WebMama Puts Mark Logic CEO Blog in Top Ten

SEO and SEM consultant Barbara Coll, author of WebMama's Look at the Web blog, and CEO of the Palo Alto consultancy WebMama.com, Inc., did a blog post today entitled 45 Online Resources for CEOs which named this blog to a list of top ten management blogs to which readers should subscribe.

WebMama's top management blog (of which I count only 7 -- I guess she's saving 3) list is:

David Kellogg, CEO
Mark Logic, Inc.

Joel Spolsky, CEO
Fog Creek Software

John Mackey, CEO
Whole Foods

Guy Kawasaki, Managing Director
Garage Technology Ventures, founder of Truemors and Alltop

Jonathan Schwartz, CEO and President
Sun Microsystems, Inc.

Mark Cuban, Owner
Dallas Mavericks

Craig Newmark, CEO
Craigslist

Thanks WebMama, that's quite some company. Next thing you know I'll be on Dancing with the Stars. :-)

Mark Logic and the US Army's Warrior Knowledge Base

Check out this story, published today, entitled Mark Logic Technology Key to Success of Army's Warrior Knowledge Base.

Here are a few excerpts:

“Connecting our 90,000-plus members with relevant documents in the ... repository within seconds, and then sharing that content with other forum members, is our goal,” said Mark Uhart, Warrior Knowledge Base integrator and knowledge management consultant. “The [Warrior Knowledge Base] shows promising capabilities that have already demonstrated great utility ...

The metadata assigned to these types of documents, and the fact that content can be discovered and viewed page-by-page without having to download the file, allows discovery and use of only the relevant content. [The Battle Command Knowledge System] has the capability to store and manage content in 37 different languages.” ...

We knew we had to provide content discovery and access at the page level to be of any value to forward-deployed soldiers and the DOD/DA contractors and civilians who support them,” Uhart said. “They have neither the bandwidth nor the time to download a large file only to discover it’s not what they wanted."

Basically some great arguments in favor of fine-grained search and retrieval, multi-channel content delivery (to handhelds) and web 2.0-style participation / user-generated content.

Related articles

• Mark Logic Innovation Awards Photo [via Zemanta]

Happiness as a Business Model

Just a quick post to highlight an exceptional PowerPoint deck (that I suspect is making the viral rounds) on happiness as the first-principle from which you can create a successful business model. It's by Tara Hunt, founder of Internet consultancy Citizen Agency, and author of a blog called HorsePigCow.

I like the deck not only because of the psychological first principles from which its drawn, but also because it is an A+ demo of what I'd call "the new PowerPoint" style.

People are so burned out on old-school PowerPoint that I see many people make one of two mistakes: (1) throwing out the baby with the bathwater, dispensing with slides altogether or (2) forgetting that slides are a useful medium for creating "written presentations" -- i.e., decks that stand alone, intending to be read / clicked-through and not necessarily as materials to support a live presentation.

| View | Upload your own

Moderating Gilbane Panel: Digital Publishing Platforms

If you're attending the Gilbane Conference this week at the Westin in San Francisco, then please don't miss the session I'm moderating on digital publishing platforms.

The session title is Digital Publishing Platforms: Magazines, Newspapers, and eBooks. It's on Friday morning June 20th at 10:15 AM. Joining me are:

• Andrew Savikas, director of publishing technology at O'Reilly Media, general manager of the TOC conference (tools of change for publishing), and regular contributor to the O'Reilly Radar blog.

• Cimarron Buser, senior vice president of marketing and business development at Texterity and author of a blog called Digital Publishing Insights.
  

If you're attending, I'd suggest as prep the two posts I blogged about the other day from Scott Karp

• What Newspapers Still Don't Understand about the Web
• What Magazines Still Don't Understand about the Web
  

LinkedIn Raises $53M and Posts Video on Valuation

Today, the business social-networking site LinkedIn announced that it raised $53M at just over a $1B valuation. In addition, they posted a YouTube video with Jeffrey Glass from Bain Capital, David Sze from Greylock, David Cowan from Bessemer, and (Mark Logic board member) Mark Kvamme from Sequoia. Per TechCrunch this round brings the total funding up to $80.5M.

Related articles

• Confirmed: LinkedIn raises funding at $1B valuation, plans to buy smaller companies before going public [via Zemanta]
• LinkedIn gets its billion-dollar valuation [via Zemanta]
• LinkedIn secures $53m European boost [via Zemanta]

Mark Logic Innovation Awards Photo

Here's a photograph from last week's user conference with representatives from organizations recognized in the first annual Mark Logic Innovation Awards.

From left to right we have:

• Telly Stroumbis, The Boeing Company, for work within the US government
  
• Dave Kellogg, Mark Logic Corporation
• Ryan Shevchik, Booz Allen Hamilton Inc., for work within the US government
  
• Hank Hoffman, Congressional Quarterly Inc.
• Etienne Taylor, Clinical Trials Semantics Inc., for work with the American Cancer Society
• Darin McBeath, Elsevier, B.V.
• John Walker, John Wiley & Sons, Inc.
• Carl Hixson, The McGraw-Hill Companies
• Alex Moulas, OpenConnect Systems Inc.
• Charles Greer, O’Reilly Media, Inc.
• Alex Humphreys, Oxford University Press
• Lisa Bos, Really Strategies, Inc.
• George Florentine, Flatirons Solutions Corp., for work with United Airlines
• David Sewell, University of Virginia Press
• Max Schireson, Mark Logic Corporation
  

Also recognized was the US Army for its work on the the Battle Command Knowledge System (BCKS).

What Publishers Still Don't Get About the Web

Just a quick post to highlight two great posts by Scott Karp of Publishing 2.0 on what traditional newspaper and magazine publishers still get don't get about the web.

