Category: Defrag 2007 (Page 1 of 2)

Well, a few days have gone by since Defrag ended, and I always like to let things settle a bit before I have my final say on a conference. This is that post, which also captures some highlights from the rest of the second day’s program.

The "Next-Level Discovery" panel featured an eclectic group, including an IT professor from Berkeley (Marti Hearst), a general manager of Dow Jones Factiva (Lou Paglia), the inventor of Jabber and now working at Wikia on a project to create an open search engine (Jeremie Miller), and the CEO and cofounder of code-search site Krugle (Steve Larsen).
Brad Horowitz of Yahoo was the moderator. Here’s a sampling of what we heard:

Hearst said discovery is about finding things you weren’t expecting to find. But, today, in the enterprise, "people are about search." She said today it’s about navigating, whereas in the future it will be just "saying what you want." Hearst said we’re starting to see the convergence of two trends: massive collections of implicit user behavior, and better handling of longer queries.

Lou Paglia is the GM of the corporate research division at Dow Jones Factiva, which "normalizes thousands of news information sources from around the world for corporate research people." He said that in the enterprise today, Facebook is a way of discovering, as is checking stocks, for example. Jeremie Miller, open source guru that he is, said that, for next-level discovery to happen, we need better connection of independent parties. "The fabric must grow there."  Steve Larsen spoke of efforts to make search smarter, and that there are now more than 100 vertical search sites (his company being one example). "These could do to the major search sites what cable TV did to the networks."  These vertical search sites are about specific interests of consumers, he said. Because they constrain what one is looking for, you can assume implicit things about them. And you can present results that provide context, the real meaning of discovery."  Horowitz said Flickr (a Yahoo property) mines the "interestingness" of people and "reflects that back to the community" — and that this is "better than just explicit things like ratings." There’s value in these implicit behaviors, and it lowers the barriers of participation.

In the discussion period afterwards, Factiva’s Paglia spoke of the "hidden web" and how it’s a big area of need in the corporate world. He said his firm is seeing a change in the paradigm. "How do you get beyond the article? What’s the actual knowledge in the news? What about the data element to help managers make decisions? We want to help people get to where they want to go."  Krugle’s Larsen said his company’s code search engine can even link all the way back to the market requirements that caused a given piece of code to be developed. Wikia’s Miller said "the dark web will cease to be known … information wants to be found." Doc Searls asked from the front row, "What about searching old web pages?"  Miller said "these are a real treasure, but it’s very hard." Factiva’s Paglia, however, said "we’re getting there." Another audience question was about social search. The panelists noted that one company working in that area, Me.dium, was here, and that Microsoft has such a project. Horowitz said Del.icio.us (also a Yahoo property) is "headed more this way."  Esther Dyson asked whether more structure was needed for search. Horowitz said it depends on who’s providing that structure. Larsen said visualization is one way to give search more structure. Paglia asked, "Why should we have to search at all? You should just be able to log onto your CRM and be told what things you need to action on today." Horowitz closed the panel by commenting that "it’s amazing how ossified search results are!" People are locked into a certain look or way that search results are presented to them, but implied this needs to change, to get better.

The Elephant in the Room
A special add to the program, who flew in just for the second day of Defrag, was technologist Kevin Marks of Google.
Formerly with Technorati and Apple, Marks said his area of responsibility at his new employer covers basically everything except search and ads. The word was that he came to event in response to one of the impromptu Open Space discussions the previous day, called " Closed Private," which took issue with Google’s new "Open Social" spec — as it might relate to the enterprise, at least.  And the word was Kevin showed up to explain his side… 🙂  A great recap of his talk was done by my friend Sean Ammirati — it’s here on  Read/Write Web.  Also, for those of you into the details, here’s Kevin Mark’s slide presentation. A key thing many people may not be noticing is that two of the already announced partners for Open Social are Oracle and Salesforce.com. So remember this, friends, as Kevin said: "Business is social, too."  I say that’s a key takeaway from this event. And I think it’s great that Google chose to show up here. It really speaks to the significance of this first-year gathering called Defrag.

