Comments on: Google designer leaves, blaming data-centrism

"I've grown tired of debating...minuscule design decisions," design lead Douglas Bowman vents upon his departure from the search giant.

[kgsbca]

I get so fed up with slow-loading web pages, I'm glad somebody cares about performance. I don't care about the width of a border, or the shade of blue, and I hope google doesn't replace Bowman with somebody who only cares about how things look. Maybe his next job can be for a real estate broker, they favor form over substance.

Bowman wanted subjectivity in google's subjectivity in their designs, but he doesn't realize that the problem with that is then it becomes a question of whose subjectivity. He decried the lack of debate, but how do you debate subjective taste?

[sanenazok]

Good for Google. Don't leave design choices to "artists" who don't realize that adding in a "nice" element that adds 0.00001% more processing power per transaction will noticeably cut down performance. Sounds like Mr. Bowman didn't know how to work with professionals outside of his own field.

[paulej]

Most engineers today do not care much about performance. If they did, they would not be writing so much Java code.

[someguy-]

Wow, someone seems jaded. What you don't take into account is that Doug is an experienced designer and knows what he's doing. He doesn't talk about adding "nice" elements at all in fact, from his blog it sounds like he wants to get his job done and avoid the minutia (that the engineers are adding- ha!). Being that he worked at Wired - I don't think he's a stranger to technology. Truth of the matter is, he doesn't tell engineers how to code - engineers shouldn't tell designer how to design. In fact, it sounds like the engineers there want to be what you call "artists", however any designer knows a good design isn't about just looking "nice". Talk to a designer - you'll open your eyes and come out of the cave you are living in.

[sanenazok]

@someguy: yep that's the problem with designers - everyone else is "living in a cave." He may not be telling coders how the code, but what do you think makes his design choices possible. Fairies? No, it's code. So if he tells them to do something - like add a border or use a particular color - that has to be translated into code. If that code creates inefficiencies, they are obligated to tell him. Obviously, he couldn't take this sort of feedback and so he took his playthings to a new sandbox.

[Sourdust]

@sanenazok, what code inefficiencies are created by selecting a shade of blue? You just pick one and move on. Inefficiency is wasting a developer's time coding tests for 41 shades of blue and then analyzing the results. Inefficiency is wasting someone's time debating the merit of a 3 pixel border v. 4 or 5 pixels.

[thelemurking]

Sorry, but you are not going to notice any difference in loading time based on the size of a border. If you do, then ditch your 28.8 modem and get with the modern age of the internet. It seems rather a waste of time and resources to determine which shade of blue is most mathematically correct. Is it really that hard to say oooh this blue looks better than that one and leave it at that?

[Art Dir]

Sounds like these code monkeys are testing load time differences that human beings can't perceive. Maybe they're trying to optimize total usage by the entire net. Even so, they are forgetting that ultimately, computers don't use the internet, people do. Unless they come up with a way to quantify, gather, store, and analyze subjective issues of taste, they are chasing their tales in trying to determine aesthetic choices from computations. They seem not to even recognize the existence of aesthetics at all. They don't seem to recognize that most animals in a particular species of hairless ape that has opposable thumbs and horizontally matriculated wrists develop emotional responses based on visual stimulation. They do so to one degree or another for just about every experience they have. That includes coders who have their noses buried to deep in numbers to recognize their own very human, very non-data driven behavior.

The most efficiently loaded web page is a page of all black text on a white background in a single column with no other visual elements and all navigation as linked text. Try looking at that for a while and see how long it takes before your eyes feel like they're going to bleed and your brain starts to wander off to decide what it wants to put in your stomach for lunch tomorrow, or to find anything else it can to avoid seizing up from boredom. Of course design will influence code and vice versa to some degree. It's best to let each side do what it knows best and leave the other alone or the result will be web pages that don't load at all because no one will care to waste their time looking at them in the first place.