• What Magazines Still Don't Understand about the Web
• What Newspapers Still Don't Understand about the Web

The posts -- and the comments -- should be required reading for traditional publishers.

The Entrepreneur Age Myth

The romantic comedy Sleepless in Seattle popularized a faux statistic about age and the single woman, saying "a single woman over the age of 40 has a greater chance of being killed by a terrorist than of getting married." Another equally invalid urban myth relates to age and the entrepreneur.

Reading Valleywag you'd think that the only entrepreneurs are Digg-style party animal 20-something year olds. Or, if you really want to feel over-the-hill, consider the 12-year-old Jason O'Neil, profiled here by Forbes.

While I have no problem with young entrepreneurs, I think the media -- in its disproportionate coverage of them -- creates the mis-impression that if you over age 30 then you have a greater chance of being killed by a terrorist than of starting a company. So I was happy to find this post on Infectious Greed that brought some data to the problem.

Excerpt:

Perhaps surprisingly, the report shows that U.S. tech entrepreneurs are, if anything, older than expected. People founding tech companies over the last ten years had an average and median age of 39-years, nowhere near the age that makes for good stories about dorm room entrepreneurs -- and older than many of us might have thought.

Related articles

• Tech Start-ups Aren't Just for Wunderkinds [via Zemanta]
• Age and the Entrepreneur [via Zemanta]
• Fred Wilson post on the Mid-Life Entrepreneur Crisis
  

Elsevier's Article 2.0 Contest

Darin McBeath of Elsevier recently told me about their upcoming Article 2.0 contest which puts entrants in the driver's seat of defining the next generation of scholarly articles.

The idea is both simple and clever. Elsevier will expose, via a REST interface, the full content (in a rich XML format) of 7,500 articles and challenge entrants to build applications that leverage and present that content. Excerpt:

Each contestant will have complete freedom for how they would like to present the scientific research articles contained in the Article 2.0 dataset. We will encourage the use of XQuery, but this will not be a mandate. By leveraging these APIs, the contestant becomes the publisher and can render scientific articles to meet their needs including integrating the article into existing applications or combining it with other web service APIs.

I suspect that some of the people who read this blog (and many who read Matt Turner's Discovering XQuery) may wish to enter. First prize is $2500, second prize is $1000, and third prize is $500.

The contest starts on 9/1/08 and ends on 12/31/08. Awards will be announced on 1/31/09.

I've been invited as a potential judge and, provided they aren't too many moonlighting Mark Logicians entered, I'll be able to do so without recusing myself.

The Hippest MarkLogic App Ever -- Pop Culture Universe

You might wonder if I'm testing the new co-occurrence feature in MarkLogic 4.0 by uttering "MarkLogic" and "hip" in the same sentence, but OMG, the folks over at Greenwood Publishing are taking pop culture srsly with their upcoming new product Pop Culture Universe.

Greenwood describes Pop Culture Universe as "an authoritative, yet irresistible, digital library of information on American and world popular culture, past and present—in a package as dynamic as the topic it covers." The product integrates over 350 different titles/sources into a single online resource that covers the "icons, idols, and ideas" of popular culture.

The folks at Greenwood suggest a number of ways to use the product including studying:

• What makes a person or thing iconic?
• Film and society
• Music as history and literature
• Sports and society
• Diversity in American culture
• Art and architecture in American and World culture
• Science in pop culture
• Advertising and consumer culture
• Politics and the Presidency in popular culture
• Television and American society
  

Library Journal has a glowing pre-review of the new offering, slated for launch this summer, available here. Greenwood are also the creators of Cooking with the Bible which I blogged about here.

C u l8r!

Liautaud Lands as General Partner at Balderton

Bernard Liautaud, co-founder of Business Objects, and CEO for the company's initial 1.5 decades, las landed as a general partner at Balderton Capital in London. See the press release from Balderton, suitably in English here or French here.

Balderton manages $1.5B and has now four general partners: Tim Bunting, Mark Evans, Bernard Liautaud, and Barry Maloney.

Balderton was an investor in MySQL, acquired by Sun in January, and on whose board Liautaud sat. They are also an investor in Instranet, a provider of multi-channel knowledge applications, founded by former Business Objects engineering head, Alexandre Dayon, and former French operations chief Jean-Noel Grandval.

Related stories:

• Liautaud joins Balderton Capital (FT)
  
• Huge coup for Balderton
  

Notes from the Social Media and UGC Panel at the Mark Logic User Conference

(Not typed in real-time but summarized from hand-written notes after the event, so this is less verbatim than my other user conference posts.)

Panelists

• Max Schireson, Mark Logic (moderator)
• Stewart Shelline, The Church of Jesus Christ of Latter-day Saints
• Barbara Lange, IEEE
• Mark Uhart, CSC (works on the US Army's Battle Command Knowledge System)
• Shannon Holman, McGraw Hill Education

The twin pillars of publishing used to be authority and control. The Internet wiped out control. So now publishers are clinging to authority. That's one thing that interferes with openness to user-generated content (UGC).

(One panelist hates the term UGC, preferring participatory media.)

It's all about balancing authority and control. Particularly in the top-down organizations on the panel how do you do that? Leadership has to recognize that it's not really about a loss of control, it's about engaging participation to solve shared problems and attain shared goals. Leadership also has to realize that the fastest way to kill a discussion forum is for a senior leader to weigh in with a strong opinion.

(At LDS) not all our content needs to authoritative. We have (literally) a mountain full of genealogical data on microfiche, on roll film in Utah -- it's like 2B names. We realized that we need to capture that content and started a large-scale collaboration process to do so. In doing this process, we realized there was a huge amount of redundancy and inconsistency in the data. It's simply too big a problem to be authoritative right now, so instead we've put up collaboration tools so the members can work together to resolve problems and conflicts in the data. Basically, we need to let the audience sort it out.

