brad,

i was in the mgmt consulting profession and very skeptical of frameworks that show stacks/progressions or anything that have neat little boxes...

but i think you have articulated this very well. the craigslist example is a very well put. i think most people "get" the idea that the community/data is the key differentiator with the other stuff being "almost" commoditized...

but the question becomes- HOW DOES ONE ACTUALLY GET THE COMMUNITY GOING?

obviously, there are is no one size fits all governance/values system. different governance/values are more applicable for different communities. e.g. the free for all Myspace crushed the stricter Friendster governance model but then again a stricter model worked for facebook...

the governance/values is all about the details and ties into fred's earlier post (and ari paparo's as well) about what fits for what target audience...

Posted by nr , January 29, 2007 07:20 AM

Great post. I think we are getting towards the end of the current era of IT history, but that a new one is just starting. The successful web services of today started the ball rolling by not *just* building the data that you described, but by building the communities that build that data. An incredible feedback loop. Agree that governance and values are critical going forwards, and would add individual customization. Tailor content to meet the values etc of each individual user, e.g. x versions of the same basic ad delived to x different users based on both data that is mined behind the scenes, and data that they directly input. Amazon etc do this to some degree today, but I think they are only scratching the surface of the next era of both IT and media.

Posted by galeal zino , January 29, 2007 10:59 AM

I concur… This is well said…

I also see/sense a move to toward data intelligence on sites where algorithms are written which analyze patterns and help predict new patterns and events.

I would love to find an site that analyzed my project plans and proposed a new composite project plan based on a collection of prior plans I've worked on (ooo wait I'm slowly working at it myself… very slowly). Anyway, intelligent tools will point out areas where we need improvements and will be a competitive edge moving forward.

~ Cheers!

Posted by Rick Martinez , January 29, 2007 12:25 PM

Great post, Brad. For me though, I don't think you've taken a step up by the stack when you focus on governance -- if you are merely optimizing the data product, you're still focused on data.

As you describe in the craigslist example, good governance impacts the efficacy with which data is produced -- but the core offering (value proposition) of good governance is still data -- it may be fresher, more complete, more accurate data, but the product is still data. Same with values -- they're just good self-governance, or good governance on autopilot.

So, maybe the "content is king" phase of the IT revolution will continue for a while longer -- now entering its second term with increased focus on the efficacy with which data is produced -- and it comes with the knock-on innovation in governance and values required to achieve that efficacy. As an aside though, I must challenge the value of production efficacy as a differentiator in a peer-produced world where the community already bears most of the costs of production inefficiencies.

OK - so rather than just criticize this model, let me take a swing at what I think the next offering for consumers on the stack will be (after data) -- I think its time -- more specifically, I think it's "leisure time" -- the next phase of opportunity in IT will relate to giving consumers more leisure time. The types of opportunities I'm talking about are in industries like A-I, robotics and telepresence -- and we're probably still a decade away from any real scale there.

Thanks for the though-provoking post.

Posted by David G , January 29, 2007 01:27 PM

Interesting post.

Instead of, "What's Next?", I would have called it, "Back to the Future!".

In some ways it seems like we may be coming full-circle.

When "online communities" began, they were differentiated by their governance systems and values.

In an era where data is ubiquitous, perhaps the key differentiating factors will again be governance systems and values.

For a fascinating discussion of this topic, see Jonathan Zittrain's, "The Rise and Fall of Sysopdom", http://cyber.law.harvard.edu/10hjolt495.html

Among other things, Jonathan traces the rise and fall of system operators or "sysops" in shaping the early online communities or, as he puts it, "From Friendly Gatherings to Restless Mobs to Friendly Gatherings: The Growth, Trashing, and Revitalization of Online Communities".

Posted by Simon Venture Capital Brazil , January 29, 2007 01:54 PM

Using a "stack" to divine "what's next" may lead us astray. It seems more intuitive to look for the next constraint. For example, Umair Haque's theories about attention scarcity (www.bubblegeneration.com). This viewpoint frames the mechanisms of governance systems and values as tools that preserve attention. Many other tools exist and a "stack" may inadvertently exclude them.

It may be more instructive to look at the steps-of-education to forecast what's next (sorry, can't find the reference):

1. info exists
2. someone becomes aware of the info
3. the info is internalized
4. the info is mechanically practiced
5. the info is expressed and combined in new an unique ways
6. new info returns us to step1

Computers and networks have done a very good job helping organizations progress along these steps, however, the networks have just started to help individuals. Search engines help with steps 1-3 while items like the iPod and social sites integrate information with lifestyles to accomplish 3-5.

