Liveblogging a presentation at An Event Apart San Francisco by Doug Bowman
(His entry slide reads d3zign2 sca13. I’ll look later for the inverted z he’s using. [From Azeri? No, I was thinking of ƨ from Zhuang. Inverted z isn’t in Unicode.])
If it weren’t for Jeffrey and Eric, I wouldn’t be in the exact position I am today. Thank you for your support and encouragement several years ago for the Wired redesign.
Rather than give you a formal presentation, I’d rather just assume the 200 of you are my close friends. (Runs through his “career so far.”) Started at a small design agency, Mentus, in San Diego. 20 people. I moved to Hotwired in San Francisco in 1996, I think. Around 150 people, publishing around eight or nine different channels – a site about cocktails, politics, pop culture, search engine Hotbot. Transformed back into the Wired brand a year after I was there. Renamed Wired Digital, acquired by Lycos. Acquired by a Spanish company, Terra. I was the only one on the team who spoke Spanish, so I travelled around the world speaking Spanish, telling jokes about the team in Boston. I decided to kind of give that up. I started a little company of one, Stopdesign, but I got to work with some great people along the line – Happy Cog, Adaptive Path, Eric – through subcontracting. About two years ago, I got contacted by a little company for a contract opportunity, which later became an offer for full-time employment.
I never thought I’d work inside a large company again, but the idea of working for Google really appealed to me – to have a grand impact on a global scale, design problems the likes of which I’d never seen before. I’d have taken it had the salary only been a dollar a year. I had to think about it for a while, but this opportunity to lead the visual design for Google was phenomenal and I’d be stupid t pass it up. Been there a year and a half.
I can’t give you all the inside story of Google, but IC can give you some of the design decisions we have made there and the ones I’ve pushed into the company and tried to change Google a bit. Compared to companies like Apple, which has their design 75%, 80% figured out, Google doesn’t.
Scale: It’s not until we start to step back from that data and we realize it’s different from what we thought it was. You see a larger picture, a more divers set of statistics. You make different decisions based on the scale at which you view the data. Ben Shneiderman: “An interface is worth a thousand pictures.” (At a picture is worth a thousand words,) that’s a million images per interface.
Anticipate growth. (Shows world population growth, Web growth, including data points for 1960 through 1990.) Products are gonna grow, not just the popular products. Languages, everything you can think of, is going to go into building this phenomenal Web. We though in the ’90s we were just on the verge of that. Now in the 2000s, we’re still on the very tip of this growth of the Web.
Companies that have dealt with scale well: McDonald’s. Their actual line of business? Place. It’s all about location, location, location. McDonald’s is sin the real-estate business. Starbucks. Ikea (shows visitors to stores rather than number of stores). You know how efficiently packed they are. They take advantage of every single inch in that box. That lets them scale up more quickly than if they were wasting space with packaging and whatnot. ANd bringing fairly nicely designed products to the masses, the mainstream.
Gaming consoles. (Shows Xbox, PS3 plateaux, Wii growing straight up.) This is the kid of growth and scale I’m interested in. Designed for the masses. They tapped into something that the other two didn’t think of – not just the high-end graphics cards. You just start going through motions and you can start playing. That has broad appeal and lots of social impacts as well.
The invitation model for a lot of products, particularly with Gmail: People thought it was to create a cool, desired factor. No, it was mainly to control the growth. Any product they release gets millions of hits in the fist couple of days. They set out an invitation model to more accurately predict growth. Gmail has stabilized its servers enough to go on an open-sign-up model.
Golden Gate Bridge: Rubber cones change the direction of lanes, with two empty lanes late at night for safety. Times Square: Everything is screaming for my attention; advertising sliding down a slippery slope, with one ad going up and the next one having to scream for attention. This has parallels for the Web. If Google were known for this nice clean Spartan interface, what would happen if it were Times Squared? (Shows horribly cluttered CNN-style Google search-results page.)
We’re starting to talk at Google about the one-billion-object user: All your E-mails, chats, MP3s. All the things that are involved in the cost of switching to another computer. This kind of scale is going to be hitting us very quickly.
(BTW, I don’t understand these designers onstage wearing brown shoes with blue jeans or nonbrown trousers.)
How would you design an elevator control panel for 500 storeys?