And From Sponsor Yahoo….
Jeremy Zawodny of the Yahoo Developer Network gave an interesting presentation, saying his areas of interest were Open Source, Social Everywhere ("the ‘Net always has been"), Web Services, and Web Platforms.
His "Future" slide shows where he’s focused, so I include it here.  For more, see Jeremy’s blog.

The Guys With the Money
How could there be a conference without a VC panel?  It just can’t happen. But we were very lucky to have partners from arguably three of the most successful firms investing in Web 2.0 technology today: SoftTech VC (Jeff Clavier), Union Square Ventures (Al Wenger), and the Foundry Group (Brad Feld). Here are some highlights:

Feld said there’s a notion now that it only takes $2-3 million in invested capital today to become a profitable business in this space, well below what’s traditionally been the case. Clavier noted that "the $100k you spend today to develop your prototype is like $1 million ten years ago." Wenger reminded us, though, that "if you fall behind after you hit a vein, that’s problematic." 

Feld said his first investment as a VC was only $50k (Net Genesis), and the he only invested $100k in Harmonics (which developed Guitar Hero). "A typical seed investment now is $250-500k, where it used to be $5 million," he said. "It’s faster now to get to a proof point." Later, we learned that Clavier really doesn’t invest in "enterprise software" [though one would assume he’s interested in social networking as it applies to business, or why was he at an event that was more enterprise focused?], whereas Feld chimed in that "he loves enterprise software!"  He thinks it’s amazing that so many VCs "don’t like it anymore."  Consumer-oriented Web 2.0 deals have certainly taken much of the attention away, but a major trend is emerging whereby those technologies are now leading advancements in enterprise productivity.  And Feld inserted here his take on that, which I think is a major takeaway from Defrag: "I like what I’m seeing in enterprise IT spending increase projections for 2008 — double digits, maybe 20%," he said. Most VCs want to fit into one or the other category (consumer or enterprise), "but I’ve never been able to separate them." Wenger noted the huge difference in the two is that, in an enterprise business, "you must focus on building a sales team." No small undertaking, to be sure, nor cheap.  Feld noted that, of course, very few companies become successful on both sides.  His advice to startups: "Don’t overcommit too early which side you’re going to focus on."

As the afternoon’s sessions continued , we heard a very good presentation by Nova Spivack, founder of Radar Networks, called "Towards a Usable Semantic Web". He referred to the "third decade of the Web" as being about "transforming it from a file server to a database."  For a selection of his slides, see my Defrag pix on Flickr. Nova’s firm has introduced a new service called Twine, which he says represents the next level in knowledge management —  "knowledge networking." Nova defines that term as "social networking + semantic web + collaboration + search."

My Recap of Defrag Version 1.0
This first-year event was great, in nearly every way.  In fact, I can’t think of a single thing I’d change when I fill out the comment form — except maybe to publish a list of attendees. [Unfortunately, the sponsor wiki sites were not fully populated with everyone in attendance — which amazed me. Perhaps they were expecting everyone to sign up on their own?  They didn’t.] It was really the people attending this event that made it such a rich expertience.  That was the key, and I say hooray to the producers for getting so many key players out for this one.  At one point, producer Eric Norlin said it was like "getting the band back together" from Esther Dyson’s great PC Forum events, now retired. I mean, when you have people like David Weinberger, Doc Searls, Chris Locke, Esther, Jerry Michalski, and so many other early Internet pioneers in the same room, it can’t help being a thought-provoking experience!  I was taken by what Steve Larsen said about the event. He told me: "I was at the first PopTech conference, and I also attended the early PC Forums. This event has the same feel."  Steve also pointed me to a blog post about those early days of this core group of Internet and online community pioneers, which gives you a feel for what things were like back then. It’s good to look back, but I also think Defrag made some history of its own on these two days in Denver in late 2007.

As I spoke to Eric Norlin on the second day, I think he and
a colleague really hit on the essence of Defrag #1: "It’s all about
questions, not answers."  That really wraps it up as best as anyone can. And I know that I’ll be back again next year for more.