"What's next" are tools that preserve attention. They deliver pertinent information and help us integrate it into our work, play, and social exchanges. Governance systems and values are simply part of this equation.

Posted by Vada , January 29, 2007 04:09 PM

That's a great observation about the framework I was taught back in the early 90's about basic Computer Science. As many people point out in the comments, all pictures of processes have their limits.

It seems there are two 'basics' here though. One is that the stack was built for how a computer does it's processing through 7 levels (at least as I was taught). That's still relevant, and the point was how a computer does its processing and that's not changing yet because it's still relevant.

However, the business side is about understanding what a customer wants to get done. That's the other basic and that too isn't changing. Yesterday's long expose of Ballmer / Microsoft seemed to me to be about their issues understanding more deeply how their customers get value out of their products vs. a mentality of tools they provided. When Ballmer questions the value of Gtalk not generating any money, to me, it was asking 'how do I make money out of that?', when a better question is, 'what do users get out of this?'. Link that to cheaper IT 'stack' and you get scale like CraigsList. You don't need to charge for everything, only the things of particular need. Don't worry about what's in it for you, worry about what's in it for the other guy and the business will come. That article didn't make it sound like Microsoft was going to make the switch from that stack mentality of selling their products to owning the results their products.

What's next is more restructuring of distribution channels because they don't provide the value/cost position anymore. The relationships are indeed still just 'data', that's still there, but the implications are less friction in distribution systems and greater benefit for the consumers of data. Through the net and software more relationships can be managed directly and effectively than was previously possible. Classifieds, music, news are all re-aligning to a new way of delivering what customers have always wanted.

I think things are still just starting.

Posted by Lloyd Fassett , January 29, 2007 05:57 PM

I think you are right on with this. The transparency inherent in the Internet, combined with the commodification of the LAMP stack, moves the arena of competition. Everyone's web site will be "fast enough" or "reliable enough" -- those will just be the price of entry. Sure, the product is data, but mostly in the sense that everything in the browser is data. The customer can't usefully evaluate the quality of the end product -- if they could do that, they could recreate the product themselves -- so they instead evaluate the quality of the process that created the data product.

The question will be, are your data sources "authentic enough" or "accurate enough" or "honest enough?" You can see this happening now with Wikipedia and its clones/competitors. You can also see this emerging with critiques of Google and their attempt to prevent Googlebombing. Are the architects of the pagerank algorithm "independent enough?" What is their motivation? Those are governance questions.

Collaborative filtering embeds a concept of governance, in the form of "one user, one clickstream." Sure some of those clickstreams may be biased, but as long as they're real, not robotic, they won't pollute the results. Netflix recommendations (to take one example of many) rely on this effect (or, more precisely, we rely on the recommendations because we believe the clickstreams of other Netflix customers, and thus the resulting recommendations, are genuine).

Interestingly, many industries that deal in intangibles like advice and evaluation struggle with this, and governance issues have moved to the forefront recently. Whenever you see a company describe its "code of conduct" or talk about "compliance," they are dealing with governance.

Posted by Tom Hughes , January 29, 2007 08:48 PM

It's an interesting point. Obviously, it's less and transition than a merging: without the data, who cares about the platform?

But the platform is no doubt worth perfecting. With BricaBox(tm) we're trying to perfect the user generated nanocontent platform, but that's just one example.

Other companies trying to perfect the management of user generated data abound and will surely influence "What's Next."

Posted by Nate Westheimer , January 29, 2007 09:03 PM

Stacks are great metaphors and great ways to compare technical approaches, but I don't think it captures everything in play. And placing governance as a block in a stack, I think, misses a lot.

The technical stack is just that. I build apps within the walls of at least 4 different stacks (more, depending on how purist you are). Governance of any sort can be instantiated in any of them.

And again, depending on how you bend the term 'govern', you're reaching into customer service, marketing, and operations, at least. This is no longer technical, outside of enabling technology.