Google Reader deals with this an interesting way. They’ll pull in the first 20 items to render the interface really fast. As you start to scroll down, it’ll start grabbing more results. It’s known as the infinite scroll: You can keep scrolling without noticing the effects of delays. We’ll be seeing this more in certain applications and certain sites that want to render data but don’t want to wait for all of that to come in.
Factor in flexibility. (Shows gas-station sign around Brisbane with limited space for dollars and cents.) Knowing that you’re growing in scale. Even if you can’t imagine adding a zero onto your numbers, start planning for that now. Start planning for a row of zeroes. Y2K was a famous example of this; designers never accounted for the fact that we’d flip past the year 1999.
(Shows Amazon tabs.)
(Shows date format in a little cartouche on a blog. Bumps up text size.) Eventually text pokes out beyond it. Factor in the idea that those things are going to grow in scale. You don’t ever have a lockdown on text size. (Shows how Simplebits gets it right.) If you can’t make the text scale, at least make the UI and the chrome scale.
Make it fast. I didn’t know how much attention was paid to speed till I got there (Google). Even a tenth of a kilobyte or 5KB, when multiplied by millions of users, can result in 1.9GB to 95.4GB of bandwidth.
(Had done a little of the HTML and CSS on Google Calendar.) But we had large issues with latency. (Each event has its own separate code and colour values in CSS.) Google pre-caches weeks on either side, let’s say 12 weeks. With 0.005s latency, times a million users it’s 467h of latency.
Reliability in terms of nines: Only 90% reliability means you’re down 36d a year. Six nines would be near perfection in reliability (99.9999% means 31s downtime).
Internationalization: Google in 57 languages. But also localization – 117 languages for 157 domains. [Um… Canada in Canadian English?] (Shows Tongan, Klingon, Elmer Fudd, and other translations, rather trivially.) Not only does this affect search, ti affects all our products. We know that every product will have to be translated very soon after release; it isn’t fair to users to make them wait to get their own version. We have to plan for tabs to flux in size and shape. (Shows Chinese with single-character tabs.) This is why Google will rarely use text in images.
(Shows how Blogger signup page had to have image button converted to text. Shows six language versions, including Vietnamese with character-encoding errors.) I would have liked the arrow to have expanded or contracted with the length of the word.
Our users don’t always want something simple. They want powerful products; they want them to appear simple. (Shows success of registration process in pre-Google Blogger step by step. Last step asked for FTP server information and had 3% success. After Google acquisition, huge dropoffs even in the first stage.) The majority of people would come to the Blogger front page and just leave. That’s what the redesign in 2004 was all about. Redesign increased successful sign-ups threefold.
(Long discussion about gestural interfaces, mentioning Wii, Minority Report, Media Lab experiments to stretch and squeeze Google Maps.)
Spot the invisible opportunity: How do you know when an idea is successful? It’s very rare that people can kind of spot those opportunities. (Transcontinental railroad vastly increased urbanization and industrialization on the West Coast.) Las Vegas produces nothing you can actually buy; it’s all in the gaming. This idea of producing nothing and yielding money is interesting to me.
What have I been doing for the last year and a half? Everybody’s been asking me that. There are a lot more employees in this company than I expected to deal with. Everybody has an opinion on visual design. It is the hardest thing to push through improvements and get any kind of changes even on the smallest things in the Google interface. How do we objectively evaluate such a subjective topic? We can’t just ask what people like; there’s so many diverse opinions, types of products. Their products are so lacking in consistency, and there are so many improvements we could make in these products.
Focussing it just on the brand itself. (Shows giant list of “brand attributes”: organized, accessible [ha!], quirky, ubiquitous, personal, daring, easy to use, ubiquitous again, smart. Then they added future attributes and ones to avoid: sea of blue links, flashy, gratuitous, dominant, busy, patronizing, overwhelming, underwhelming, decorative, wealthy, pimped, overproduced, heavy, slow, flourish [sic], flashy. Grouped those into logical clusters: fast/modest/instant, fun/friendly/playful, organized/trusted/accurate, innovative/daring/edgy, the last being where design is the farthest away. Set up moodboards showing external designs.)
How to combine these four clusters? Trying to use desirability methods, testing with users, to determine if we’ve designed appropriately to these adjectives. Anticipate future growth.
Tibor Kalman: “Just look at the Web designers: They can single-handedly reach the world.” It’s important to keep in mind that the Web has the power to reach everyone in the world. We can start thinking about scale now.