Pattern Finder

Blinkx: More than Meets the Blinking Eye

On Friday I had a discussion with Suranga Chandratillake, Blinkx's CEO.  It was an interesting conversation because it ran counter to the popular perception.  Blinkx has been portrayed in the popular media as just one of a number of companies offering desktop search, including Copernic, Google, Microsoft, and X1.  The story typically has the cast of a "David and Goliath" story -- along the lines of, "Interesting technology, but just wait until Google and Microsoft get their act together."

From talking to Suranga, it's clear that the desktop search product was a "let's get our feet wet" product -- merely the first stage in a much longer planned journey.  I don't think I'm giving too much of the game away if I say that Blinkx has a very multimedia mindset.  The company is convinced that searching across multimedia content is the next big problem.  Driven by ubiquitous Internet connections, the growth in broadband, cheap disk space, and increasingly inexpensive and powerful PCs, both companies and consumers are going to want to find the interesting snippet in the podcast or the video broadcast.

Such multimedia search has been touted for a number of years.  It will be interesting to see if we are indeed near the tipping point.

Now You See It, Now You Don't

On Thursday, May 26th, Autonomy's Web site had a press release in its press release section announcing that the BBC had selected Autonomy technology for its Video-on-Demand search tool.  Today, a day later, it's gone.  Poof!

It's unclear what the reason is:  perhaps Autonomy got de-selected or the powers that be at the BBC hadn't OK'd the release.  For those curious what it said, here's the summary:

BBC Selects Autonomy's Technology for Video-on-Demand Search Tool

Cambridge, England - 25 May, 2005

Autonomy Corporation BBC Broadcast, one of the commercial companies of the BBC, has announced its selection of Autonomy's technology for the development of its Video-On-Demand (VOD) search tool. The tool will enable the automation and search of media and automated publishing of content over multiple platforms, including broadband and mobile. Autonomy provides the backend retrieval for the BBC's VOD tool, enabling retrieval and recommendation through BBC Broadcast's online store of media.

[Update:  Almost a week later, the press release finally made it to the site again.  Comparing the two versions, the earlier one didn't give credit to an integrator:  "Kontiki, the content delivery specialists, will provide the user interface and the delivery to multiple hardware platforms."  Oops.]

GreaseMonkey Gums Up the Works

As a faithful Firefox user, I cruise the Firefox Extensions page from time to time, and came across GreaseMonkey.  This is an extension that lets users modify how a Web page displays via JavaScript.  Happily, a number of folks have already posted pre-built scripts.

So, GreaseMonkey is nice for a user.  But a disaster for Web analytics.  All of a sudden, users can modify a Web page so that it behaves differently for them.  An enterprise assumes that users are seeing what their Web server has delivered to them; all of a sudden that assumption is out the window.  So start getting a little bit more nervous that your Web analytics system isn't telling you the truth.

Now the flip side of this is that folks use GreaseMonkey to make the page more useable (from their point of view).  In short, GreaseMonkey is a system that makes customer voice very concrete.  Think about it.  Customers have become so unhappy with your site that they've taken the time to recode it.  So if you find a GreaseMonkey script posted for your site, take a look at it and discover the behavior customers would prefer.

The "Blogginess" of Newspaper Readers

An interesting post by Ethan Zuckerman on how "bloggy" certain newspapers are.  You need to read the post to understand the methodology, but the Christian Science Monitor comes out on top with a score of 134.90; the LA Times is at the bottom with a score of 11.21.

FAST & EMC/Documentum Get Together

On Monday, May 23rd, FAST announced that FAST's InStream product would be integrated into EMC/Documentum's 5.3 platform.  This OEM announcement is a huge win for FAST -- I would say it's even more important that FAST's recent Siebel announcement (Siebel can't get its act together, EMC is cruising).

Under Joe Tucci, EMC clearly sees the future, what with its purchase of Documentum and forays into virtualization.  EMC has figured out that the problem for the future is managing information, not data -- for example, that files should be backed up based on their business value, not how old they are.  I knew Tucci at Wang Laboratories, and I think he's similar to another Wang alumnus, John Chambers.  They both are consistently underestimated.  John turns on his West Virginia accent, and non-Southerners figure he's a hick.  Joe is his amiable self, and folks figure he can't be that bright.

However, based on what he did at Wang and some short conversations I had with him at the time, he is one smart cookie.  In short, I think FAST has hooked up with a company that (1) will be going places and (2) has an installed base to be envied.  The EMC/Documentum deal gives FAST permission to show what it can do for mission-critical content management applications.  If a pharmaceutical sees how search can help its business, it will probably throw more money in FAST's direction.  Not a bad place to be.

The Switch from Technology to Process

On Monday, Eloqua Corporation released the Spring 2005 version of its Eloqua Conversion Suite.  (Eloqua, a Canadian company, is apparently targeting its sales effort to the Northern Hemisphere.  In Australia, this time of year is Fall.)  The application uses a variety of technologies -- Web analytics, sales force automation, and direct marketing -- to discover leads and track them through the Interest Pipeline and ultimately the Sales Pipeline.

