Is Web 2 causing the dissolution of the mega-corp?

by Kurt Cagle

I've been wearing my entrepeneurial hat of late, and I have no doubt that its skewed my perceptions somewhat, but I've noticed an interesting phenomenon that's occurring somewhat out of the main spotlight of the media circus that's the tech industry. In the mid-to-late 1990s, while it was possible to start companies out of a garage, and SOHO offices became something of the rage, most such companies so started were either services oriented companies that were intended to remain relatively focused on the local communities, or they were companies with plans to grow to be the next Microsoft in ten years. One or two man custom development did of course occur - software's only real requirements is that you need to have a machine handy - but in most cases, such customization involved a large company subcontracting out to smaller ones.

Lately, however, this is changing, a change driven largely I suspect by the open source development model, but driven in a way somewhat different than the "New World Order" folks of yesteryear predicted. The virtual corporation as envisioned earlier (and expounded upon in such lengths by Wired Magazine) was predicated upon the notion that programmers, graphical designers, and other "creative" types would end up as glorious freelancers, independent entrepeneurs in control of their own destinies, the vanguard of the move towards the new frontiers of business.


2006-06-19 06:40:18
No argument with most of this except I agree with Ayn Rand in some parts, not the John Galt As Superman to be Helped and Related With - you have to read the autobios on Rand.

Where the BigCos excel is at the Big Projects. So far, most 'networks' have a tough time building the million-piece-part systems. Saying 'simple is better' is no defense to that. There are projects that are very large and that way by requirement. So while we may see things as programmers see them, it still takes a village of many types. If the network-as-virtual-company is to evolve, it must acquire many skill types and integrate them to do the big jobs, and that may be what happens: all bigCos become Temp Agencies and the high prices go to the best Temps.

Kurt Cagle
2006-06-19 08:28:13

I agree with Rand in some places - there are certain aspects to Libertarian philosophy that are very compelling ... there are others that are repellent. I've often been fascinated by the works of LE Modesitt, who has a very libertarian core, but all too often brings up the rather troubling question of who in fact is the more evil - the millions who are involved in petty evil or the one superman, in the name of good, who is capable of ending the lives of those millions.

I've actually seen very little evidence that bigCos are in fact any better at Big Projects. BigCos bring with them marketing savvy (making it look like they are making big progress) and the ability to bring to bear lots of resources largely simultaneously, but often at the cost of top-down administration and development that is fundamentally brittle when dealing with large systems. Indeed, what I see emerging instead is the concept of the virtual megacorp, in which you end up with one vcomp that acts largely in an oversite and management role, and other vcomps that attack specific parts of the problem. (Note that I'm not saying that such vcomps are little OSS factories - they could be formally incorporated entities doing completely commercial work - only that the approach that seems to work more effectively when dealing with a Big Project is to view it as a transient ecological system in which smallCos and vcomps "live" within their respective niches.

- Kurt

2006-06-19 14:19:52
Compare this

to this

Communication by insiders vs communication among outsiders (strongly coupled vs weakly coupled) and the goals and strategies for each.

IMO, it is still the BigCo with the deep pockets and the excess capacity that manages big projects best when the unknowns are many. The problem of the BigCo is that it can actually create more of those than really exist because of inefficient communications or a lack of shared goals among the many little fiefdoms that are inevitable. Because many virtual companies don't last long (aren't supposed to), they don't tend to develop deep control hierarchies that 'protect information' to 'protect turf'. Communications are better. They also don't play as many 'what-if' games and that has pluses and minuses.

A virtual company can manage a big project but when it gets down to fast and confidential issues, face to face, fact to fact meetings are productive as long as the players can play well together. Over the years, from the German Rocket Scientists to the XML tribe, what I've noted is that practice by a team with a goal over a lot of cycles is the best predictor of success. Colocation helps but isn't essential. Shared values are essential. Think of this as the mixture: a holding managing company with subcontractors. It doesn't matter if it is virtual or not: node competency is the issue. That is true of any networked organization.

Rand almost has it right. She unfortunately accentuated the individualism and the 'strong man' theories when in fact, it is the shared values and practice that make the difference. Where social networks go wrong is when they ossify and continue to control events too far past their prime. Too often I've seen extremely valuable opportunities lost to the 'we were the experts in our day' decisions that don't notice the sea or the environment has changed.

I've seen billions lost that way at least three times. Just call me Cassandra.


2006-06-20 03:23:46
Just to say, great post, and comments too!
Sylvain Hellegouarch
2006-06-21 05:38:32
Same as Kit. Great great post. I loved the Star Trek analogy :)

2007-07-21 12:49:13

2007-07-23 22:39:14

2007-08-09 11:27:04