[bammgabriana]

@thelemurking: the small differences may not matter to you and me, the end user. but put them together with biliions of people accessing your server and you will see the difference in server load and bandwidth.

[1812dave]

Just give me a fast, unclutter, easy to navigate web page. I could care less what subtle colors are chosen as long as they aren't obviously hideous. I see the validity of Bowman's complaint! They are bogged down in minutia. Perhaps that's way some many projects remain "beta" forever, at Google. MOVE ON GOOGLE! Finish a project!

[NPGMBR]

Honestly, how much faster does a page need to load with the processing power of computers increasing with every new generation and companies offering access to the net offeringn ever increasing speeds to compete?

[liquidfairy]

This boils down to usability 101, not performance of load times folks, but performance of one design element over another based on usability analytics such as the results you'd get from a&b testing. I've found myself in this same predicament, completely appalled by an almost cynical dissecting of design ideas. I completely understand the frustration with having to stop every two steps in your design process to spend countless hours quantifying any creative decision. Design is not a science and attempting to break it down as such will stifle creativity and the exploration of new ideas. I am a web designer who is in favor of and very much immersed in quantifying design and weighing out decisions based on SOME statistics but I strongly believe a line should be drawn and there should be a very fine balance between these in your creative process. Making creative decisions based solely on performance and having to justify every design element is a robotic way of being "artistic and creative"; it is unnatural and goes against our very own fundamental and very passionate core which questions the standards, pushes the limits, sets new trends and boldly "goes where no man has gone before". We are creatives, and when questioned where this creativity comes from the answer shouldn't have to be quantitative; my creativity stems from everyday life, derives from a bottomless well of years of experience in the field and from my own inborn creative abilities. If that answer doesn't suffice, then without a second to waste, it would be time to move on just as Douglas Bowman did. Congratulations to Douglas! He has my admiration and respect.

[miroslodki]

I think one of the elements we tend to forget when we do our a/b testing
(aside from the overall aesthetics) is understanding to what extent users really care and for how long

an immediate response might not have value for the longer term direction
what if Google tested gogle or goggle and found people clicked to the latter choices
would it change?

I think there needs to be wisdom applied to any data - lest we experience a derivative moment
cheers'
Miro
http://miroslodki.wordpress.com

[1812dave]

ack! make that "uncluttered"....

[1812dave]

Furthermore, how does Google know what any shade of blue will look like on every monitor?? There is plenty of variation in color from monitor to monitor. Like I said earlier, they are bogged down in minutiae that holds them up from making the BIG decisions.

[thomasbrenneke]

What big decisions? Rolling out a sea of datacenters, creating a fabric of utility computing and free services that completely destroys the competition who has been at it much longer?

There's a proper balance between aesthetics and engineering. Based on their product, the only component I see missing is adequate customer service.

[NPGMBR]

Hmmmmm....makes me wonder if this is the reason Google is still the BETA king.

[mark_925]

All the things that he complained about I thought sounded like good ideas. I hope they stick with the engineers.

[Art Dir]

This type of thinking is the same kind of genius that inspired hiring physicists because they're experts at pushing around data and developing complex formulas to design wonderful products like unsecured mortgage backed derivatives.

[rapier1]

Design matters. Just ask Apple.

[DBdweeb]

Likewise, design matters, just as Google. I and many others abandoned Yahoo for just one simple reason... All those blinky ads were insane. Simplicity and elegance are part of the design because it's ALSO part of the function. Design for subjective aesthetics at the expense of function is just bad design.

[rapier1]

There is a qualitative difference between good design and bad design. Good design is, in part an artistic and creative expression that gives value to the product. It can appeal in a fundamentally visceral way that isn't actually reducible to a set of metrics. However, if your base assumption is that you can measure the effect of design then you immediately reduce the design choices to that subset which is quantifiable. It necessarily excludes the exploration of a wide range of design possibilities.