Posted by Jamie , January 29, 2007 11:41 PM

interesting post brad :)

i'd agree, the underlying tech stack probably matters only in terms of cost (both to buy and also to manage, tho likely more focus on the latter).

what's next?

i agree data is likely the next layer... but more specifically, it's around data interchange & APIs, rss & microformats, and some kind of event model that works between sites & services to enable triggers for other processes. the overall infrastructure is mostly there, just not very rigorous / reliable yet.

what else matters?

well, if we're not talking about machine-enabled functionality, i'd still say these 3 things matter for consumer-facing services:

1) simple & useful product features -- (note: "features" perhaps similar to your "data", but i prefer the user-centric view). these days, product development should be super iterative based on initial & ongoing market research, user feedback (qualitative), web metrics (quantitative), and lots of a/b testing.

2) effective & measurable marketing -- these days the marketing mix is broader than ever, but also more measurable than ever. however, marketing skills/delivery are a lot more tech-driven. SEM, SEO, widgets, blogs & RSS, and good ol' links & email. even some PR still works, tho less critical than before i think. built-in viral features like Geni.com are probably the best marketing... that and simple word of mouth from customers that love the product (see #1).

3) cost structure / revenue / rebates -- how does the company attract customers & generate revenue? do they offer the service paid or free? how about less than free? (ie, rebated revenue to users). finding an effective monetization strategy and using that to reduce costs (both fixed operational & variable customer acquisition) is still relevant, even in a "freemium" web 2.0 world. users count, so even a "free" service can acquire value thru usage & content. however, users + real money work even better.

lots of other interesting stuff happening, but for consumer-facing services those are the important ones. security & privacy for certain companies, pretty UI for others, scalability & infrastructure for high volume sites.... but almost everyone has to deal with the 3 noted above.

- dave mcclure
http://500hats.typepad.com

Posted by dave mcclure , January 30, 2007 12:19 AM

Some thoughts:

If you look at the progression of IT, there is a very clear pattern: the "next step" always builds on the foundation provided by the previous step.

Systems software was enabled by hardware. Application software was built on system software. Application software layers on top of system software, and so on...

Which brings me to point #1: There is always room to build upwards when you have a foundation to build on.

Just as IBM was wrong about a handful of computers being enough for the whole planet, any predictions of the end of progression in IT will be equally wrong no matter when they are made.

Another critical thing to note is that the previous step typically becomes invisible after the "next step" becomes established. For example, we used to have to care about what hardware we were using because the system software had certain requirements. For a typical user today, the web simply doesn't seem to care -- you just need a browser.

And that is point #2: The history/baggage/requirements of the IT landscape always becomes simpler for the average user once you reach the "next step". Furthermore, the "next step" often solves problems that the average user was only just becoming aware of.

So looking at today's world of IT, data is clearly front and center.

Which brings me to point #3: I believe Brad is _exactly_ right to assume that the next step will build on that data. The history of IT makes that almost certain.

I strongly believe that today's tools for utilizing data are inadequate. There are a number of critical underlying reasons for this. One of the biggest is that current systems are built on concepts of sparse and static data. In today's world, data has become superabundant and is highly dynamic.

Addressing this IS the next IT frontier, and its applicability is incredibly broad.

Brad: I'd very much enjoy trading some further thoughts with you if you'd care to drop me an email...

Posted by fewquid , January 30, 2007 01:03 AM

Very insightful, although I tend to agree with one post that governance could also be categorized as a way to improve data volume and quality. As for the "end of IT history" (isn't the prediction of the end of history always the harbinger of a new phase of it?), it seems to me that one major next step is the collection and interpretation of data from the real world. Traditionally, all data is entered and organized symbolically by people. Increasingly, we have sensors and cameras everywhere and the data can come through directly. As another post mentioned, we need algorithms and techniques for analyzing data, particularly if we want this raw data in a usable (meaning symbolic) form.

Posted by Dave Jilk , January 30, 2007 09:28 AM

Mr. McClure hit it right on the head, my *hat* is off to you :)

I really think RSS will have an integral role in the next generation of applications. I look forward to using it in our application to auto-notify sellers of interested buyers.

Craigslist really is a great service, but I feel that a niche site dedicated to each of the categories truly is the way to go. The "personals" section is obviously less popular than MySpace, the "for sale" section is less popular than eBay, and so on.

As a whole, however, Craigslist has the ability to do a little bit with a ton of categories... In comparison, a service like Etsy can do a whole lot with one category.

Fred Wilson really got me thinking about that when he posed the question "Doesn't Craigslist already service that market?" in a recent e-mail.

