Special forces teams

Innovation and speed to market. Ask any CEO these days and you’ll probably get these two responses to keys of success for the company. The high tech industry has always been a fast moving industry. But are there management organization structures that could speed it up even faster? Faster innovation, better customer focus, quicker to ship?
Most software companies divide themselves up into business units. For example, Microsoft has a Windows, Office and MSN business units. From there a fairly typical matrix organization is established along functional management lines (VP Engineering, VP Marketing > Manager Engineering, Manager Marketing > Software Engineer, Product Manager). VPs organize their employees around projects and disband them when they’re done.
If there is anything that slows down these organization types is management layers. The more managers you have, the more meetings you need, the longer it takes to get things done. Presumably, if you could reduce the need of middle management, perhaps speed to market and innovation could be positively impacted?
An article in Time magazine about special forces (Secret Armies of the Night) got me thinking about whether special forces could offer an analog for corporate organization. Rumsfeld has been the real driver for special forces getting a more prominent place in today’s military. His idea is that a few special forces with the right gear and intelligence should be able to take the place of a 3,000 man brigade. They can move far faster than a typical army division–critical for fighting quick moving world conditions.
So, what does a typical special forces unit look like? They appear to be made up of a squad of 6-20 soldiers and are trained for highly specialized engagements (terrorism, psychological operations, light infantry strikes). Any one who read or watched Black Hawk Down saw both the Army Rangers and Delta Force at work. Teams working within their functional expertise area have a certain amount of autonomy to do whatever it takes to complete their mission-though missions seem to be fairly targeted and short term in duration. For example, during the recent Iraq War, the Naval Special Warfare Task Group (five separate teams of 20 Navy SEAL commandos) were sent out to secure Iraqi oil-refining and shipping facilities. That mission was accomplished. πŸ™‚
I do see a possible analog for the software industry. Let’s take Microsoft as an example again. During the browser wars, the company made a full assault on Netscape. The Windows business unit assembled a team to build a competing product (Internet Explorer) and set on a schedule of software releases, business partnerships (namely with AOL) and strategic moves (integration into Windows) to destroy their competitor. The goal was fast releases which a fairly small tactical team was able to accomplish. Just as the Army has special forces units, I think standard matrix based organization can have their own strike teams.
But the major difference between the military setup and a corporate setup is the ongoing nature of the teams. The military special forces teams make sense to be ongoing. We’re likely to always have a need for small tactically focused teams special trained in something (counter terrorism). But in the corporate world, a team trained in a very specialized area doesn’t make as much sense (at least in the software industry). Back to our Microsoft example, would it make sense to have that same Internet Explorer team stay together forever? Those “A” players were eventually disbanded and placed on other important projects. The one thing that would seem to be consistent to special forces, is teams are self directed. But they are only self directed in the sense that they decide how to achieve their objectives. Teams have a mission that military command or company senior management in this example assigns to them. They don’t come up with that on their own.
I think the notion of special forces teams have a place in the software industry. They’ve been around informally (like when the original Apple Macintosh team was formed and flew a pirate flag over their building on Apple’s campus). But it’s my theory that strike teams only work under the following conditions: 1) they obtain a mission for executive management, they don’t create their own; 2) they have a fixed time frame of existence, they are not ongoing; 3) they are self directed only in how they achieve their objective, they don’t act as an independent mini business unit; 4) they are a temporary group of seasoned personnel, staffing teams with new or junior members slows them down.


