Hacker News Reader

Comment by ams6110 | Hacker News Reader

It isn't popular because it's more expensive. Cubicles are cheaper than offices, and desks in a row are cheaper than cubicles.

It's a beancounter mentality, possibly understandable in a startup with limited funds for offices, but a false economy in the long run.

View on HN

I work in an open floor plan where I'm right on the aisle and hate it. Unless I'm focused 100%, it is human nature to check your back when someone is behind you. I can now identify the entire office floor's staff via their walk. It is a good way to keep abreast of issues and socialize, but waaaaay less productive than when we has cubes.

Anecdotally speaking, this streamlines the process for your butterfly fairy manager to float in and unnecessarily disrupt everyone's work at once with random out of context comments during their daily bypass appropriate communication channels.

This is often repeated on HN threads as a plausible explanation for open plans, always without evidence. But it makes very little sense as a theory. Companies do not have a financial inventive to consciously cripple their very highly paid, very difficult to hire developers' productivity. People are their number one expense and asset, vastly moreso than office space. Far from cutting corners, many open plan companies also spend lavishly on amenities to keep engineers happy.

A more plausible explanation is companies do open plan because they believe it will help foster collaboration, culture and team productivity. You may disagree but the "beancounter" theory makes no sense.

It's easy to measure the money saved when you go from individual offices to an open plan. It is a lot harder to measure the lost productivity of your programmers.

So, agreeing with that, are you aware of a specific terminology or theory of easy-to-measure vs. hard-to-measure costs and benefits?

It's a question which has been bugging me something fierce of late.

That's not really true. You can look at developer velocity and bug fix rates, to name a couple of readily available metrics. It would not be some big mystery if open plan was killing productivity.

Also, it does not explain why a "beancounter" mindset about office space would be paired with lavish expenses on free lunches and other de rigueur amenities in said offices, as is typical of open plan startups. Conclusion: it's not a beancounter mentality that drives the decision.

> "You can look at developer velocity and bug fix rates, to name a couple of readily available metrics."

How? If an office starts out open plan (which many of them do), are you going to move to new offices just to test the theory that it'll make your employees more productive? Plus, bug fix rates are a bad metric, bugs are not equal in complexity, and even if you try and guess at the complexity before working on a bug that's only an estimate, and one that doesn't take into account other important factors such as the level of churn in the codebase and the level of familiarity the developer has with the part of the codebase they are asked to fix.

Just because you have to do something that is suboptimal when you're starting out because of resource constraints, doesn't mean that you should enshrine that sub-optimal situation as ideal.

As an absurd example, Jobs and Woz made the first Apple computers in a garage. When they started making some actual money and could afford something better, they didn't insist on their engineers all working in garages.

Open offices are very reminiscent of the kind of environment that I remember from college, where the default studying option seemed to be going to the library, staking out a seat at a big table, building a fort of books, and "studying," with the emphasis on appearing studious and some sort of bizarre status hierarchy about how many hours you held station there.

The new Apple campus, designed by Jobs, will be open plan.

http://iphone.appleinsider.com/articles/16/10/06/apple-to-un...

What if there is an increase in bug fixes because the open office decreased the quality of work resulting in more bugs?

There are plenty of companies that have shifted formats, and a gross change in productivity would show up in those metrics.

Which metrics? There aren't any accurate ones I know of. Anecdotal evidence is possible, but I don't see how programmer productivity can be measured with scientific rigour.

>You can look at developer velocity

Developer velocity compares what you thought the complexity of a problem was vs how much time it actually takes. It conflates the level of optimism with the speed of development and glosses over how complex the stories should be.

Smart developers get overly pessimistic when pointing stories if they're measured by developer velocity rather than actual business value delivered because it makes them seem faster.

>and bug fix rates

Last time that metric was taken seriously developers deliberately inserted bugs.

Pretty much any development metric that drives some kind of reward will be gamed (maybe even unconsciously) by the developers. It's just basic psychology.

Those are issues with tying compensation and performance evaluation to metrics, not with detecting changes in them due to office environments.

>>Companies do not have a financial inventive to consciously cripple their very highly paid, very difficult to hire developers' productivity.

I highly doubt that:

a) most companies have done studies to measure correlations between office plan types and developer productivity

b) they even know how to measure developer productivity accurately

On the other hand, they can easily measure headcount per square foot. It's a simple division that even the stupidest executive can perform. The other stuff is just hand-waved away.

Also, the managers who sign off and make the final decision on which office plan to go with tend to spend most of their own work time meeting and talking to people, rather than needing to sit and concentrate on something for several hours. I expect this influences a cognitive bias when these managers consider the productivity level of an open plan office.

They don't consciously cripple, they just can't connect the dots. The facilities person tells the CFO, hey, I found you this great office for 1/3 the price. Then the VP of Eng, who never codes anyway, spends 100% of time in meetings replies, awesome - we'll have an open floor plan for better exchange of ideas. Win-win, right?

You're implying that management bases its decisions on scientifically founded research. I've never noticed that to be true just about anywhere. Most corporations don't implement best practices in, well, anything. Why should office layout be any different?

Or, with apologies to Andrew Lang, they use scientifically-founded research the way a drunk man uses a lamppost: for support rather than illumination.

Right. I mean, this is the industry that occasionally thinks it's a good idea to install rootkits as part of game or music drm. The idea that they're too smart to work contrary to their own long term interests is laughable.

Well all the research basically shows that productivity goes down when distractions increase, so you won't even see anyone quoting research to justify their decision. The executives who make this kind of decision generally know they will always have a nice office so it won't hurt them (and maybe the unsaid reason is that deep down they know that only them having an office helps the little people to understand their place in the org).

It doesn't help when the cool kids like Facebook endorse this kind of idiocy - other companies feel it is safer to just follow what others are doing. We need to have a new major tech player endorse the idea that the goal of a company building is to make the workers as productive as possible - you get better work from them, they enjoy it more and it will become easier to retain or recruit people.

Companies may not want to do that but the individual "tail" functions will optimise for there needs and ignore the wider needs of the company.

Look Boss I saved 10% on office costs - ignoring the fact that productivity has gone down by 20%

It's not about what companies financial incentives are, but what type of behavior they reward their staff for.

"Noone ever got fired for buying IBM"

The thing is, the higher rents for more floor space and separate offices are very tangible metrics for beancounters, whereas suboptimal developer productivity isn't. And that doesn't change when a company stops being a startup (by whatever definition).

My own anecdotal experience is that I have never, ever, seen private offices for devs. The private offices I know are all given as a privilege (for some boss), not for productivity.

I have been a software professional for almost 30 years. I have worked in shared offices, open plans, cubes, but have never in that time had a private office of my own. Often my immediate managers/leads did not have offices either. The execs always have offices, yet in many cases are rarely in them.

Reply on news.ycombinator.com