2003-04-25 Google, Innovation, and the Web Craig Silverstein (Google's first employee after the two founders) http://conferences.oreillynet.com/cs/et2003/view/e_sess/4325 It's all about: -> Process, "how do we do what we do". * Google ingredients Mission statement: "Organize the world's information and make it universally accessible and useful" He mentions the phones-in-supermarket, product look-ups on the fly thing, as a personal goal/interest [Cory mentioned this; Microsoft guy mentioned this in the context of AURA too]. The mission statement stops Google getting distracted. "do things that matter" -- they believe that what they do really matters [it's true] [this is the ethos quadrant of the business process graph, with the axes long/short-timescale and formal/informal] "relentless focus on the user" [also ethos] ...google were willing to wait with adverts unless non-obtrusive targetted ads could be developed. labs.google.com: They have an idea, they want to make it a product, but they're not sure how to do it . it might be cumbersome to use . it might have too many features, and they don't know which to choose So they get feedback in the form of logs and comments One reason for user focus is that switching cost for search engines is very low (although all the short urls are taken, so they're protected there). "brilliant people have good ideas" good people understand the value of what they do, and use their time properly to make cool things. "a creative environment helps" the kind of place you'd like to spend a lot of hours. [you know, this is kind of slow. etherpeg is pretty busy at the moment] * PRODUCT LIFECYCLE 1 - ideas come from everywhere both good and bad ideas! (from employees, and the world at large) but the goal here is to collect ideas, that's all quick view of the touchgraph google browser -- this is the kind of idea google try to pick up on, too 2 - design for users they consider this very early in the process [and there's a sudden burst of words on sticks on etherpeg...] an example: - the first google beta "anorexic" (a compliment!) . "I'm feeling lucky" was already there . very lightweight. they still count the bytes on the homepage 3 - compile, discuss, prioritize the first step of the product proper. they keep a list of the "top 100" (actually it's 230). it's a prioritised list of the projects they'd most like to do. they use a system called "Sparrow" (from Xerox PARC), a freely editable webpage, to record ideas in their idea archive. there are product discussion forums every week or two about the new ideas. ideas are fleshed out here, understood better. evaluation next -- graph of Google's Wireless traffic, for web enabled phones, palm pilots etc. it boomed at christmas last year, and now it's growing really quickly. so wireless gets reprioritised, it's now further up the list. 4 - small teams are fast and agile teams of 3 people, that's how product development works. same people doing design, codes, testing, developing launch plan and maintenance. here's a photo of the team who crawl the web: there are 2 people. [!] 5 - communication is key 200-300 engineers, 100 small teams. this could be a problem. so lots of communication. products are beta'd and discussed over the entire company. code reviews... people can't commit code without someone else checking it and saying it's okay. 6 - tools to organize they have a search engine running over their internal website that makes it easy to find documents. guess which one... and web forms that have all the checkpoints for launch plans, etc. every engineer in the company writes a mail to a bot once a week; the bot collects it and puts it all on a webpage. weekly summaries. this guy, craig, he reads it all every week (takes 30-40 minutes). feedback helps, communication helps. weblogs are used internally to help communicate. they use blogger. 7 - test, experiment, iterate this is getting close to the launch point. user testing. First user study on the pre-beta, years back: "I'm waiting for the rest of it" and there are some funny ones. again, labs.google.com oh, and a jobs plug. he reckons that people who aren't the best hires are the reason other companies aren't as good as they can be. so they're *really* careful about only letting the best people in. [that's weirdly true. because people don't leave, they become part of the company. you've got to be really careful.] back to the iteration thing. showing iterations of google news.. gradual tiny changes in the design Top Queries: "it's useful to have a gestalt", a view of what's going on new slide "the web changes everything" is this true? maybe, maybe not. some things it has changed: . there wouldn't be Google before the web . oh, other stuff new slide "stay true to your mission" (original name of google was "backrub". [hey, i remember when "backlinks" were a really important concept. now we think of them *as* links.])