5 Comments on “Special forces teams”

  1. Amazon.com’s initial Auction site was a swat team operation like the one described. It was secret inside the company, collocated all operatives onto a single floor, and plucked skilled people out of their orgs (managers couldn’t say no). The result was a great site builtin 60 days — but initially a total failure in the market (though said structure now is nearly 20% of amazon’s sales).
    It was grueling work, and I wish as much effort had gone into the business strategy as the functionality of the site. Your comparison to Black Hawk Down is appropriate — as it doesn’t matter much how skilled the Delta Force is if you rope down into a city filled with people who know where you are and are ready to attack you. Recon and signals intelligence are also important. And it matters how you declare success. Was the operation in Mogadishu a success? With out a doubt — but the Clinton administration couldn’t live with the cost of casualties and turned tail. The number of subsequent casualties from turning is without a doubt in the tens of thousands, as it sent the message that America wouldn’t accept casualties in a war — leading all sorts of people to make bad judgments about engaging America. The stakes aren’t nearly the same — but thankfully Amazon pushed marketplace ideas around until they found their nitch. Had they done the same with PlanetAll and Junglee — and not declare failure and retreated — Amazon might be farther ahead on those fronts than they are today.

  2. rm -r * says:

    Both are excellent points, and I agree with Andrej’s conclusion about defining a mission and directive. The one point that still remains sticky is that millitary operations and shrink wrapped software are generally a one-time operation – with something like a gigantic commerce website you have an integrated system that needs to be maintained and running underneath (You also have this with boxed software because you create patches and updates).
    How is that reconciled with the special forces? As someone who has done support both for Amazon (US and Intl retail development) and the Army (172nd Support Battalion) I know it’s incredibly hard. Once the special forces are in, someone needs to pick them up or supply the food and ammo for the occupation – or provide the technical support and integration into the existing code line.
    This takes intense planning and also a willingness to honestly look at all options and change course when needed (I wouldn’t say that neither Amazon or the military do this very well). However, from my perspective the bottom line is: it takes a lot of technical planning and behind the scenes support to create a viable base of operations for these special forces to be deployed from.
    How do you sucessfully build that in the software industry without bloat? Are most support services a commodity that can be outsourced? What is the right balance? My experience at Amazon has led me to believe that this support layer is pretty much non-existent. If that happened in a military setting, we’d all be dead.

  3. Andrej Gregov says:

    Great comments. Speed to market has been on my mind quite a bit lately as I’ve been leading a team that I haven’t been able to get launching new features anywhere close to their potential and capability (through no fault of their own). Josh mentions building an entire new store and platform at Amazon in 60 days. Today, we hardly get a requirements document written in that time frame.
    Looking at the Microsoft example again, the IE team had an infrastructure already built before they started work-namely windows. For a commerce site like Amazon, the equivalent would be the platform. These key infrastructure examples I assume work in the Special Forces discussion. But I do realize this is a big assumption. πŸ™‚
    It’s well known in the military that being the sponsor for a fancy new weapons system (B2, F22, etc) wins promotion. Sponsoring a troop supply ship-boring. But they’re just as critical is the front line weapons. Problem is, infrastructure bores sr. management just as much as military brass. Tough issue. I’ll keep my eye out for infrastructure done right and post if I find anything.

  4. rm -r * says:

    This article in the Atlantic Monthly, “Peace is Hell” by William Langewiesche
    (http://www.theatlantic.com/issues/2001/10/langewiesche.htm) is a great look into what it takes to support tactical forces in the military.
    These things need to be considered, and like you said, they are boring issues. Sr. Management needs to have a balance of vision
    and realism in order to succeed. The latter is sorely lacking.
    And this type of support is expensive. Most software industry studies state that integration accounts for roughly 30% of all IT costs. That is HUGE, and will probably be the norm for many years to come. Because just as soon as you have developed a ‘platform’, it is already starting to become outdated and you need to be thinking about the next thing and repeat the cycle.
    The upside is that there will always be work πŸ˜‰

  5. jana says:

    Andrej, glad you are making good use of your vacation time. I wanted to add a comment to your Special Forces post. Earlier this week as I was flipping to PBS Kids I caught a little news segment that said how the government is having a tough time keeping the special forces enlisted. The turnover rate is very high and it is not good for the military to continue to lose these highly trained and experienced military troops. Anyway, something to think about before turning deploying a Rumsfeld strategy at A to the Z.