am not. Are you?
Stephen Arnold, October 6, 2008
Xalo.vn Search Engine Launched
October 6, 2008
A happy quack to the reader who alerted me to the successor to Vinaseek. The new Web search engine is Xalo.vn, which indexes content in Vietnamese. The service according to the news item sent to me by my loyal reader is here. The article appeared in April 2008, but the English version is now available. The news story said:
Xalo.vn is designed to search in seven areas: Web, photos, news, music, blogs, forums, and small ads.
The company received funding from a variety of sources totaling about US$2.0 million. Other Vietnam centric search engines include PanVietnam and Vinaseek. Let me know if this link directs you correctly.
Stephen Arnold, October 6, 2008
Google: Mathematical Logic or Opportunism
October 5, 2008
Beta News posted a prescient article by Scott M. Fulton, III “Was the Yahoo/Google Deal a Ploy to Weaken Yahoo?” You can read this article here. The hook on which the article hangs is Google’s deal with Yahoo to sell ads for the Stanford-spawned America Online clone. Mr. Fulton does a good job of summarizing the original plan, the recent Department of Justice skepticism about competitors cooperating with money at stake, and the effort Yahoo has made to explain its side of the story. For me, the most significant passage in the write up was:
Google made the deal in bad faith, as an anti-competitive measure to maintain Yahoo’s subordinate position in the marketplace.
I have no enthusiasm for online advertising. The way the for-fee messages are presented reminds me of early Sunday morning cable TV in Kentucky or the wacky ads on Sirius Radio’s NFL channel for colon cleansers. Distasteful regardless of format and presentation.
My mind seized on Mr. Fulton’s comment and allocated a few senior moment time slices to these ideas:
- Was Google gambling with little or no forethought, acting out of a spirit of goodness and a touch of Google Yahoo against Redmond? After disrupting to some degree the Microsoft bid for Yahoo, were Googlers trying to help Yahoo or put the company in a position from which escape would neutralize the firm’s strength? After all, if Yahoo’s ad system worked, Yahooligans wouldn’t need Google?
- Was Google implementing an opportunistic series of unrelated tactical moves, roughly similar to going through the standard moves in a chess game until the more interesting end game begins to take shape? Would Google view its deal with Yahoo as a game, ultimately indifferent to the outcome even though the path leading to the outcome was a useful learning experience?
- Did Google apply mathematical logic to the Yahoo deal? Was Google able to assign values to certain variables and then run the numbers in order to determine that [a] if the deal went through, Microsoft would be even farther behind than it now is, [b] Yahoo was sufficiently desperate to overlook the fact that Yahoo’s own ad inventory would be subject to Google indirect influence and thus out of the control of Yahoo’s own semi-tarnished wizards, and / or [c] once the deal was inked Google could turn the dial on ad algorithms and starve Yahoo a basis point at a time?
I am on the fence. What’s your take? Google the Good, Google the Greedy, or Google the Godzilla? Bring facts, not just opinions, to your post on the Comments section for this Web log.
Stephen Arnold, October 5, 2008
Splunk: Log File Search Engine
October 5, 2008
The name “Splunk” rang a bell for me. I checked my archive and learned that Splunk was Splunk Server, a utility that indexes logs from Apache, OSS X, sendmail, and routers, among other sources. The system runs on Mac OS X 10.4. You can download a copy here.
A reader in the UK sent me a link to a new search engine, also called Splunk. The article was written by Colin Barker and titled “Search Company Splunk Launches in UK” here. Mr. Barker reported that Splunk has about “750 paying customers.” That’s a hefty number of customers for a company with which I was unfamiliar. Mr. Barker wrote:
Splunk is a search company that offers “flexible input methods [to] index logs, configurations, traps and alerts, messages, scripts and code and performance data from… applications, servers and networks devices”, Haynes said. The company’s product is aimed at corporate users, and works by searching across the network and answering IT managers questions about who is using the network, what resources are available and so on. Haynes claimed it does this very quickly and inexpensively in comparison to standard systems-management tools.
I visited the Splunk Web site here, and my confusion went away. Splunk is “the IT search engine.” The company’s system makes it easy for information technology professionals and system administrators to “search an navigate IT data from applications, servers and network devices in real-time. Logs, configurations, messages, traps and alerts, scripts, code, metrics and more.” The company’s catch phrase is one I like: “If a machine can generate it, Splunk can eat it.”