Tuesday, the final day of Defrag, opened with Andy McAfee of Harvard Business School speaking on the topic of "Enterprise 2.0." I love learning from these Harvard guys. They have such a…delivery, or something. A way of making the topic seem, well, just so compelling and fascinating. They must give these guys stage training. They just keep me forward on the edge of my seat — I love that. Andy referred at one point to his colleague Clayton Christensen, another great thinker and presenter I’ve heard a couple of times. But I must say, after hearing Andy for the first time, he’s in the same category. (An insight of Christensen’s that Andy dropped here, one he also uses as a guidepost, is this: "Managers are voracious consumers of theory.")

"I think we’re at a time parallel to ten years ago," said Andy, "the first stage of the Internet." What he’s been hearing a lot from Corporate America lately is CEOs asking their CIOs, "What is this Web 2.0 thing, and what’s our strategy?"  The implication is this trend we’re calling "Enterprise 2.0" is for real, and I guess that’s why it’s a major underlying focus of Defrag.

McAfee put up a slide that he labeled "the knowledge worker’s view of the enterprise." 
He said this diagram provides a foundation for understanding how the new Web 2.0 technologies help "build ties" in the organization. For the center of the target, "strong ties," the prototypical tool is the wiki. He noted that email is "broken in many ways," and that wiki technology can be a better, simpler solution for collaboration. "I keep coming back to what Ward Cunningham says: ‘What’s the simplest thing we could possibly do?’ "

Andy said that a more important area of focus, however, is the next concentric circle: "weak ties."  He quoted a sociologist that says we spend a lot of time working on our strong ties. "That’s okay," said McAfee, "but they’re not likely to be sources of non-redundant info." We need networks to build our weak ties. "That’s exactly what we should be doing inside organizations," he said. "And this speaks to the strength of Facebook inside companies." [I wanted to ask at this point why, then, are some companies banning its use at work? We can only assume they’ll eventually see the light, or create their own internal Facebook-like networks.] "The prototypical tool here is the social newtorking system," said McAfee. "It builds bridges to other networks…and what emerges is access to non-redundant information."

Bloggers Are ‘Brokers’
So, what of the next circle, the "potential ties"?  This is the area of the unknown pool of people we all know are out there that we want, or hope to, connect with to expand our worlds. "But how do we find them?" In this classification of ties, Andy said the key is the broker. "Brokers have real value in bridging people." And the prototypical tool here? That, he said, is blogs. You would think he’d be referring to internal company blogs (and Andy does support that practice), but he went beyond what happens inside the firewall. "The blogosphere is an incredible way to expand my knowledge and my network," he said. "And it’s not terribly time intensive. "The typical knowledge worker is not going to spend hours reading blogs, but they can certainly learn how to monitor the most important ones to them."  He showed what he called a great example of an internal, idea-sharing blog — that of a construction-related firm called IntraWest. He also spoke of an internal "working prediction market" at Google. FInally, he showed what he said was an excellent example of an intranet, that of Avenue A | Razorfish — and, since we rarely get a glimpse of company intranets as outsiders, I’m including his slide here, at least so you can see the types of information included, if not the detail.
Blog posts receive a prominent position at the right. McAfee said the base technology for this site is MediaWiki. "Companies are deploying technology at all level of the bullseye."

We Need Strangers
But what about that other, final circle — the one McAfee labeled as "None"? Why was it included? He said this is the group that could be assumed to be "people that have nothing to say to each other." But, he said, let’s stop here. "Think about capital markets. This is a way for complete strangers to value a company." And that brought him back to the subject of prediction markets. "There are lots of these on the Internet," said McAfee. "For elections, sports events, many things." He also cited, and showed, the example of the Hollywood Stock Exchange, which is an entertainment prediction market where you can buy, sell, or hold your favorite movies and stars, in a completely virtual marketplace. This was the point where McAfee laid out his greatest challenge. "Why aren’t such markets employed more inside corporations? Especially since firms spend so much time and effort on forecasting. They work!"