It's odd that people hate taxonomy but like tagging. (Cool seemingly paradoxical observation; probably relates to psychology and our desire to self-express but not in a constrained framework.)

(At BCKS) it's all about socializing and transferring knowledge. A reservist may work at a publisher, be trained in logistics, and find him or herself serving as project manager for the construction of a school. Because people need to be able to wear different hats, we need a way for the last person who had a job to share information with the next one. Yes, we have training programs, but we want to enable direct knowledge sharing as well. Anyone can file an after-action report after they've learned something and share it.

We have a similar problem at LDS -- we have a lay ministry (i.e., not career professionals) so the job of "bishop" (local pastor) will move though a community through various members with each serving maybe a 3 to 5 year term. And while we also have training for new bishops, it certain can and does happen that a new bishop who's two months on the job suddenly finds himself having to perform a funeral. We want to be able to get him all the information he needs to be able to perform that duty. Training is part of it. Sharing and collaboration is another.

Another problem the panelists agreed on was the catch-22 in profiles and personalization. Simply put: until you tell me something about yourself, I can't get you useful information -- but until I've given you useful information you don't want to tell me about yourself.

Another interesting idea: sometimes, less is more when it comes to search results. (At LDS) we're trying to get down to 10 results per query. Not because we want to limit information, but because we know our content so well and because (through your profile) we can know the member so well, that we should be able to deliver pinpoint results.

In this vein, one panelist expressed that they were impressed with the precise results delivered in the United demonstration at the conference. (Where again the intent is not to drown the pilot in possibly useful information but instead to find *the* useful information in a given situation and provide it.)

Everyone agreed experimentation is important, but how -- asked an audience member -- do you do it?

"Uh, we use Mark Logic," said one panelist (I loved that one)

The basic idea is to start small but there are two ways to do that -- build trivial new functionality into an existing system (e.g., enable comments on news stories). The other, arguably better, way is to take a narrow slice of your content and then build out something really cool. One panelist noted that they were very impressed with the demonstration by Platts at the conference in this vein; take one deep slice and build something impressive.

Also it's important to draw correct conclusions from experiments: "just because no one clicks on the advanced search tab doesn't mean they don't have rich information needs." Maybe you just have a bad advanced search tab. Or an advanced search tab isn't the right way to solve the problem. (Where, e.g., I'd argue that iterative query refinement and faceted navigation beats the advanced tab most every time.)

Notes from Telly Stroumbis WITS Talk

Boeing technical fellow Telly Stroumbis did a great talk describing a prototype he made using data from the National Counter-Terrorism Center's (NCTC) worldwide incident tracking system (WITS) database. He basically took a public database and, in his spare time, implemented an (infinitely) better search and discovery interface than the real one that's on the NCTC site.

While not for the faint of heart (there was plenty of XQuery code), Telly showed the power of XQuery to perform tasks such as faceted navigation, auto-completion (which he pointed out would have helped Andrew McAfee in his failed attempt to find an academic paper and which generically avoids blind queries), tagcloud generation (with frequencies) -- and even report generation using XQuery. He also demonstrated the power of recursion to make code powerful and compact. My conclusion: if you know what you're doing you can write highly re-usable, very compact XQuery code and get a lot done in very few lines.

Concrete example: the key code behind his faceted navigation "fit on one screen in low resolution."

On report generation, his idea is that it's pretty easy in XQuery to build a report in XML and then transfer that XML over to a cheap reporting tool (e.g., JasperReports) for formatting. His demo in that regard was impressive.

He concluded saying that "while the Mark Logic folks hate it when they call it an XML database," Mark Logic is "actually much more than an XML database. With its analytics, you don't need a separate discovery product. With analytics and XQuery, you don't need to buy a seperate reporting product."

(He also referred to his talk last year where, on hearing Mark Logic's claim to support sub-second query response at scale, he decided to verify it, building a 10 TB prototype, running benchmarks against it, and basically "throwing everything they could think of" at the system -- and happily (for me at least) it passed with flying colors.)

Notes from Andrew McAfee Speech

Here are my notes from the Andrew McAfee speech at the Mark Logic user conference. First reacted to web 2.0 term with skepticism. Wanted to walk away from the term as quickly as possible. Silicon valley over-reaching and arrogance. Wanted to get to studying what he'd done previously ... studying large-scale implementation of IT projects and their business impact.

First time he visited Wikipedia, the term he typed was skinhead, hoping to see the failure of Wikipedia as (what he asserts as) the two completely different types of skinheads battled it out in flame war. What he saw instead was a fantastic article on the topic. And he thought: there's something different here. Maybe the 2.0 stuff isn't oversell or hype.

Definition: enterprise 2.0 is the use of emergent social software platforms within companies, or between companies and their partners or customers. Different use-case for software. Not trying to automate tasks, eliminate people, enforce rules. Instead about helping people work better together with very few preconditions. Emergent means you're not trying to dictate terms in advance.

Even though you're not trying impose structure in advance, it emerges. Web 1.0: the Internet is the world's largest library -- the problem is the books are on the floor. No hands go up in response to "who finds their corporate intranet easier to use and navigate than the Internet."

Enterprise 2.0 is a discontinuity. The technologies are novel. Not an incremental improvement: collaboration, knowledge sharing, collective intelligence, search and discovery. Used to think that the IQ of a group was half the IQ of group's lowest member. No more. Collaboration can produce great results.

He was on a panel yesterday at the Enterprise 2.0 conference. Panelists were the CIA, Pfizer, Sony, and Wachovia. Last place he would have expected open, free-form collaboration was in the Intelligence Community -- but they are big users and have great success with Intellipedia.