The company has added some management horsepower, and it is focusing on a specific search niche. If you want to index and search log files and other system generated data, check out Splunk.
Stephen Arnold, October 5, 2008
Hosted Microsoft Exchange
October 5, 2008
Email is the lubricant for many orgnizations’ business dealings. A reader sent me a link to David Hamilton’s “USA.Net Leads Email Market: Report” article on TheWHIR.com. You can access the news story here. I don’t think too much about email unless I’m working as an expert witness. Then email becomes a big deal. Mr. Hamilton reported that USA.Net is the “top player” in the Microsoft Exchange email hosting market. The USA.Net outfit was new to me. After some checking, I learned that the company offers a customized, scalable, and almost 100 percent crash proof hosted Exchange service. Mr. Hamilton reported tht USA.Net processes 38 million email messages a day and manages 80 terabytes of data from its 50 data centers. You can get more information about USA.Net here. The firm is a unit of Perimeter eSecurity. My question was, “How does a customer search email on this service?” I ran a query on the USA.Net Web site and the results list provided me with no concrete answer. Let me know if I have missed something. Finding Exchange email is as important as sending and receiving Exchange email in my opinion. I know I’m missing the obvious. Readers, any thoughts?
Stephen Arnold, October 5, 2008
Exalead’s High Performance Platform: CloudView
October 5, 2008
It’s no secret. When I profiled Exalead in one of the first three editions of Enterprise Search Report that I wrote, I likened the company’s plumbing to Google’s. The DNA of AltaVista.com influenced Google and Exalead. For most 20 somethings, AltaVista.com was one of a long line of pre-Google flops. That, like prognostications about Web 3.0, is not exactly on target.
The AltaVista.com search system was a demonstration of several interesting technologies developed by Digital Equipment Corporation’s engineers over many years. First, there was the multi core processor that ran hotter than the blood of a snorting bull in Pamplona. Second, there was the nifty manipulation of memory. In fact, that memory manipulation allowed Oracle performance in the system I played with to zip right along in the mid 1990s as I recall. And, the DEC engineers were able to index the Internet with its latency and flawed HTML so that a query was processed and a results list displayed quickly on my dial up modem in 1996. I even have a copy of AltaVista desk top search, one of the first of these scaled down search systems intended to make files in hierarchical systems findable. On my bookshelf is a copy of Eric and Deborah Ray’s AltaVista Search Revolution. Louis Monier wrote the forward. He used to work at Google, and, what few people know, is that Mr. Monier lured the founder of Exalead to work on the AltaVista.com project. Like I said, the DNA of AltaVista influenced Google and Exalead. In 1997, some AltaVista engineers were not happy campers after DEC was acquired by Compaq and then Hewlett Packard acquired Compaq. In the fury of the HP’s efforts to become really big, tiny AltaVista.com was an orphan, and an unwanted annoyance clamoring for hardware, money, engineering, and a business model.
François Bourdoncle–unlike Louis Monier, Jeff Dean, Sanjay Ghemawat, and Simon Tong, among others–did not join Google. In year 2000, he set up Exalead to build a next-generation information access and content processing system. What I find interesting is that just the trajectory of Google in Web search was affected by the AltaVista.com “gravity,” Exalead’s trajectory in content processing was also touched by the AltaVista.com experiment.
A result list from Exalead’s Web search system. Try it here.
When M. Bourdoncle founded Exalead, he wanted to resolve some of AltaVista’s known weaknesses. For example, the heat issues associated with the DEC Alpha chips was one problem. Another was rapid scaling, using commodity hardware, not hand crafted components which take months to obtain.
Exalead now has, according to the company’s Web site, more than 170 licensees. Earlier this week (October 1, 2008), Exalead CloudView, a new version of the company’s platform and new software features.
Paula Hane, Information Today, provided this run down of the new Exalead features:
Unlimited scalability and high performance
Business-level tuning and management of the search experience
Streamlined administration UI
Full traceability within the product
WYSIWYG configuration of indexing and search workflows
Advanced configuration management system (with built-in version control)
Improvements in the relevancy model
Provision for additional connectors with simple and advanced APIs for third-party implementations
You can read her “Exalead Offers a Cloud(y) View of Information Access here. The article provides substantive, useful information. For example, Ms. Hane reports:
One large [Exalead] customer in the U.K. can’t say enough good things about the choice of Exalead—its search solution was up and running in just 3 months. “After performing an extensive three-month technical evaluation of the major enterprise search software vendors we found that Exalead had the best technology, vision and ability to fulfill our demanding requirements,” says Peter Brooks-Johnson, product director of Rightmove, a fast-growing U.K. real estate Web site. “Not only does Exalead require minimal hardware to work effectively, but Exalead has a strong, accessible support team and a culture that takes pride in its customer implementations.”