Eloqua has been taking this process-oriented approach for a long time.  I remember talking to Mark Organ, the President, right after I'd published the industry's first large scale report on Web analytics and Mark already had that sales process focus in mind.  The majority of the companies doing online technology bits were technology-focused, not process-focused.  However, as the technology matures, I'm seeing more and more companies take such a process focus.  In fact, Manticore, which had been focused on Web analytics, is now taking a page out of Eloqua's book and taking on more of a sales lead bent.  It will be interesting to watch the technology-to-process transition accelerate, as I'm sure it will. 

Douglas Adams: A Designer of Systems

The movie The Hitchhiker's Guide to the Galaxy opened last week to great fanfare and excellent reviews.  However, the event was a bit bittersweet for me; the author, Douglas Adams, died several years ago, and consequently the IT world lost a wonderful visionary.

I know, because fifteen years ago I worked with a bunch of people intent on bringing the Hitchhiker's Guide to life.  I was working at Wang Laboratories, at the time a $3.3 billion computer company, a pioneer in pen-based systems.  In 1990, a set of developers worked at creating The Guide, a pen-based PC that we would now call a tablet PC.  (For those who think tablet PCs were a relatively recent invention, Wang and NCR both shipped production versions in 1990).

Douglas's book was the inspiration for The Guide, which led to some interesting design decisions.  Most other companies building a tablet PC at the time (NCR, Toshiba) looked at a pen PC as a laptop with a pen-enabled screen.  Consequently, the designs were kind of kludgey.  However, Douglas's spec called for a much more integrated system.  Therefore, besides a PC chip (386), The Guide sported a DSP, a specialized chip that could do many things.  For example, the DSP enabled the modem as well as a voice recorder.  The system also had a chip whose sole purpose was to clean up the digitizer information.  Consequently, The Guide's handwriting recognition was head and shoulders was above competitive systems, since the data from the digitizer was smoother and less noisy.

Because of The Guide, I got to meet Douglas Adams in person.  He'd been speaking at a conference in California; the development manager of the group happened to attend.  He went up to Douglas after his talk and asked, would you like to see The Guide?  We've built it.

Douglas leapt at the chance.  So on the way back to the UK, he stopped off in Boston and came up to Lowell, where we demonstrated the system.  He was quite impressed, given the state of technology at the time.  However, it pained him considerably that it ran on Microsoft Windows; he was an Apple convert through and through.

After signing some books and discussing the system, the 12 of us convened to a Japanese restaurant in Nashua, NH, where Douglas recounted, in hilarious fashion, the origins of some of the stories in the book.  He was tall (probably around 6'5"), funny, loved technology, and was quite a fellow.  But perhaps most importantly, he thought about systems from the outside in -- he thought about what would be nifty to have as a user, rather than starting with the technology and then retrofitting it to the user.  In short, he was a great systems designer, even if the only specs he wrote were novels.

UW Desktop Search Report

The University of Wisconsin E-Business Institute has released a 15-page report on 12 desktop search tools.  While I quibble with their specific rankings (they put Copernic Desktop Search at the top, while I prefer Yahoo! Desktop Search, their #2 pick), I think the general order is pretty good.  For example, they ding Blinkx for hoggish indexing, and that was my experience.  (If you want to slow a fast PC to a crawl, load on about three of these puppies at once and watch the drive go nuts as they all index your content.  It brings back the days of the 386).

The report notes in perfect passive voice-ease that, "When all the results were reviewed, it was determined that most of the desktop search tools were still too immature for significant business use due primarily to a lack of mature security and overall manageability."  I concur, with reservations -- that with the speed of technology improvement in this area (and the release of Apple's Tiger and Microsoft's Longhorn) -- desktop search will, over time, be business critical.  To be fair, UW says the same thing, but points to IM as the example of a personal technology that was eventually adopted by business.

Progress Buys EasyAsk

On Monday, Progress Software announced it was buying EasyAsk, an enterprise search company several miles down the road.  The two companies know each other well.  As the press release notes, "several PSC directors serve on the board of EasyAsk and several PSC directors hold, directly or indirectly, equity interests in EasyAsk."

EasyAsk has concentrated on e-Commerce search.  The press release states, "EasyAsk is the leading provider of natural language question/answer and eCommerce search solutions" -- a clear case of PR hyperbole.  While "leading" is incorrect -- FAST, iPhrase, and Endeca would certainly disagree -- is it one of the players, and counts among its customers Ann Taylor, Lands' End, Sony, and Talbots.  However, it hasn't been minting money, as its $9.2 million sale price attests.

The owners probably figured things aren't getting better anytime soon, and decided to (1) get money (a bird in the hand is worth two in the bush) and (2) work under the larger Progress umbrella, which has deeper pockets and a much larger installed base.

Progress is one of those quiet database companies, somewhat like Sybase, in that oficionados like the technology, even if the company doesn't have a large mindshare with the general public.  Progress cruises along, growing each year, selling to VARs, making them look like heroes.

What's interesting about the acquisition is it signals that search is starting to become embedded in other applications.  Search has been standalone for a number of years because the technology has been so specialized.  However, as it becomes more mainstream, it will now no doubt be absorbed by applications, in the same way that workflow became absorbed a decade ago.