Facebook has business value. Why -- think about it -- do you go back and see your feed so many times per day? (Most audience members admitted that they do.) One of the landmark papers in organizational psychology was The Strength of Weak Ties by Mark Granovetter. He thinks go over to the HBS library and get it.

What have enterprise 1.0 business information providers provided? Completeness, precision (if you ask the perfect question, I'll give you the perfect answer), security/access control, structure, and consistency. Gave long example of failed search using a library search product based primarily, I think, on the fact he was looking for a landmark paper that was over 30 years old. (And ergo de-prioritized by the search engine.)

Aside: a student asked him if he should invest in Google IPO. He explained search was the ultimately unsustainable competitive advantage market. As soon as better algorithm comes along, competition is one click away. Yeah, they're red hot right now, but trust me, you don't want to invest in the IPO. She calls him from time to time from the island she owns in the South Pacific.

The student tipped him off to Google Scholar. He did the same search that failed before (Granovetter weak ties.) Got a desired result set. Published in 1973, can see number of citations (>5000). Liked what he saw: reverse citation index, related articles, faceted list of other authors, import into endnote ("this is like crack for an academic"). There were also some junky citations in the middle of the list, with typos and such. And his key point is: he doesn't care. He got so much of what he did want, that he didn't about some garbage thrown in.

Now has a new view on information services. Used to be: if I don't get what I want shame on me. Now, it's shame on you.

What are enterprise 2.0 business information consumers learning to value? Forgiveness, sufficiency, convenience, serendipity, triangulation, dynamism. Pretty sure I'd get back a different page today than I did back then in response to my weak ties query. In a 1.0 mindset, that would be disturbing. In a 2.0 one, I expect.

Metadata is becoming emergent. If we think we know in advance what the right metadata is, we're probably kidding ourselves. Preferences of business information consumers are shifting (his students start on the outside world and come into the library world only when they have to). Users / knowledge workers will vote with their feet. Only way to get them not to use Google would be to block it at the firewall.

If incumbent information providers lose volume, will they be able to maintain preferred status and pricing? Google Scholar was evidently built by one person working in his 20% time.

Jill O'Neil took him on with the first question -- some of the things you're praising Google for are information issues that librarians have been dealing with for over 100 years. Maybe good enough works for organizational psychology but in other areas, medicine, engineering, good enough isn't what we need, good enough is inadequate.

He agrees that in some disciplines completeness and precision is essential -- but are we in effect over-delivering and at what cost? "My own work is not a matter of life and death."

Asked about authoritativeness. Said to remember that pagerank-style authority is very similar to (and in fact was based on) the notion of academic citation. Only difference is citers don't have lots of letters after their name.

Our preferences are not the ones that matter here.

How to get 2.0 thinking in a 1.0 entrenched culture? He suggests (1) gathering hard data/evidence, and (2) engaging the millennials, the younger folks, in the company who will bring this different perspective. Make a task force of the youngest people around you to help drive this.

(It hadn't occurred to me that as an academic that Andrew is actually a consumer of journals and searching services and that he'd take this angle on his speech. While it was rather provocative for publishers, my belief is that the people who attend this conference *are* the change agents within their organizations and have independently arrived at many of his conclusions. So hopefully it was seen as reinforcement as web 2.0 ideas both in the enterprise and in publishing.)

Agility Panel at the Mark Logic User Conference

Here are my notes from the agility panel at today's user conference.

Panelists:

• Marc Strohlein, chief agility officer, Outsell (moderator)
  
• Larry Tunks, CIO, Congressional Quarterly
• James Wonder, director of emerging technology, American Institute of Physics
• Alex Humphreys, director of business technology services, Oxford University Press
• Kurt Cagle, managing editor, XML.COM

About 30-40% of the audience said they’ve tried agile methodologies. A lesser number characterized their companies' culture as agile.

Company culture must enable agility; it must come from the top down. There must be a willingness to make mistakes and learn from them. Hard in blame-oriented culture.

There is a fear of lack of control with agile. How can you be flexible and have control? How can you mitigate that far?

What should you do if you want agility and the company culture doesn't support it? Quit? Disagree with the notion that agile culture needs to come top-down. It's about how people work together. It's a hurdle that you go from a predictable process to a less predictable one. Way to overcome is to be clear about what that process is. Communication is key.

Trust is important in agile. It comes from having a cohesive team and a well defined vision. Trust gets built as well. Be sure your first agile project works as that will start the cycle of building trust. Trust needs to be built. Where is it built already? Can you put a core of people together and build around that?

Work hard to break down silos in the company. Programming staff speaks in the language of the newsroom, not the language of technology. Less structure is more intimidating; what is my role; how am I defined here?

It's impossible to over-stress the issue of trust. Agile is predicated on the notion that software development is unpredictable. Things are unpredictable. Everyone involved has to trust that everyone involved is doing what they need to be doing all the time. Agile can fail when there is a pre-existing system predicated on adversarial relationships. Hard issue when dealing with outside service providers -- i.e., how much is this going to cost? Should a consultancy say "I don't know" (the agile answer) or $247,000. Which is the better answer?

Can get around this by adopting outsiders on the team and treating as part of the team (i.e., don't do traditional fixed-bid consultancy but more time and materials on join-the-team basis).

Mentoring is an important part of this, comes from XP (extreme programming). Mentoring isn't just having two guys look over each others' shoulders. Mentor should act as filter and guide to the person who's new to the organization. Tap the mentor to get the person either to a point where they're comfortable or to a place where we say "Joe's not cutting it." Need a way to establish trust for people coming into the organization at any given point. This works way better than team-building management exercises where we go build bridges in the rain.

Agile needs to take into account business realities. These businesses need to make money. Getting business people to buy into this notion that the budget is either $250K or $5M isn't going to fly. What will fly is delivering projects. (Sense the agile means unpredictability and that's good but you can't go so far as to put your hands in the air and never commit to anything.)