(Note: A happy quack to Ms. Hane, whom I am quoting shamelessly in this Web log post.)
Phil Muncaster’s “Exalead Claims Enterprise Search Boost” here does a good job of explaining what’s coming from this Paris-based information access company. For me the most significant point in the write up was this passage:
The new line features a streamlined user interface, improved relevancy and the ability to extend business intelligence applications to textual search…
In my investigation of search company technology, I learned that Exalead’s ability to scale is comparable to Google’s. As Mr. Muncaster noted, the forthcoming version of the Exalead software–called CloudView–will put Exalead squarely in the business intelligence sector of the content processing market.
You can get more information about Exalead here. A fact sheet is also available here. Exalead’s Web index is available at www.exalead.com.
I have to wrangle a trip to Paris and learn more about Exalead. I hear the food is okay in Paris. The French have a strong tradition in math as well. I remember such trois étoiles innovators as Descartes, Mersenne, Poincaré, and Possson, and others. In my opinion, Microsoft should have acquired Exalead, not Fast Search & Transfer. Exalead is a next generation system; it scales; and it is easily “snapped in” to enterprise environments, including those dependent on SharePoint. I think Exalead is a company I want to watch more closely.
Stephen Arnold, October 5, 2008
Minimsft: A Biting Comment
October 4, 2008
A happy quack to the reader who sent me a link to the Minimsft Web log post for September 25, 2006. You can read the full write up here. Minimsft is allegedly a Microsoft employee who writes about the company. The editorial angle is that Microsoft should be a more lean and mean innovation machine. The hook on which the write up hangs is a change at Microsoft regarding executive compensation. The modification, filed with the Securities & Exchange Commission makes a provision for cash payouts for certain executives. Here’s the comment that I found revelatory, if it is accurate:
Microsoft is dying from the inside, and the folks sucking it dry have zero motivation to change things. It’s working out pretty damn well for them.
Microsoft’s recent decisions about search seem to be a move in the right direction. If these shifts are window dressing, the likelihood of Microsoft closing the gap between it and the GOOG is sharply reduced. Thoughts?
Stephen Arnold, October 4, 2008
The Goose Quacks: Arnold Endnote at Enterprise Search Summit
October 4, 2008
Editor’s Note: This is a file with a number of screen shots. If you are on a slow connection, skip this document.
One again I was batting last. I arrived the day before my talk from Europe, and I wasn’t sure what time it was or what day it was. In short, the addled goose was more off kilter than I had been in the Netherlands for my keynote at the Hartmann Utrecht conference and my meetings in Paris squished around the Utrecht gig.
I poked my head into about half of the sessions. I heard about managing search, taxonomies, business intelligence, and product pitches disguised as analyses. I’m going to be 65; I was tired; and I had heard similar talks a few days earlier in Europe. The challenges facing those involved with search are reaching a boiling point.
After dipping into the presentations, including the remarkable Ahead in the Clouds talk by Dr. Werner Vogels, top technical gun at Amazon, and some business process management razzle dazzle, I went back to the drawing board for my talk. I had just reviewed usage data that revealed that Google’s lead in Web search was nosing towards 70 percent of the search traffic. I also had some earlier cuts at the traffic data for the Top 50 Web sites. In the two hours before my talk, I fiddled with these data and produced an interesting graph of the Web usage. I did not use it in my talk, sticking with my big images snagged from Flickr. I don’t put many words on PowerPoint slides. In fact, I use them because conference organizers want a “paper”. I just send them the PowerPoint deck and give my talk using a note card which I hold in my hand or put on the podium in front of me. I hate PowerPoints.
Here’s the chart I made to see how the GOOG was doing in terms of Microsoft and Yahoo.
Source: http://blogs.zdnet.com/ITFacts/
The top six sites are where the action is. The other 44 sites are in the “long tail”. In this case, the sites out of the top 50 have few options for getting traffic. The 44 sites accounted in August 2008 for a big chunk percent of the calculated traffic, but no single site is likely to make it into the top six quickly. Google sits on top the pile and seems to be increasing its traffic each month. Google monetizes its traffic reasonably well, so it is generating $18 billion or so in the last 12 months.
