Business Intelligence: Revenues Drift Down
June 21, 2008
The business intelligence revolution has come and seems to be headed for knee surgery. The nimbleness is gone. If Information Week has it right, business intelligence in the US needs help to get back on the revenue treadmill. Mary Hayes Weier’s essay “Business Intelligence Software Growth Shows Dramatic Drop in U.S.” is here. Ms. Weier’s key point is, “Sales growth of BI software in the United States…sputtered to just five percent.”
BI, as the aficionados prefer, is the stuff that fires the synapses of smart managers. That may be true, but the market is dominated by a handful of companies who seem to be intent of sucking the oxygen from the market, stunting smaller BI vendors, and competing with one another.
Software giants like IBM, Microsoft, and Oracle seem to be moving to a super sized version of Microsoft Office only this time the suite contains application servers, databases, analytics, and search. Will this strategy of delivering a dump truck filled with software help or hurt business intelligence?
My thought upon reading Ms. Weier’s essay was that the effect of on premises systems of such complexity will hasten the emergence of cloud-based solutions. The reasons are easy to identify:
- Information technology departments are no longer able to budget reliably. The slightest glitch can chew through a budget. More complexity means more glitches and less control and predictability in IT spending. Solution? Shift to the cloud and a price list.
- Vendors will find their marketing assurances losing efficacy. Customers cannot afford systems that do not work as advertised. For many years, licensees have been reluctant to grouse. That is beginning to change. Even stage managed vendor trade shows are becoming tough to hold to the party line. Going forward, licensees may become more vocal in their criticism of software and pricing policies. Buggy software “sells” consulting services until licensees get savvy about this ploy.
- Smaller firms may find it easier to explore alternative delivery, pricing, and support models. A small vendor has narrow margin of error. On the other hand, a smaller outfit can make a change to a business quickly. When there are large numbers of competitors, one or two of these outfits may find the keys to the kingdom. The giant firms will be unable to adapt quickly and in effect become more vulnerable.
The business intelligence wave has come, hit the shore, and is now receding. Companies want to make decisions based on data, and the winners will be firms who can make complexity less painful from the cloud. My hypothesis is that a shake up is coming. It may take many years, but the dominant companies will be the BI equivalent of Toyota and Honda surrounded by a small number of specialists. Whom do you think will emerge as the BI winners? I am going to put my money on the GOOG for these reasons: big league analytics in an easy to use package, cloud capability, and big data to complement the puny data sets most companies crunch for their current BI analyses.
Stephen Arnold, June 21, 2008
Boomers and Millennials: Implications for Enterprise Search
June 20, 2008
Enterprise Search and the Age Gap
Employees, contractors, and consultants are becoming younger. For enterprise search, aging boomers are leaving the work force and younger employees moving in.
En route from San Francisco to the less civilized environs of rural Kentucky, I made a list of the differences between Millennials and Baby Boomers. Millennials are all digital all the time. Baby Boomers have luggage stuffed with printed books, paper calendars, and blank notebooks in which one letter at a time can be written using a pencil. For simplicity, I will call the Millennials the younger workers, and the Baby Boomers the aging dinosaurs. Keep in mind that you may be 25 and as mired in books and microfilm as an ossified Baby Boomer. The categories are not absolute. The two part division is intended to make it easy for me to communicate my thoughts about the changes wrought upon search as as Baby Boomers become the minority in organizations and Millennials become the majority.
I want to alert you that any one under the age of 35 will probably be annoyed at my thoughts. But this is a Web log, and I am going to capture these notions before I pass out from the brutalities of a red-eye flight seated next to the lavatory. In short, another red eye, another Web log essay about enterprise search from a different angle.
The generational differences mark a clean break with key word search and retrieval systems of the past and point to more sophisticated and complex information access solutions more youthful enterprise system users require.
Seven Differences between a Young Professional and a Near Retirement Professional
Difference 1: Under 35s don’t read anything long. I have the impression that the under 35 enterprise search user wants short, chunky information from search systems. Systems that return long documents that have to be printed out, annotated, and studies are not what users of search systems want from their information access systems. Over 55s (yes, I am generalizing) may not like long documents, but I for one will slog through this stuff. There may be gold in those hills, I think.
Difference 2: Under 35s want to have search suggestions, assisted navigation, Use For references, and See Also hints. Over 55s like me don’t have much resistance to formulating a query, scanning results, reformulating the query, scanning results, and finally narrowing the result set to a useful collection of documents which can then one-by-one be reviewed. I love shortcuts, but research is research.
Difference 3: Under 35s seem to have the uncanny ability to do several electronic tasks at once. At the Gilbane conference I watched as professional journalists listened to a speaker, sent messages on a BlackBerry, and chatted with the person sitting next to her. I am lucky if I can listen to the speaker; forget the digital activity. Over 55s are less adept multi taskers. The reason the BART train was speeding and crashed into a stopped train appears to have been a young train driver who was chatting on a mobile and controlling the subway train. I prefer single task focus to avoid collisions.
