"War Rooms" Double Software Productivity

matt20 writes "Teams of workers that labored together for several months in specially designed "war rooms" were twice as productive as their counterparts working in traditional office arrangements, a study by University of Michigan researchers has found. Say goodbye to little cubes; it's war baby. I used to get tons done in a living room full of other people watching tv, doing homework, and programming, but the biggest problem is always choosing the music.

War room is great, but...

Now only if that crazy guy in the wheel chair would stop tring to salute Hitler all the time, we could really get some work done.

extreme programming

This reminds me of the claim of extreme programming [extremeprogramming.org] that working in pairs increases productivity. I think it's just because you feel more guilty screwing around when the other guy is working, so you both end up working. Kind of a prisoners' dilemma, I guess.

What do you mean flawed code?

If you wonder why it's called a "war room", wait till they start debugging each others code. I've always said that there's no fighter like an overprotective programmer.

luckman

i know what this is all about

anybody who knows anything knows that "war room" is simply a euphemism for "bong parlor". the reason people are more productive is because they're all too baked to talk to one another, focusing their energies on programming instead (which, as everybody who knows anything knows, is really easy to do stoned).

the old cubicle system didn't allow for huge hookah-parties, thereby forcing employees/programmers to smoke out of their own small pieces, which didn't really get them that baked, just enough that they couldn't concentrate on anything anymore.

as a side note, picking the music is never difficult in a bong parlor-- no matter what you pick, everybody will start bobbing along to the groove and saying, "dude this is pretty sweet. what is it?".

love,
grizzo

www.grizzo.com [grizzo.com]
it's 100% grizzo

Re:i know what this is all about

is grizzo.com hiring?

yeah, but you better have good headphones

i've worked in lots of rooms where the idea was "let's get everyone who's working on this project into 1 room so they can all work together easily" -it was nice when you had a question, you could just shout it out. but you're interrupted so often by other people's phones ringing or their conversations that i think i ended up less productive. if you're put into one of those big offices, you'd better be able to tune out background noise easily. then again, that's probably pretty much the same w/ a floor full of cubicles.

it's a nice way to create a feeling of working as a team, but i think that instant messaging & lunches together or something like that works just as well.

The reason it works

It works because workers surf/pr0n/slack less. If your boss could just move his eyes over and see you were reading slashdot when it was crunch-time, you'd be in big trouble, hence you work more instead of surfing. Not to mention the people that look at pr0n behind their closed office doors.
Ours is a generation that likes to surf and take lots of 'mini-breaks' when we are working by ourselves.
Having your boss sitting with you constantly changes the workhabits to create better productivity.
I'm not saying everyone does it, but I'm sure you have people at your office doing it, and 'war-rooms' would make them more productive...

--

CompUSA

When I once worked at a CompUSA I noticed a "war room" in the business sales divison. They had a blackboard with lines drawn on it and an actual army helmet with the words "$1 billion in sales by 2000".

They never did make that goal, or so it would seem. They appeared to be at war not with other computer sellers but the customers.

Re:extreme programming

Remind you of Highschool? One guy slacks off while the other works, and they share the credit... Also consider the stress factor of War Rooms, they may seem more productive, but in the long run this environment may cause more burn-outs and a high personel turn-over.

War room worked fine, until...

...Until the web designer decided his Super Soaker was more effective than our nerf weapons.

He chased one programmer into the server room. This resulted in an entire rack filled with fried boards.

So, it might be effective... as long as general stupidity is taken into consideration.

Peopleware and IRC as War room

In the book Peopleware [slashdot.org] DeMarco and Lister theorize that this works whereas everyone just packed in working on different shit doesn't because everyone's in the same mode at the same time. When you're designing you're all designing, and when you're rushing for a deadline and coding like mad everyone is.

I telecommute and we use IRC as our war room. It works great 'cause I can tune in and out w/o hassle.
--

"You can't fight in here, this is the War Room!"

I had one "war room" development experience. Not sure if was the company's idea or Anderson consulting.

No cubicles, no dividers, and no monitors that faced into walls or corners. Everything was public and open to inspection at all times. At first, the lack of privacy was maddening. Even if you had time to surf for porn, you wouldn't dare. The noise was a problem, but I found that you quickly adapted. Most people were pissed to fuming at the beginning but this passed.