Given a potential history of IT over-promising and under-delivering how do you get senior management to buy in? Start on per-project, per-team basis. Make some waterfall-y concessions where necessary. Have a design and development contract to make legal happy. Allow for the fact that it won't be a model of agility but there is some benefit in acknowledging that we don't know everything up front.

Agile tends to work differently on the XML front. In imperative development, you're building incrementally, component by component. With XML, you're having to shift away from building components and move closer to the idea that we are designing incrementally what we are trying to do. With XML the design part is often the most difficult part of the process. In normal databases, you have large numbers of tables, that using created along the way as part of the development process. With XML, you say I have the semantics of the workflow worked out. With XML, the workflow tends to be very much oriented to publishing and moving entities around, regardless of what those entities are.

Isn't it waterfall that requires trust and a humongous leap of faith? I'm going to make a big spec and leave me along for 12 months and I'll deliver -- that's trust! Isn't agile really about less trust -- constant iteration, constant guidance, working together on trade-offs between resourcing, funding, and requirements.

Yes, waterfall requires trust, but when you move to a cross-functional team with less well-defined interactions, where people swap hats, the marketing people need to trust that the development people are doing that for the right reasons, that they're not trying to encroach or take anything away from there. Great thing about agile is there are no six-month black holes.

Needs to come down from a conceptual discussion to a real business problem that needs to be solved.

The chasm between waterfall and agile doesn't really exist. Agile is the reflection of what has been happening over the past several years, contracting modularization of content, workload, expectations, and requirements. Don't forget that the reality of the development process has changed. 15 years ago, it really did take 2 years to build some apps. With the new reality of the Internet, with services orientation, you don't need to build all the infrastructure. Your role as a development changes from from-scratch builder to integrator. Development cycles themselves have been contracting dramatically.

What's happening to the science of information management, which should also be a part of the development of content applications?

Consider open source development. Two things are happening -- distributed open source teams just can't do waterfall and the level of abstraction of programming has gone way, way up. You now have non-programmers in some cases working on these products and systems. You can get stuff done fast in Jango and Rails, but how do you make sure you're not rebuilding the wheel every time because it's over-componentized. In effect, today, we're moving back to the centralized computing of the 1990s.

The biggest benefit of agile is the simplicity. You can communicate. You can see, touch, and feel it. Build up on it. Rather than over-design something and then whittle back in an adversarial manner.

Check out Bob Glushko and his work / book on document engineering.

I've seen agile work and I've seen it fail miserably. Two things I've seen. When it works, there is a long-term roadmap. Sometimes, marketing says need X product by Y date. Then development says, I guess that's what we need to do. Any product needs a given set of features to be successful. With agile, you think of those features on a progressive scale. You can see a lot of waterfall in agile.

Yes, it goes back to the roadmap. It's easier to develop a mature product with agile than a brand new product. What's needed is a good common understanding of what's on that roadmap. As you're creating XML you have to have some idea of what you're going to do with it, even if you're going to change it, or you won't define a structure or schema that's helpful.

XML is ultimately about data modeling, and to a lesser extent about process modeling. It's the Tao of the programming world. Imperative programming is "go." XML programming is "be." If you ask a programmer how long it will take to write a program, they will tell you any number that comes into their mind. Realistically, until they know what the model is, until then, you don't know what you're building. That model becomes your roadmap. One problem we create is to conflate the process with what we're trying to model. Just as in building a house, until we understand it, it doesn't matter what methodology we're using. Need to get to declarative, be-full modeling constructs.

Understand your model. Understand what you're trying to do.

Two years from now it will be great to see the impact of Vista on this whole debate. Won't be the first time that a Microsoft initiative changed the corporate approach to development.

"I like agile projects because they're fun." Marc Strohlein. That says it all!

Tag for the 2008 Mark Logic User Conference

Just a quick post to request that people tag their blogs, pictures, bookmarks, et cetera related to this year's Mark Logic User Conference with the tag "MLUC08" and to please tag your Twitter tweets with the hashtag #MLUC08.

That way, people will be able to find items related to the user conference more easily.

For example, you could use Summize to track tweets about the conference with this query. (OK, I'm cheating a bit as it covers both the hashtag and the phrase search "Mark Logic" but it's effective.) Real infojunkies could sign up for the RSS feed of this query.

You can use Technorati to find blog posts about the conference with this search. Or, you could find photos of the event on Flickr with this search.

Get Your Program for the Mark Logic User Conference

Well, it's almost here -- the Mark Logic 2008 User Conference, to be held at the Intercontinental Hotel in San Francisco. Hopefully, you'll be one of the nearly 400 people expected to attend the event.

• The final conference program is here (52 pages, PDF) -- or see below to see it in Scribd's iPaper format.
  

• Pre-conference technical training starts tomorrow. (It's not too late to sign up and the architectural overview -- intended for newbies -- starts tomorrow at 1:30 PM.)

• Don't miss the Welcome Reception! Drinks are on Really Strategies, makers of RSuite/CMS, starting at 4:30 PM in the Spa Terrace.

• Jill O'Neil, director of planning and communication at NFAIS, will be blogging from the conference on her blog, here (and probably Tweeting as well).
  

• I look forward to seeing you all there!
  

Mark Logic User Conference Program 2008 - Upload a Document to Scribd

Read this document on Scribd: Mark Logic User Conference Program 2008

The Convera Transition: 1Q09 Results (Updated)

(Updated: I read the earnings call transcript and made a few changes.)

One of the items I'm tracking in this blog is Convera's burn-the-ships attempt at transition from a Federally focused enterprise search provider to a vertical search platform provider.