Another Google Should: Buy the Associated Press
June 20, 2008
I enjoy “Google should” essays. Google has money, technology, the number one global band, and the ability to move like a ninja. Wired’s Web log carries Betsy Schiffman’s interesting essay “Forget the New York Times: Google Should Buy The AP”. You can read it here.
The idea is one way for the Associated Press to jump from the tracks and avoid the same fate as a coin placed on the railroad tracks so the wheels can flatten it. Playing on train tracks is fun; letting the wheels of the locomotive rework a penny is a sudden transformation.
The most interesting point in her essay for me is this statement: “The flip side of the equation is that web companies are picking up where the newspapers left off.”
That nails it.
The implications for enterprise search are significant. More and more organizations want to create a Folger’s blend for their Intranet or behind the firewall search users. For fee content has been available to organizations for many years. Now, why bother? Even the high value information such as financial data are becoming more findable. Stock traders need their fancy Bloomberg terminals and Reuters data. But for a snapshot of a competitor Google Finance works well for me. (Yahoo, I fear may be slipping off my radar due to organic issues at that company.)
In June 2007 I made the suggestion that the Associated Press should find a way to “surf on Google”. Perhaps the tie up between Google and AP should become more formal, as Ms. Schiffman suggests. She’s on the right insight vibe as I. The Google is more than Web search and advertising. I am more convinced than ever that my describing the company as a “supranational corporation” is an understatement. The GOOD is our own informational revolution. Instead of sitting in the Black Country in England we are in Data Country. News is one piece of raw material in this new world.
Stephen Arnold, June 20, 2008
English Invade France: Autonomy Snares Lyon Library
June 20, 2008
Autonomy, arguably one of the top two or three vendors of enterprise search, landed a big fish. In fact, the company snagged the second biggest library in France, the Bibliothèque Municipale de Lyon. You can read the full story on CityAM here. (No quotes. The AP sabre rattling echoes in my ears.)
Why is this important? Three reasons:
- France has some serious search, content processing, and text analytics vendors. The ones that merit a close look in any search bake off include Exalead and PolySpot, but there are others about whom I have written in Beyond Search and this Web log.
- Libraries are strapped for cash, so these organizations look for a vendor that offers text firepower and a good deal. My sources tell me that Autonomy’s pencil sharpening carried the day which, when combined with a video search capability, melted French hearts the way the sun softens camembert.
- User expectations for search are soaring. At the same time, dissatisfaction with search systems is rising as well. Lyon’s technologists have sent a bright signal that Autonomy can deliver a better solution and one that will leave users smiling and the users with big grins
So Autonomy has invaded Lyon. The company will work overtime to make this sales win the foundation of other attempts to win business. I will be monitoring the Lyon implementation, the reaction of the French technologists, and the number of wins that Autonomy can achieve.
Stephen Arnold
June 20, 2008
Update 1
A helpful (though reticent reader) has alerted me that the Autonomy Lyon win is not new. You could have fooled me. Here’s the news release that I saw on June 16, 2008, and I certainly thought “news” meant “news”. To my aging self, the news release appeared to originate in Cambridge, UK and San Francisco, California, and I understood the news release to report that event. My anonymous email writer pointed out that the library had been a Verity customer and this was a multi year extension. My anonymous writer suggested I do more research before commenting about news releases. Point well taken, but I’m trying to link actions in search to the needs of users, not analyze the veracity of what appears on PRNewswire. At my age, it is a habit (obviously a bad one) to assume that a “news release” contains news.
Stephen Arnold, June 20, 2008, 8 50 am
Gilbane 2008: Three Things I Learned
June 20, 2008
The Gilbane 2008 conference ended. Attendees seemed happy. In fact, as the exhibit tear down began, some attendees were sitting at tables in the registration area exchanging business cards and planning enterprise application moves. At times, the snap of business cards rivaled the noise in and around a Las Vegas 21 table.
I lost track of the sessions into which I poked my beak. Knitting together what I heard from earnest lecturers, in the break chats, and from exhibitors who smiled continuously for two days, I learned three things.
First, content management systems designed to end hassles with content for the Web and other outputs don’t work very well. One person told me, “CMS. My goodness, what a disaster.” I don’t think this youthful looking person was exaggerating. CMS is one of those software systems that is supposed to allow an organization with neither publishing work processes or people who can write very well do both and generate content automatically.
Second, enterprise search does not work very well. For the first time, a number of different people were talking about the problems of search, but what I heard boiled down to two issues: [a] users don’t use the system and [b] the system is tough to fix. One person told me, “I was surprised how many people admitted that their search systems were not what the companies thought they were buying.” Popular silver bullets and amulets included taxonomies, social search, and semantics. All incantations to keep search evils at bay.