The most amazing thing was the teaming that went on. You would think this sort of forced teaming wouldn't work, but it did. Programmers that normally played their hands close to the deck became show offs. Spontaneous groups would form for discussion or demos or to show off some nice coding tricks. By simply removing cubicles, a totally different dynamic was created!

I now work alone much of the time and I miss my "war room" days. Maybe more companies will follow if the productivity claims are proven. Maybe in the future, programmers will be placed in open glass enclosures to be shown off during company tours. As long as those touring are advised to keep their hands away from the programmers, there should be little injury. Most programmers might be surprised that they would actually thrive in a fishbowl of an environment. I know I was.

Hawthorne Effect ???

How much of this alleged increased productivity was simply due to the Hawthorne Effect?

Researchers many years ago at a GE plant in Hawthorne, England wanted to demonstrate the effect of improved lighting. So they increased lighting levels, and lo, productivity went up.

The problem came during the check-back when they lowered lighting levels to the original lux. Productivity went up even further!

It turns out the Heizenberg's uncertainty principle applies to people as well: If you measure and watch something, people react to the closer attention.

I worked in one of these...

I worked with two other people in a "mega cube" (with 6' high permenant "walls"). We dubbed it the "Playpen". The company firmly believed in giving people the resources they needed to do our jobs, so we had:

1) A very large whiteboard on one wall - with no furnature in front of it.

2) A spare computer and desk for "guests" to use during technical discussions (also used as a second terminal for the residents if they needed to run something that took a lot of resources)

3) It was a corner office in a tall office building, so it had an awesome view

4) Each person had their own phone

5) Nice workstations with 21" monitors

6) A comfortable "poof chair" (it is sort of a "full body" bean bag)

7) A shared bookshelf, so that you could borrow each other's books.

8) A collection of office toys, including a rubber-band powered plane (OSHA wouldn't have liked us flying that in the cube; too bad) and a bat suspended from the ceiling (it claimed to have a "soothing motion" - it didn't).

It worked VERY well since the three of us that shared the office all worked on the same projects at the same time. This environment was easily the most productive environment I've worked in.

People have mentioned "noise", though. It was true that music could be an issue. I recommend that companies buy GOOD headphones for every employee - a pair of $200 headphones can sound better than a $1000 set of speakers; once everyone has a set of these, you won't be able to pay them to listen to music on crappy computer speakers. The headphones should allow outside sound in and have at least 25' of cord (use an extension if you must).

As for ringing phones, that WAS annoying! It wasn't too bad, though, because we also had a "mini room" (actually two spare offices) across the hall. These rooms were used when people needed to have a long phone conversation, as they could go in and shut the door. This also gave some privacy. It was considered rude to talk for hours in the megacube, unless you were talking to everyone else there.

The furnature consisted of whatever we could dig up. I would recommend nice desks (single piece, not a U or L shaped desk) with LOTS of small tables. The ones that we had were 3' by 3' tables that could be configured however we wanted. If you wanted a "L" desk, you just grabbed three of these and put them on the left of your desk. I actually had a wrap-around desk build out of these. The nice thing is that you can reconfigure your space as appropriate for your work. We could, for instance, build a conference table in the middle of the room in a matter of minutes. All those nice "executive" desks really fall short in the ability to adjust to the work environment - they are nice for people who crave status symbols, but not for many others.

As you can see, though, this didn't save the company any money. The three of us had about twice the space we would have had if we lived in cubes. Not many companies could justify buying a poof chair for a space like this. Most environments I've worked in refuse to buy the most modern workstations for programmers, and 21" monitors are, sadly, rare. But, we were much more productive and I believe that our space and equipment cost less than additional employees would have.

I would also say that some of the positives of this environment came accidently. For instance, the company didn't think that being cheap on a bookshelf would increase productivity, but it did!

Re:XP favours a similar approach

What's interesting is that another rather sophisticated software development book, Software Project Survival Guide [construx.com] by McConnell says that one or two person offices are much better than more open, less private cube farms. He cites "After 15 Years," an essay by Tom DeMArco and Timothy Lister, that was published in the book Peopleware: Productive Projects and Teams. They claim that workers who work in the top 25% of environments are 2.5 times more prodcutive than those in the bottom 25%. Maybe the addition of being in extremely close contact is enough to overcome the distractions.