(See prior posts: Honey I Shrunk the Company, Fast Hires Autonomy's Convera Guys, What Happens When You Sell Your Revenue, Honey I Shrunk the Company II)

Yesterday, Convera announced their first-quarter fiscal 2009 (1Q09) financials, so let's check in and see how things are going:

• Revenues of $0.40M
• Year-over-year growth of 24% compared to 1Q08 revenues of $0.32M
• Consecutive growth of 44% compared to 4Q08 revenues of $0.28M
• Ending cash and equivalents of $31.4M
• Cash burn of $5.3M
• COGS of $1.8M on running the hosted services, suggesting gross margins of -348%
• 6 new Excalibur-supported vertical sites (for a total of 45)
• 1 new publisher launching an Excalibur-supported site (for a total of 25)
  

My analysis:

• The revenue numbers are still tiny in an absolute sense; usually the law of small numbers would apply, meaning that off such tiny figures we'd see multi-hundred percent growth rates. At 24% compound year-over-year growth, it takes Convera 5 years before they're doing million-dollar quarters. Yikes.
  

• The 44% consecutive growth rate out-pacing the 24% year-over-year growth rate is usually a good sign as it implies acceleration. However, due to the seasonality of software revenues, I believe that year-over-year growth is the more reliable (and easily interpreted) growth metric. Ergo, I put more faith in the 24% than the 44% as the indicator of real growth.

• That said, an up-quarter from 4Q to 1Q is usually a good sign since software (and to a lesser extent SaaS) companies tend to have back-loaded, seasonal sales that result in a saw-tooth revenue curve where 1QN+1 revenues are often a bit less than 4QN, even when the company is experiencing strong growth.
  

• That said, Convera's 4Q08 was worse than their 1Q08, providing an easy comparison point for the 1Q09 numbers. Simply put: does 44% 1Q09 consecutive growth mean "great 1Q09," "bad 4Q08," or a bit of both? (Sometimes, the easiest way to interpret all these relative growth rates is just to make a chart, which I eventually did, and included above.)
  

• The cash burn rate is sustainable over the mid-term, but not the long-term. They have $31M in cash ($35M including $4M held in escrow) and they are burning $5M/quarter. At that rate, the cash lasts 7 quarters.

• Overall customer acquisition to-date seems good, and 25 publishers signed-up is fairly impressive. But, in my opinion, publishers are drawn -- like moths to a flame -- to the something-for-nothing idea that they can invest little and get a vertical search site. The question is will Convera ever be able to charge enough to run a profitable business? Right now, Convera is subsidizing their customers' sites to the tune of more than $20M/year. Going forward, they'll either deliver enough value to extract enough revenue to run a profitable business, or their customers will scatter like cockroaches when they try. Time will tell.
  

• The new customer acquisition/deployment figure of 1 suggests deceleration and is, in my opinion, not good.

• The company says 75 vertical sites are under contract (compared to 45 launched) with 25 publishers. This sounds good and tends to indicate broad acceptance of the strategy.
  

• But the earnings call transcript reveals that a single customer accounted for 80% of revenue during 1Q09 and this same customer was 82% of revenue in 4Q08. I sure hope Convera keeps this customer happy, since they're doing only $100K/quarter in revenues without them. This undercuts the credibility of the claims that support broad success.

• In the transcript, CEO Pat Condo gives revenue growth guidance for 2Q09 of "over 25%." It's not clear if means consecutive or year-over-year growth.

• There's a lot of happy talk about traffic both in the press release and in the earnings call transcript. I've disregarded it for two reasons: (1) it's a busy week and I've not had time to fully parse it, and (2) the kind of traffic in which I'm most interested turns to revenue and should show up in the financial statements.

• In the transcript, they say that total 1Q09 expenses net of non-cash charges for depreciation and stock-based compensation were $3.6M. They guide that the same expense metric will run between $2.8M and $3.4M per quarter in the coming year. This, and other comments, suggests they have done some heavy cost-cutting, and thus that the cash will last even longer than my previous calculation suggests.
  

Overall, I'd say the experiment is still in progress. Convera has plenty of cash to keep it running, so let's see what happens. Personally I'll be watching customer acquisition and deployment, revenues, revenue concentration, and cash.

(And, by the way, Convera, you can revise government out of your safe harbor statement; you sold that business to Fast some time ago.)

Ultimately, I'm cynical on the notion of cheap-as-chips vertical search. I believe the answer for publishers is not to focus on lowering costs, but instead to focus on creating value. For more on that riff, read (the end of) this post (Blind Eyes, Industry Analysts and Lessons from B2B) or see my slides from the recent E-Publishing Innovation Form in London, embedded below.

| View | Upload your own

John Kreisa Upcoming Speeches

I thought I'd take a moment to highlight that Mark Logic's director of product marketing, John Kreisa, will be speaking at two upcoming events.

John will be speaking at the recently sold-out Web Content 2008 show in Chicago on June 17-18, 2008. The title of John's session is Making Web Content Agile. I've never been to this event, but in browsing the program it looks quite interesting. If you don't have a ticket for this show, you'll have to wait because the next one is in Clearwater Beach, Florida on 2/17-18, 2009. Maybe John will scalp you his ticket to the Chicago show, but please give it back to him so he can get in to do his speech.

After downing a few Chi-dogs, John will reward himself for his hard work by jetting off to sunny San Diego to speak on an panel at the Burton Group's upcoming Catalyst conference. While I've known the Burton analysts with whom we work for some time (Peter O'Kelly, Guy Creese), I must say that in my prior lives, I've never worked with Burton Group before and was rather surprised to find that they are a relatively quiet, but nevertheless major-league, analyst firm -- when I spoke at last year's Catalyst show in San Francisco, I believe there were well over 1,500 people in attendance. To give you an idea of the scale, the top-level agenda has 156 sessions listed.