Best Regards,

Robert

Posted by Robert Dewey , January 30, 2007 10:09 AM

Brad-
This must be one of the most significant blog posts I have lately read. Thanks for a well-thoughtout point.

I think you are on to an important idea. Thoroughout the comments, I see many are asserting that governance is more important once there is community and data, but we've started seeing independent governance moves, as well. MyBlogLog is an example, and so is RapLeaf. They both sit outside of data and community and help govern it, in the process acquiring their own data and community.

Posted by Cem Sertoglu , January 30, 2007 02:08 PM

Brad -- This is a really important point. As you know, I've long argued that "Data is the Intel Inside" is the hidden underbelly of Web 2.0, and in a post the other day, I argued that there's a cycle of open and closed technologies that progresses according to Clayton Christenson's "law of conservation of attractive profits." But I'm not sure I buy the idea that it's governance that is the new differentiator.

Craigslist gets its "lock in", to the extent it has it, from its critical mass of data. Its lightweight governance (i.e. its business processes) are definitely a competitive advantage, and you have a great insight into how important that is. But if someone emulated those processes exactly, they still would have a hard time displacing craigslist, because of the critical mass of listers. That is, governance gives advantage, but not proprietary advantage. It's not comparable to the kind of lock in that IBM had with hardware, that Microsoft had with software, and that Web 2.0 companies are now achieving with their proprietary databases.

That being said, one of the commenters on my post suggested that when data is open, proprietary algorithms can still provide lock in. And governance is a kind of algorithm. So perhaps you might broaden this concept a bit.

Commenters also pointed out that infrastructure may be a source of lock in. WalMart is a good example. Their integrated supply chain is a key part of their competitive advantage. It's a combination of data, business processes, and infrastructure. The question is when you open up each of these, what advantage is left?

When the IBM PC opened up hardware, advantage moved to software. When open source opened up software, advantage moved to data. When some future open data movement opens up data and makes it mobile, what is the source of lock in?

Posted by Tim O'Reilly , January 30, 2007 03:37 PM

Thanks for introducing the word governance to my understanding of what is happening on the internet right now. It puts what we are working on in perspective.

After reading this post, I think the strategic positioning of Second Brain will be even more focused on governance of data from other web services. There is no point trying to copy functionality from other services that already work well. Instead, let the users see everything through our service.

Posted by Lars G. Teigen , January 31, 2007 06:38 AM

Wow - these are all great points. I agree that the stack metaphor breaks down when we get to data governance and values, and that time is going to become an increasingly important dimension. I agree with Dave that a user centric view is the only one that really matters, but I disagree that that means that features are more important than data. I think of features as software functionality, and differentiation in web services as a function of data richness and utility (from the perspective of the consumer). This probably a difference in semantics not substance.

I think Tim's final question is important but I would change "lock in" to value. In an open data world what is the source of value? When we are drowning in data, value will come from the way the data is gathered, organized and presented.

Tom makes this point better that I did. His example of Netflix and the inherent integrity of collaborative filtering is a good one. We will begin to care deeply about the governance system that produces the data we consume because, our faith in the way the data is produced may be the only way we can make meaningful judgments about the integrity and authenticity of that data.

Posted by Brad Burnham , January 31, 2007 08:45 AM

Brad:
Terrific post. I think you can compare the Craig's List governance to the NY Times editorial governance - which is what we buy the TImes for as much as the "data" provided by the reporters. And the level above that is the board/shareholders. So after governance comes ownership? Perhaps spreading the wealth again (a la Revver???)

Posted by Howard L Morgan , January 31, 2007 10:01 AM

Hi Brad, I think you are 100% right here:

"I think Tim's final question is important but I would change "lock in" to value. In an open data world what is the source of value? When we are drowning in data, value will come from the way the data is gathered, organized and presented."