[1812dave]

FUNCTION matters! Look at the new Shuffle. No controls on the body, so you are stuck with using headphones that have the controls on the cable. LAME! And I'm hardly the only one who sees the idiocy of that design.

[nicmart]

Google is both careful and daring, innovative and practical. I give it high marks.

[DBdweeb]

So many development efforts get bogged down in subjective stuff from "damagement" and end users it's tragic and a terrible waste of time. I almost lost a contract because some clinically insane dude thought my yellow border color was racist. I changed the color and kept the lucrative contract. While there needs to be a balance between form and substance, beauty and function, I find the over obsession with looks occurs way more often than nanosecond performance compulsions. If the application looks good data integrity is just assumed.

[1812dave]

Apple RELIES on design, since it doesn't have the function arena sewn up. Fanboys drop big bucks on Apple products, YET Apple's market share still sucks and apple is the pits when it comes to making reliable s/w updates. You get more bang for the buck and in this economy, design isn't enough. A fair price is more important that design, and will more important as the recession deepens. Besides, havent all the yuppie fanboys bought all the Mac computers they need already? Apple sales are way down this quarter. serves them right, IMO.

[zato_3]

It doesn't long for Microsoft to send it's creeps to anything on CNet that can turned against Apple.

[j0hnc]

Let's see... AAPL is up $16.24 or 19.03% for the quarter. Not "way down" IMO. MSFT is down for the quarter though.

[3zzy]

..and what creates "fanboys"? Apple indeed does a good job. I love my iPhone, my Macbook and my Cinema Display. They function as I desire and I'll buy their new products if I can, not to just GIVE AWAY my money, but to buy their products and increase my productivity.

[u1927]

What is a "code monkey", Stephen? Do you mean the really smart folks that made Google raging success? I need to get me some of those.

[jc364]

I am wondering how a company can put a measure on the effectiveness of 41 different shades of blue. It seems like such a minuscule change that it would be impossible to determine if the color made any difference at all. The effectiveness of design can be measured, but the unit of measure is way too big to make any conclusions like that.

That said, I do enjoy Google's choice in blue for Chrome's active window, which happens to be #589CEE. Much better than if they had gone with #589CEF.

[nthmost]

> I am wondering how a company can put a measure on the effectiveness of 41 different shades of blue. It seems like such a minuscule change that it would be impossible to determine if the color made any difference at all. The effectiveness of design can be measured, but the unit of measure is way too big to make any conclusions like that. <

Then I'm guessing you don't understand how evolution works either.

[1812dave]

u1927, so you think the shade of blue they choose is what makes Google successful? How about starting with servers and algorithms that index the web? Perhaps that has a bit to due with their rise to the top?

[llorban]

You can find a significant difference between anything if you have millions of users to test it on. It is really possible that there is no meaningful difference between 41 shades of blue, and you really can't tell from what sounds like a kitchen-sink approach to analysis and experimentation.

[rshah29]

Both sides have a good point.

My question: how many Google engineers commented on this thread. ;)

[zvonr]

this reminds me of:

Camel = Horse race designed by committee
or
http://www.despair.com/meetings.html

nothing unusual, part of the experience of working for somebody else... :-)...

[danaronm]

I have working in this kind of environment and it can be time consuming, petty, and void of common sense. Spending time, money and effort on details this minute are a misguided effort to turn a fault into a virtue and undermines the value of the individuals involved..
A good leader would recognize this shortcoming and effectively guide valuable manpower into a direction that tackles more difficult problems, such as accuracy of the search, speed of the search, & costs.
This pixel/color fixation reads like Google employees are a bunch of children who have just discover that they have a new toy.

[sanjayb]

I few years ago I may have been in the who cares what it looks like department. But seeing what products are successful like Apple for example shows that people do care about how things look. Look at Cnet's prize fights and reviews. They do put some criteria on the sexiness of a product cuz people do care. If you can design a good looking product that works great then more than likely that product will sell well.