Third, consultants feasted on the attendees looking for silver bullets. The UK outfit 451 and Gartner were making sales left and right. Lesser souls were also dragging in nets filled with prospects. I steered clear of the consultants because the flashing white teeth and the broad smiles frightened me.
Quite a lesson for me.
Stephen Arnold
June 20, 2008
Newspapers: Descent from Mt. Olympus Continues
June 19, 2008
No references or quotes from the Associated Press for me: The news, however, is easy to find, and it has grave implications for enterprise search systems dependent on for-fee content from traditoinal publishers.
A quick read of these stories makes clear that unless traditional news organizations staunch the bleeding, traditional newspapers may be reduced to a shadow of their present selves. This is more startling than the Subway fast food weight loss advertisements. Newspapers may be even more trim that the once-chubby Jared.
You can start your learning about the financial plight at these two links:
- Editor & Publisher, “Behind the McClatchy Layoffs”. One-second summary: Ads go away.
- Bloomberg, “Tribune, MediaNews May Wind Up in Default as Ad Sales Evaporate,” One-second summary: Default possible
Enterprises depending on branded news may want to increase their intake of unbranded news available via the Internet. I heard that Factiva is stepping up its professional services activity. Good move. Content licensing may be in for some rough water.
Stephen Arnold, June 19, 2008
SharePoint: The Digital Maginot Line
June 19, 2008
Internet News has a must read story about Microsoft SharePoint here. Richard Adhika’s “Search, Social Networking Key in SharePoint” nails the identify crisis that Microsoft faces with this server product. SharePoint is search and social networking, The story casts into sharp relief that SharePoint is a polymorphic product. With millions of users, three flavors of search, and dozens of Certified Gold Partners selling add ins and add ons SharePoint is important to Microsoft.
To me the most interesting statement in the essay is:
Echoing statements by analysts and other vendors, he said the danger is that the millennials working in enterprises will “turn to outside services on the Internet,” which may breach compliance regulations and spur fears about information leaking outside the organization. Corporate IT is “increasingly thinking about how to build an internal social networking platform,” Koenigsbauer said, adding that SharePoint Server 2007 provides native support for wikis (define) and blogs, and lets users push content to mobile devices.
SharePoint, if I interpret Mr. Koenigsbauer’s comments as intended, suggests that SharePoint is a Maginot line, the line of concrete fortifications designed to protect France from Germany. Perhaps SharePoint in its search and collaborative form will work.
Stephen Arnold, June 19, 2008
Fast Search: Is This a Real License Document?
June 19, 2008
I was updating my Fast Search & Transfer files. Over the last two months, I have noticed that certain information has been removed from the Fast Search Web site. I ran my crawler scripts and saw in the hit list this link:
http://contracts.onecle.com/findwhat/fast.lic.shtml
I am not familiar with onecle.com, so I ran a whois search. The owner of the domain is listed with some data that are not particularly helpful The title of the Web site to which the domain points is “California MCLE, CLE and Continuing Legal Education.” The purpose of the Web site seems to be to provide sample agreements for people to examine.
What interested me is this document which you can view if you click here. It appears to me, and I am not an attorney and have zero qualifications to do anything other than obey the law, that this agreement sets forth the terms for a deal between FindWhat.com and Fast Search & Transfer. FindWhat.com has become MIVA, which is a vendor of pay-for-click software and services.
The Onecle splash screen. Access to the site is here.
Because I am not certain of the provenance of this sample license agreement, I will not reproduce it here. However, there were three parts of the agreement that I found interesting. Let me highlight each of these points and then offer several observations about my understanding of each. You will need to print out the entire agreement or have it one your screen as you read my post. I am not going to quote more than a sentence from the source document just as I would have done in the 7th grade when I wrote my first term paper. Miss Soapes was quite negative about using work by another as one’s own.
The Three Points
1. Maintenance and Support
The $7 million license fee may be bogus. I heard that Fast Search licensing began in the $175,000 to $250,000 range and could go higher depending on the customer’s requirements. But $7 million seems high to me. But set that questionable number aside. The agreement says in 4. Maintenance and Support:
Customer shall purchase maintenance and support services from FAST with respect to all software licensed hereunder for a three year term
The fee, if I understand this correctly, is the starting point. Additional fees will be assessed to maintain the system and support it. I learned that most enterprise software vendors charge anywhere from 15 – 25 percent of the license fee for maintenance and support, I think the cost of the installation jumps significantly. I also believe that customization of the system adds to the cost.
A segment of the alleged Fast Search license agreement. Full document is here.