John is participating in session 3124, Data Management Market Dynamics: A Reality Check on XML, Open Source, and Other Topics, on Wednesday, June 25 at 4:45PM. John's going to have his hands full because he'll be on the stage with some pretty sharp folks, including

• Bill Maimone, SVP at Ingres and previously a VP in charge of a big piece of Oracle server engineering.
• Mark Drake, product manager for Oracle's XML products.
• Zack Urlocker from Sun, and I'm guessing from the MySQL side of it. (Is he the same as this pseudonym? Dunno, but Santa Cruz strikes me as a clue.)
• Peter O'Kelly from Burton Group as moderator (who has a great blog, here)
• Bob Zurek, CTO of Enterprise DB, previously a long-term IBMer who still maintains his blog on IBM's DeveloperWorks site.
• Arvind Krishna, VP of data management from IBM, who looks like he came from the Tivoli side of IBM.
• Ram Ramanathan, a senior product manager on Microsoft's SQL Server, who's been at Microsoft since 2006 and prior to that was at GE.

Good thing John has a PhD from MIT on A Programming System for the Dynamic Manipulation of Temporally Sensitive Data. Wait a minute. That's not John, that's our founder, Christopher Lindblad. Well -- and seriously -- I'm sure John will do a fine job and bring Mark Logic's unique perspective on database management systems to the panel. (And can someone please check in with him after the panel to make sure he's okay?)

After an undoubtedly fine meal at Jack's in La Jolla and a weather-permitting trip to the sea caves, John should be making the trek back to San Carlos to return to his regularly scheduled duties.

The next thing you know he'll be asking to boondogle off to the European Catalyst conference in Prague in October. Funnily enough, I think my calendar's suddenly developed an opening at that time.

Find Out What's Hot -- Right Now -- With Summize

Check out this Twitter search engine that I recently discovered called Summize. It enables you to search Twitter "tweets" (i.e., microblog posts) so you can see what's hot in the Internet, right now.

While Technorati does a great job of searching the blogosphere and keeping up-to-date with "right now" given a blog-centric definition, Summize one-ups Technorati, giving "right now" a Twitter twist -- meaning literally seconds. I'd argue that if for no other reason, Twitter's 140-character limit on tweets means that tweet searching will beat almost any other content source if you're looking to maximize recency.

You might think: what's the use in searching 140-character messages? Well, you can learn a lot. Check out a few searches:

• http://summize.com/search?q=xquery is how I learned about both the phptek conference and the new XRX approach to web application development.

• http://summize.com/search?q=Clinton will tell you if Hillary's still running ... and from a quick skim of the tweets it appears that her recent speech left some ambiguity in that regard.

• http://summize.com/search?q=mark-logic provides you with a handful of tweets about Mark Logic, from which you can that our conference is coming up and that we recently hired some high-visibility technical team members, like Norm Walsh.

As an aside, I'd argue that Summize provides yet another example of why technology executives need to spend time staying on top of all the new applications on the web. Why? It's a bit like differential equations -- the concepts build on each other and if you skip class for too many weeks in a row, it's *really hard* to try and catch back up.

Concretely,

• If you don't get blogs, you won't get Blogger
• If you don't get Blogger and blogs, then you won't get Technorati
• If you don't get blogs, then you won't get Twitter
• If you don't get Twitter, then you won't get Summize
  

Related articles

• Summize, A Realtime Twitter Search [via Zemanta]
• The 3 Minute Twitter Challenge [Video] [via Zemanta]
• Startup Review: Twitscoop, SiteScanga, Rock Hard Times, Katapulco, Twitturls [via Zemanta]

Microsoft Document Interop Initiative Video

Check out this video featuring Mark Logic's own Kelly Stirman -- appropriately wearing an official Mark Logic shirt -- on the Microsoft Document Interoperability Initiative, a launch and series of events we participated in a few months ago related to the new Microsoft Open XML file formats.

Here are a few quotes from the Microsoft press release:

Microsoft hosted in Cambridge today a number of independent software vendors (ISVs), including Novell Inc., Mark Logic Corp., ... to launch this collaborative, community-based initiative. The Cambridge event is the first in a series of labs around the world that will bring together vendors to test interoperability between their implementations of well-known document formats [...] will test interoperability between existing implementations of Office Open XML Formats and the Open Document Format (ODF) ...

[...] said Jean Paoli, general manager for Interoperability and XML Architecture at Microsoft. “The labs are designed to bring technical staff together to roll up their sleeves and test interoperability between implementations of formats and address issues that are identified either in those implementations or in the translation technologies used to work across formats.”

[...] said Andy Feit, senior vice president of Marketing at Mark Logic. “Enhancing document format interoperability between MarkLogic Server and other products in the marketplace will make it much easier for our customers to deploy applications for content assembly, reuse and delivery.”

Kelly appears near the end, so you'll need to hang in there a bit to see him. The video is in total about 4 minutes long. Check it out!



Video: Document Interoperability Initiative (DII) - Cambridge, MA

Blind Eyes, Industry Analysts, and Lessons from B2B Trade

A few days ago I read this post, The Fast and the Furious, on CMS Watch. It was the one-too-many-eth industry analyst piece I'd read on the situation at Fast Search & Transfer that basically said "I hope all this nasty business stuff goes away so we can get back to analyzing technology."

While I know and respect CMS Watch and its founder Tony Byrne, I don't know Adriaan Bloem, and I don't know the scope of his analysis. But here's what I'd say to him and every other industry analyst on this topic:

• If you are going to purely analyze technology, then I fully understand the viewpoint that business issues in any form are a distraction to that technology analysis.

• If, however, your organization produces any maps (e.g., rankings, quadrants waves, charts) that include "execution" or anything akin to it, then you have extended your scope from technology analysis to overall business analysis and -- like it or not -- you're going to have to analyze vendors from both technology and business viewpoints at all times. Simply put, if you're going to include execution in your map, then you've given up your right to dismiss business problems, uncomfortable as analyzing them may be.
  