I'd like to see more research. Take the same team, put them in cubes, offices, and war rooms, and see how they do. It strikes me as entirely possible that the practices they talk about in the article as only being possible in "extreme collocation" are in fact applicable to any development team. Thus, the real factor is the implementation of software development best practices, and not the work environment. And there's plenty of studies that show good software process to be helpful, so it's not surprising that there was a big jump in productivity.

Well, I'm off to do some software process, by myself in my office. Gotta get those requirements written down...

Walt

slacking

Sorry, but good programmers who spend 1/10 of their day coding will outperform average programmers who spend 9/10 of their day coding every time. The way to increase productivity is to hire good programmers and give them the work environment to keep them there. The manager's job is to get the bullshit out of the way so the programmer can focus on what she does best. All the rest is touchy-feely nonsense.

Also: Whipping and video surveillance help too

Not to mention keystroke monitors, hidden microphones, and the random execution of anyone caught surfing inappropriate websites.
--

War Room Veteran Speaks out :)

Hi,

We just (end of October), came out of a 2 month War Room based project. Normally we live in lil gray cubes. We had a hell of a schedule - 2 months to build a meta-search engine for prices of Books, Music and Video, that used a commercial data source for book music and video data, and dynamic scrapers to get prices.

Three of us went into the conference room, and we got it done on schedule (Books, Music and Movies [goto.com])

Why ?

• It definitely created a hard-core attitude - the three of us were there 12-14 hours a day on average. Yes, it does do that "bad" guilt trip thing to you, but its really more kind of the fact that there is a cool hi-visibility project you are in on.
• Communication bandwidth was zero -- "What's the parameter to that API you defined ?" - instant answer.
• Fun, Fun, Fun. We were the most psyched team in the company - people used to come in just to feel the vibe (I am not a new ager at all). We decorated the room madly - cant post a picture of the door, but it looks like a totem pole. We had a TV - and watched every play-off game that was on... We drank every night, and ate at a local sports bar...

I don't think I could work in one of these 365 days a year - and I suspect that being THE SOLE War Room was kind of ego-boosting - if everyone was in one, who knows.

Also, you really have to be involved in a tight project, with the ability to tell anyone coming into the room to f*ck off if it disrupts you or is not relevant to the project. In a normal multi-person office, the day to day interrupts can drive multiple occupants mad...

Winton

p.s. There is also a similar article in the New Yorker this month (page 60, Dec 11 Issue).

IBM had it right

The IBM Santa Teresa report 25 years ago showed the right way to build offices for software developers: private offices with a door and window. They need to be near common areas for meetings. This was later supported by Peopleware.

The big problem with the Santa Teresa design is that it is an optimal solution. Since no brain power goes into finding better solutions, it all goes into finding excuses for not implementing it.

These war rooms were only compared to "traditional offices", ie those dreaded cubicles.

This article also used an oxymoron: "private cubicle".

One Rule for War Rooms

Do not all go out together for a TexMex lunch.

Re:slacking

I agree with the idea but not necessarily the ratio -- you're essentially saying that a good programmer is 9 times better than an average programmer? That makes for one hell of a curve, don't you think? That would also imply that the average programmer is 9 times better than a bad programmer. And you didn't even use the term "great" or "excellent".

I took a quick poll of my developers recently and asked "How busy do you feel you are on a scale of 1-10? Say a 5 or less is equal to I'm bored I need something to do, where a 9 or 10 is more like oh shoot even if i never leave my cube I still don't know if I'll meet my deadlines." The idea was that we were aiming for about an 8 for everybody. Some of my results:

• One of my best programmers, who I know has a handful of very critical projects on him,immediately said "Oh, a 3. What else have you got?"
• The more junior programmers had a tendency to say 5-6, whereas the senior programmers said 9. Is this because we loaded up the senior guys more? Or because the junior guys aren't as good at estimating their workload? I'm still not sure.
• TOO MANY said "For bursts at a time, a 9, and then for longer periods a 3." THIS is what causes horrendous productivity, because during those 3 periods nobody wants to take on additional projects for fear that the 9s are going to kick back in and leave them in chaos.

So back on topic I'd suggest that ALL programmers should be spending about 70-80% of their day coding. If you have good coders, then the solution is not to let them be productive for less time during the day, but to give them more to keep them challenged. (I used to have a job where I played games 90% of the time because there was no challenging work to do. I *hated* that!)