Prediction Markets Ascending?
He said he really believes that "we’re missing a play here — it’s worth experimenting." In the discussion following, Andy noted it’s not necessary to have real money at stake in such virtual markets. "Hollywood Stock Exchange is complete funny money." But there are advantages from these markets — "It’s irrelevant if real money is at stake." Google, he said, offers incentives for its employees to participate in its prediction market — specifically t-shirts, which are a very sought after prize inside the firm. "Google employees are not wanting for much," he said, "but I’ve been told there’s a real t-shirt economy going on inside Google." He also noted, in response to a question from an Intuit employee who said her company’s internal blog has cut company email in half(!), that "there’s an authoring urge by certain people, and every company has them." They drive the success of these internal blogs and other tools.

McAfee concluded his talk by saying that this concept of "ties" provides a foundation for understanding how the new Web 2.0 technologies can help create networks between people that would otherwise not form. "Managers want this playbook." He said that some people think the Internet tends to make things "too much alike, too homogenized." But he’s seeing just the opposite. "The net effect of these new Enterprise 2.0 technologies is that they will help companies and workgroups differentiate from each other."

The afternoon session on Monday focused on several interrelated topics, starting with a talk by Alex Iskold on "Structured Attention." I caught the beginning part of this one, then had to jump over to a concurrent session. Alex said there are three types of recommendation engines: personal, social (collaborative filtering), and genetic (like Pandora). "But the problem is they’re silos," said Alex, "and we don’t own them." Thus, the gist of Alex’s treatise it that we need "Web Wide Attention." The way it stands now, "No single site knows all about me." He said that has to change.

The other breakout session I then joined was one by Peter Semmelhack, founder of Bug Labs. This was a session billed as "Defragging Gadgets: Community-Driven Electronics." But Peter titled his talk "Hardware Innovation 2.0" — positing "what if we made all technologies available in Lego-like fashion?"
His startup company is developing an open platform for physical devices, and 81 snap-together modules, along with web-services based interfaces, resulting in three million combinations. Peter spoke of The Long Tail of Gadgets, which he described as "millions of devices for the few instead of a few devices for the millions."
Thus, his company is talking about niche, customized devices, to meet the application needs of narrow markets (even individuals). He even went to far as to suggest that this concept will disrupt the consumer electronics business, calling it "the age of personal electronics." He said later, however, that Bugs Labs’ target audience is only early adopters (and, I would say, only the geeky ones), "by no means mass consumer."  Thus, I fail to see how it would really disrupt an entire industry. And no one asked about where product design fits into the equation, once he showed us how the modules snap together, appearing rather kludgey. (He showed CPU, digital camera, LCD, and GPS modules, for example.) It is nonetheless an interesting concept, and Peter did say his firm is seeing a niche market for firms that are now locked into proprietary platforms and have a need to quickly and easily convert to an open platform. He also said they’re seeing interest in the area of home automation.

Esther Dyson took the stage next for a brief discussion on Attention. The problem of data privacy is a huge one, she said.  There’s data being misused for criminal reasons, but also many other semi-legitimate uses, which may not be criminal, but are still very bothersome. "Consumers really have no idea what’s going on," Esther said. "No one reads disclosure statements." 
She also commented on the strange dichotomy that advertisers can deliver such personalized messages to you, but "why can’t they deliver your personalized data to you, as consumers?"  She also commented on how consumers have changed: "Facebook and other scoial networks have changed the ambient notion of what a ‘profile’ is. We’re beginning to understand ‘selective disclosure’ — and even moreso if you’re younger." Esther also put forth the intersting notion that Facebook is "advancing narcissism — which isn’t really making advertising any better — people will enhance their profiles to get better treatment from advertisers."  She closed by saying that "Consumers want to be able to be selective with letting out their data, but they don’t know how." But, as consumers, she said "You can demand things. You have the power to demand metadata that is meaningful."