Make no mistake. This is not another post about Fast Search & Transfer (here's my most recent one for those interested). This post is not about Mr. Bloem or CMS Watch. (Happily I'm not sure if they provide a map; maps certainly aren't the first thing that come to my mind when I hear CMS Watch, who I associate more with comprehensive reports such as The Enterprise Search Report or the The ECM Suites Report.)

This post is about my general reaction to the "failure to analyze" displayed by most industry analysts / groups when Fast Search got in trouble. From where I sit, here's a not-too-exaggerated parody of what the story looked like over about a 12 month period.

• Phase 1: Rumblings of problems, Norwegian financial analyst starts to ask questions. Industry analysts: Yeah, but they have great technology and happy customers.

• Phase 2: Quarterly misses, receivables write-downs, large operating losses. Industry analysts: they have "accounting issues" but it's OK, it's just accounting.
  

• Phase 3: Restatement of earnings. Industry analysts: OK, they have more "accounting issues," but don't worry they'll turn it around soon.

• Phase 4: Criminal investigation. Industry analysts: Whew, all this business stuff is getting boring, can we talk about technology again?

Excerpt from CMS Watch:

I hope we've pretty much heard the last of it and can return to simply discussing the merits and demerits of the technology. So even though I have this lingering image of the last scene of "Carrie" in the back of my mind, after this short break, we'll return to our regularly scheduled programming.

Again, I don't know Mr. Bloem at CMS Watch, and if his analysis is purely technical then I have absolutely zero problem with this viewpoint. But if an analyst organization is in the business where profiles, maps, or recommendations have a vendor execution component, then the above viewpoint is simply not tenable.

Why? Because "execution" is invariably related to growth claims. Example:

• Company A is $50M in revenues, growing at 50%, and profitable.

• Company B is $35M in revenues, about flat in terms of growth, and has -100% operating margins.

No analyst in the world is going to rank these two vendors the same in execution. But what if company A is actually company B with a wig and lipstick?

That, ladies and gentlemen, was roughly the situation at Fast Search, and, once revealed, nary an industry analyst made a radical drop in industry map position or had an unkind note to say about their execution. To me, most industry analysts took the Fast Search investor relations messages hook, line, and sinker.

Now, let's try and integrate the above into a single argument that serves as advice for the industry analyst business: beware the fate of the B2B computer trade press.

IT B2B trade magazines are pretty much all gone. Computerworld, InfoWorld, InformationWeek, Transform, Intelligent Enterprise, PC Week, Network World, DBMS, Red Herring, the list goes on and on. I read them every day for years. I had piles of them on my desk. We laughed when we got great customer stories and we cried when the lab panned our new product. But the magazines were everywhere. They were an integral part of IT life.

Now, seemingly in an instant, they're all gone. (Yes, a few live on as skeletons of their former selves.)

Why are they gone? Because they didn't add enough value.

I'm not sure how it evolved over time, and as an idealist I tend to believe that in the early days the IT pubs had real reporters and real labs and real value (it sure seemed to me like they did but I was young then), but by the time the Internet was posing a huge threat to their business, most of the IT trade press had degenerated to the following formula:

• Hire 20-something English majors as IT trade journalists
• Have them filter vendor press releases deciding which to cover
• Write stories based on the press releases, one live analyst interview, and one to two customer interviews
• Make money by selling advertising to the vendors
• Don't rock the journalistic boat too much because of the prior point

Net: they didn't add much value. Once the Internet and Google Alerts made press releases easily accessible, the "value add" in distributing vendor news along with an added quote or two disappeared. Some say the Internet wiped out the IT trade press. I think the IT trade press wiped out the IT press. They catered too much to vendors. They cut costs and value commensurately.

And they found themselves pretty much out of business, ironically replaced by vendor press releases (which at least you know are vendor-biased), bloggers (who weren't afraid to call it like they saw it), industry analysts, and a few hybrids like The 451 Group who live somewhere in between the previously existing boundaries.

But now that B2B trade is pretty much gone, I think the industry analysts are the next ones to experience real pressure. The M&A wave of a few years back was just the start of it, not the end.

Here's my advice:

• Keep analyzing. That's the value-add. The IT world remains complex. Customers and vendors still need you.

• Make the rough calls. Avoid the tendency towards conservative, middle-ground analysis. Take stands.

• Generate a headline or two for yourselves. The old Gartner was great at stirring the pot in the days of Mike Braude. Regain that spirit. Not only should you take stands, but generate some press in so doing.

• Restore the vision. The original Forrester had great vision. With all the M&A, everyone looks more or less like everyone else. In my mind, the vision niche is now open.
  

• Respect the intelligence of your customers. (Everyone knew that the old Aberdeen was a white paper factory. Everyone knows you have vendor clients and user clients. Take a stand on how you manage those conflicts. Disclose your vendor/user breakout percentages and your conflict-of-interest policies. Everyone's a grown-up; they understand the tension. Show them how you manage it.)
  

• Be scope consistent -- either entirely include or exclude "execution" / business performance, but don't include it when it's convenient and turn a blind eye when it's not. And if you're goign to analyze the business, make sure your team has the skills to do it. (On a few occaisions, I felt I've had to explain finance/business basics to analysts who seemed not to understand revenue vs. cash, receivables write-downs, or what it means to re-state revenues.)

• Don't be afraid to present both sides of an issue. You can see both sides of the coin, explain them, and take your own stand all at the same time.
  

Related articles

• Fast Troubles Linger [via Zemanta]
• Microsoft's Fast Search Deal: Did It Do Enough Diligence on It? [via Zemanta]
• Is Microsoft Stuck With a Norwegian Herring? [via Zemanta]