In the enterprise search arena, I have only “off the record” sources. These ghostly people tell me that Google has:
- Shipped 24, 600 Google Search Appliances. For comparison, Fast Search & Transfer prior to its purchase by Microsoft had somewhere in the neighborhood of 2,500 enterprise search platform licensees. Now, of course, Fast Search has access to the 100 million happy SharePoint customers. Who knows what the Fast Search customer count is now? Not me.
- Become the standard for mapping in numerous government agencies, including those who don’t have signs on their buildings
- Been signing up as many as 3,000 Google Docs users per day, excluding the 1.5 million school children who will be using Google services in New South Wales, Australia.
I debated about how to spin these data. I decided to declare, “Google has won the search battle in 2008 and probably in 2009.” Not surprisingly, the audience was disturbed with my assertion. Remember, I did not parade these data. I use pictures like this one to make my point. This illustration shows a frustrated enterprise search customer setting fire to the vendor’s software disks, documentation, and one surly consultant:
How did I build up to the conclusion that Google has won the 2008-2009 search season. Here are the main points and some of the illustrations I used in my talk.
Cognos 8: Blurring Business Intelligence and Search
October 4, 2008
The death of enterprise search and the wobblies pulling down content management systems (CMS) are not well understood by licensees–yet. In the months going forward, the growing financial challenges in North America and Western Europe will take a toll on spending for information technology. The strong interest (based on my analysis of the clicks on the articles on this Web site) suggest that some folks are thinking hard about the utility of open source search systems and lower-cost alternatives to the seven figure price tags on some of the high profile search systems. I can’t mention these firms by name. My attorney is no fun at all. You can identify these vendors by going to almost any Web search system and keying the phrase “enterprise search” or “information access”. You can figure out the rest of the information from these results pages.
IBM baffles me. The company offers more information products and services than any other firm I track. Each year I try to sort out the product and service names. This year I noticed this information buried deep in one of the news stories about the new version of Cognos 8. My source is here,
My hunch is that IBM is creating a new map for business intelligence. On that map, IBM will point out the big X where the real high value payoff may be found. Here’s the pertinent passage from the IBM Cognos news release:
IBM’s recent CEO and CIO surveys have found unstructured corporate information such as user files, customer comments, medical images, Web and rich media content to be growing at 63%. The explosive growth of this type of business information has pushed the convergence of the BI and Search categories. It has created demand for new BI search capabilities to provide quick and easy access to both ranked and relevant BI content and unstructured information. Newly updated, IBM Cognos 8 Go! Search v4 lets any business user extend the decision-making capabilities of IBM Cognos 8 BI by securely accessing and dynamically creating BI content using simple key-word search criteria. The software works with popular enterprise search applications such as IBM OmniFind Enterprise Edition, Google, Yahoo and Autonomy so users can see structured, trusted BI content and unstructured data such as Word documents and PDF’s in the same view within a familiar interface. Users can search all fully-indexed metadata as well as titles and descriptions within a report. Search-assisted authoring and exploration gives them options to refine queries or analyze data cubes based on search terms. These capabilities speed access to the most relevant business information regardless of naming similarities between reports, helps business users quickly refine queries as required and frees IT from constantly re-creating commonly used reports. This leaves IT with more time for strategic business initiatives. The software is completely integrated with the web-based administration and security parameters set by IT administrators for IBM Cognos 8 BI. This integration provides a centralized, efficient approach to administration and security and effectively addresses two common areas of concern for resource-constrained IT departments, who want to provide more autonomy to business users, but need a single administration point and assurance that corporate authentication policies will be maintained. ‘These new enhancements to our Go! Portfolio provide business-driven performance information to help each area of the organization strategically manage the information that is most pertinent to them,’ said Leah MacMillan, vice president, product marketing, Cognos, an IBM Company. ‘Both the business and IT gain more autonomy whether employees are in the office searching, monitoring and analyzing business outcomes or on the road looking for new business updates or geographically relevant information.’ The IBM Cognos 8 Go! Portfolio of software is a key component of IBM’s Information Agenda, a new approach consisting of industry-specific software and consulting services geared to helping customers use information as a strategic asset across their businesses. [Emphasis added]
Let me deconstruct this passage using my addled goose methods.
-
- Subscribe to Beyond Search
Feature archive
News archive
-