Another great talk at the end of the afternoon was by Doc Searls, on the subject of "Vendor Relationship Management."  This is all about "customer reach" rather than "vendor grasp," he said.  "Why can’t you create your own profile data just once? Why is the free market still ‘your choice of silo’?"  Doc’s assessment is that the Industrial Age hasn’t ended yet. "And things won’t change till companies stop wanting dependent customers instead of independent ones."  He pointed how ridiculous, for example, the customer agreement of Verizon is, which is some 10,000 words. "All this to basically say you have approximately no rights at all." [Love the way with words Doc has.] "In real markets," Doc said,"customers do more than just consume. What if we could help vendors instead of just carp about them? We need to manage our preferences across whole markets." He then went on to describe his Project VRM (for vendor relationship management), which he launched in his current role as a Fellow at Harvard’s Berkman Center for the Internet and Society.

A week after all the buzz about Open Social, this next session was bound to be a popular one. Then again, all that hype was getting way out of control, too, which prompted conference producer Eric Norlin to put forth what he thought was more significant news last week.  Nonetheless, social networking is a part of our lives now, and this highly literate audience of tech and online community types was naturally all over this discussion of where social technology is headed. This panel would address some of the ways it’s playing out now inside the enterprise. [Excuse the crummy photo — I must have been shaking in anticipation.]

Aaron Fulkerson of Mindtouch talked about his firm’s open source wiki and
application platform for communities and enterprises, which is seeing rapid adoption. (I blogged about this previously here.) He said wikis are easy-to-use tools for creating
web-based communities, improving employee
communication and collaboration.  Tens of
thousands of organizations use MindTouch, Fulkerson said,including Microsoft, Fujitsu,
British Petroleum, and Stanford University. The company’s online community of users is at OpenGarden.org.

Dawn Foster, Director of Developer Relations for Jive Software, spoke about her company’s Clearspace social community software. She said her firm now uses it for their entire intranet, and that it’s all about "open collaboration." She said she gets virtually no email internally anymore — "it’s all collaborative IM."

The guy who stole the show, though [what is it with those British accents? …and weird hats], was Charles Armstrong, CEO of Trampoline Systems, all the way from the UK. He said his background is in anthropology, which led him to focus in "social computing." He’s studied how information gets distributed in small communities, and he said "we’ve been phenomenally successful in doing that." Trampoline builds analytic systems that mimic that activity. He described his system as kind of an "upside down social networking system." He said it integrates email, document management, and other information sources, to create "employee-facing tools to find expertise." Trampoline’s customers, he said, include a UK government agency, Raytheon, a TV network, and a global management consulting firm soon to be announced.

"Do we really need social networking in the enterprise?" asked moderator Eric Norlin. "Yes, because there are lots of data silos that are completely disconnected," said Mindtouch’s Fulkerson. Trampoline’s Armstrong said, "Studies show that work gets done in enterprises in informal networks, and social networking recognizes that and puts it to use."  He noted we’re just at the beginning of all this, but that "brave companies are diving in." One relatively conservative customer company of his firm, he said, has had his system running for only three months, but already has 20,000 people using it.

"What about the younger demographic in all this — what affect does that have on the enterprise?" asked Norlin. Armstrong strongly feels that, to attract and retain younger workers, social networking technology is mandatory. Fulkerson added that, "What starts in IT in the way of wiki adoption is now rapidly spreading throughout the organization."

Another question: "How do you deal with the information overload problem?" Trampoline’s Armstrong: "We help you defend yourself against it." Mindtouch’s Fulkerson: "There’s no such thing as too much information, just not enough tools to filter it."

And a final sobering question: "Is there a possibility for individual employees to remove their info from a  corporate social networking system?"  Not under current law, said Armstrong. "We do have ways to split your corporate and personal data, however." Fulkerson said his firm also has such tools. "But what if I have an idea over lunch?" yelled Chris Locke from the audience. Armstrong: "If it’s put up on the wiki or email or social networking system at work, it’s company property."

UPDATE (11/6):  For some great insight into social networking in the enterprise, see my colleague Alex Iskold’s post today on Read/Write Web, resulting from a conference last week back East. (Alex is also here at Defrag.)