What baffles me is that Fast Search stumbled into financial difficulty because revenue did not flow into the company quickly enough to off set its costs. My thought is that customers signed a deal and then balked at paying when the system could not be set up, made operational quickly, and then supported at the specified rates. Whatever the license fee, I think it was not enough to free Fast Search from of financial pressure.
2. Schedule B: Service Level Objectives
Fast Search spells out that the customer must try to figure out what the problem is. Okay, that’s reasonable. However, Fast Search then limits who can contact Fast Search about the problem. You will find this language in Schedule B, Paragraph A. The point that hit me is that if the “standard support” is not what the customer needs, then the licensee can sign up for “Premium Support”. Again more charges to make a system work. Furthermore, Fast Search offers “Resolution Objectives”. When I read these, I concluded that Fast Search may not be able to fix some problems; therefore, a work around may be provided. Some work arounds can take up to a month. My thought was that if I am using Fast Search to generate revenue for my company, I cannot be offline or down for a month. I would say, “This software is supposed to work, right?” I am not certain if procurement teams poke their noses into the legal documents for an enterprise search acquisition. An attorney, unfamiliar with information access systems, might overlook these nuances of support. When a problem arises, I can see that it would reach a flash point quickly as the procurement team tries to get the system working only to be told that the caller is not on the list and that the fix may take a month.
3. Schedule D and Schedule E: Customer Competitors and Fast Search Competitors
My thought, when I saw these lists, was that these are not timely, which casts doubt on the authenticity of this sample license agreement. On the other hand, I wondered if a licensee’s legal department would review an agreement such as this and routinely update the “you cannot work with these people” lists. Now that Microsoft owns Fast Search, I think Fast Search licensees need to revisit their agreements and I assume that the Microsoft-Fast Search team will be contacting licensees to update these lists. What I found interesting is that Fast Search listed Microsoft as a competitor along with Google, Yahoo, Verity, Autonomy, Convera, and Endeca. Now Microsoft owns Fast Search, and it will be interesting to see if a sample license agreement becomes available.
Observations
I have two observations about this agreement, which may not be a legitimate contract:
First, the fees seem designed to produce significant revenue. That is okay as long as the system works. When the system does not work, the fees become an issue. Big companies with big bills owed to Fast Search may quit paying. The alleged financial difficulties may be a result of big companies not paying their bills.
Second, I will be most interested in any changes in Fast Search’s pricing and business policies under Microsoft ownership. The changes may reveal the approach Microsoft will take with the Fast ESP technology.
If you have insights, or simply wish to disagree with me, use the comments section on the Web log.
Stephen Arnold
June 19, 2008
Forbes on Powerset
June 19, 2008
Forbes Magazine has an interesting article about Powerset, Chris Taylor’s “The Next Search Frontier: Just Ask Your Question“. I often have difficulty locating information on the Forbes’ Web site. Sometimes I grow frustrated with the pop up ads and page latency, so snag this article quickly.)
The key point in the article for me was this statement:
Powerset’s main asset is a partnership with PARC, the Palo Alto research center that incubated the computer mouse and the laser printer. In 2005, Pell discovered that PARC researchers had been working for 30 years on turning English into software code. Pell promptly licensed PARC’s research and hired the top scientists in the field, starting with Powerset co-founder Lorenzo Thione.
Xerox PARC (now simply PARC — it’s officially a subsidiary company of Xerox) has been an innovator for many years. But my experience has been that some of its better ideas are difficult to commercialize and convert into major revenue winners. Inxight Software, a PARC spin out, gained some market success and was acquired by Business Objects, which in turn was acquired by SAP. Powerset’s tie up with PARC will be another opportunity to convert ideas into revenue.
You can test drive Powerset here. Information about PARC is here.
I am accustomed to formulating queries with Boolean ANDs and NOTs. Typing questions is too much work for me. With the average query creeping up to 2.3 words on major public search engines, the idea that a well formed question will revolutionize search seems unlikely.
Natural language processing, like semantic and linguistics mechanisms, may be best suited for work behind the scenes, not in front of the user.
Stephen Arnold, June 19, 2008
Clarabridge Version 3.0 Now Available
June 19, 2008
Clarabridge has released Version 3.0 of its text mining platform. The company–based in Reston, Virginia–provides text analytic solutions to industry and the government.
Business analytics has been difficult for average managers to use. Many traditional sytems require that a specialist formulate a query for the business intelligence system. The manager then receives a report. If the report does not answer the question, then the time-consuming intermediated process is repeated. Not surprisingly, some managers prefer to obtain information other ways.
The new version of Clarabridge’s system boasts a drag-and-drop Web interface which the company labels Navigator. You can obtain more information about the system and Navigator here.
Now managers with access to tools usable wihout an intermediary need to recall enough collage math to know what the functions do and trust that the data analyzed meet statistical tests of goodness.
Stephen Arnold, June 18, 2008