I have drawn a nice little diagram representing that fact here:

Manual Trackback:
http://www.touchstonelive.com/blog/2007/02/brad-burnham-asks-whats-next.html

Posted by Chris Saad , February 5, 2007 09:16 PM

Brad – great post, as usual…
though, I am not sure I agree with the premise that the “progression” goes in one direction only (ie ‘up the stack’). Taking the queue from Tim’s post: Christensen paints a picture of various elements of a ‘stack’ disintegrating and re-integrating (ie re-aggregating) over time with value moving back and forth between those elements (thus his “law of conservation of attractive profits” referring to adjacent elements of the stack going through cycles of commoditization.) So, there is no “end of history” – just cycles of value creation!
An obvious example: the PC disintegrated the computing-device stack and pushed value to some of the individual elements of the stack – benefiting Intel and Microsoft among others. More recently, the ipod (and to some extent the mac) has reintegrated the stack arguably shifting value back to the ‘hardware’ or ‘device’ as an aggregated whole. It would be interesting to think about how value gets shifted within the internet services stack, which would include (among others) your elements of data, governance and values representing individual elements of the stack, and maybe aggregators representing integrated solutions(?).

Posted by Salman FF , February 5, 2007 10:07 PM

Brad – great post, as usual…
though, I am not sure I agree with the premise that the “progression” goes in one direction only (ie ‘up the stack’). Taking the queue from Tim’s post: Christensen paints a picture of various elements of a ‘stack’ disintegrating and re-integrating (ie re-aggregating) over time with value moving back and forth between those elements (thus his “law of conservation of attractive profits” referring to adjacent elements of the stack going through cycles of commoditization.) So, there is no “end of history” – just cycles of value creation!
An obvious example: the PC disintegrated the computing-device stack and pushed value to some of the individual elements of the stack – benefiting Intel and Microsoft among others. More recently, the ipod (and to some extent the mac) has reintegrated the stack arguably shifting value back to the ‘hardware’ or ‘device’ as an aggregated whole. It would be interesting to think about how value gets shifted within the internet services stack, which would include (among others) your elements of data, governance and values representing individual elements of the stack, and maybe aggregators representing integrated solutions(?).

Posted by Salman FF , February 5, 2007 10:16 PM

Brad – great post, as usual…
though, I am not sure I agree with the premise that the “progression” goes in one direction only (ie ‘up the stack’). Taking the queue from Tim’s post: Christensen paints a picture of various elements of a ‘stack’ disintegrating and re-integrating (ie re-aggregating) over time with value moving back and forth between those elements (thus his “law of conservation of attractive profits” referring to adjacent elements of the stack going through cycles of commoditization.) So, there is no “end of history” – just cycles of value creation!
An obvious example: the PC disintegrated the computing-device stack and pushed value to some of the individual elements of the stack – benefiting Intel and Microsoft among others. More recently, the ipod (and to some extent the mac) has reintegrated the stack arguably shifting value back to the ‘hardware’ or ‘device’ as an aggregated whole. It would be interesting to think about how value gets shifted within the internet services stack, which would include (among others) your elements of data, governance and values representing individual elements of the stack, and maybe aggregators representing integrated solutions(?).

Posted by Salman FF , February 5, 2007 11:44 PM

Brad,
Great post. I was sitting down with a Yale PhD in computer science last night and the ideas we were discussing were right in line with this. Who knows, maybe it will develop into a new venture.

Think about how this affects something like search. Today, the governance system for search (at least google's page rank algorithm) runs based on data and pre-programmed governance. The google system crawls the web and based on the data it collects there is then an algorithm that returns search results (developed from a system that figured out the importance of academic papers through reference citations). The governance system for how data enters the system and how it is used is hard coded by google engineers.

Contrast this with the emerging ideas of social search. Here the data may be gathered and contributed to the system by individual users. There may be a governance system that determines how the contributions from users are valued into the system and that system may evolve over time based on user behavior and contribution value. Ultimately, this might provide search that is more specific and more intelligent. The two systems would both have data and the data is valuable, but the governence system could provide a large differentiating factor (building this type of system can be very complex hence the interest of a Yale comp sci PhD).

So there's value in data, but the value could shift to the governance system that evolves how the data is used and in the long run value comes from utility right?

Posted by Sean Glass , February 14, 2007 06:01 PM

I agree. I think the only thing left on top is "values" (i.e. what ethical and moral framework supports the highest utility to human beings). Why have democratic free enterprise societies outlasted communism? Because - despite all their faults, these societies harness the collective aspiration, talents and resources of their people with far greater efficiency. Moral values underpin these societies: such as honesty, free will, mutual obligations and freedom of speech. These observations provide insights into the root cause of grinding poverty and other massive global problems. For example, the correlation between the social status of women and extreme poverty throught the world is too significant to ignore. Anyway - I think I have said enough!

Posted by Anthony Peterson , May 2, 2007 02:06 AM