First and foremost I want to apologize for the headings on this article, they are horrible. They all look the same! It has to do with this template and the style sheets which I didn’t think I was going to have to fix before posting. I’ll have to work on that.
Secondly I’d like to mention that this post is not verbatim per Marissa Mayer, but more of my interpretation of her presentation. I didn’t change too much so don’t worry. I’m also sure that I missed some stuff because her presentation was moving a bit fast towards the end.
A Glimpse under the Hood of Google
Think about the Everyday
You solve big problems when you think about the everyday.
On an everyday bases there’s a set number of tasks and knowledge people wish to gain. iGoogle provides this, it allows everyone to personalize their search page and provide them with what they want without any extra clicking.
iGoogle also provides a great way for distribution and advertisement. Distribution - Any developer can create a gadget; Advertisement - this alone sells (advertises) you as a developer.
Lower the importance of something and make it possible, make it more likely. If you enjoy crossword puzzles, it’s not likely that you’ll set your home page to the New York Times crossword puzzle because there are other more important things the home page can be set to. With iGoogle, setting the cross word puzzle is not such a big deal, because it’s just one of the many things that can be placed on your home page.
Keeping it Simple
History on the Google Search Page
A lot of people like the simple Google Search page; an obvious call to action, a very clean look. Occam’s razor for logic says: The simplest design is probably right.
Sergey created the original homepage for Google Search. Marissa Mayer as been asked a thousand times over “what was your logic behind the Google Search page?”
When Marissa asked Sergey what his inspiration was, he looked confused and after a short pause answered: We didn’t have a webmaster. And, I don’t do HTML.
The Google Search page has gone through a number of subtle iterations; one of these iterations provided the search page with the copyright statement at the bottom of the page.
When the search page was first going through some one-on-one testing with end-users they found that many users were not used to the simplicity of the Google Search page.
The page would load and the end-users testing the page would stare. They would sit and stare, after about 40 minutes of staring Marissa would finally ask “is everything ok?”
The end user would response “I’m waiting for the rest of it.”
They found that the simplicity of the page made end-users feel as if there was more waiting to be loaded onto the page. The copyright symbol was added as a punctuation mark. Letting the end user know that the page is now done loading and you can commence with interacting.
Another frequently asked question about Google services was “how many people work there?” Marissa would answer about 80 employees. The next question was always “are you from the psych department?”
Advanced System made Simple
“beautiful code” >> Load Balancer (decide which logic server to use) >> Mixer (organize query for most valid results) >> Google Web Server (Relevant Advertisements + Search) >> Mixer (right relevance and order) >> Google Web Server (write html to page) >> Query travels back
700 - 1000 machines get hit with this query (takes about 0.16 seconds)
Our users don’t need to understand how complicated our systems are in order to use them.
The Science of User Experience (Interaction) Design
Design has become much more of a science than an art. You can mathematically figure out what design works best with end-users.
Split AB Testing (derive the best design)
When a change is made you know what you’re gaining, but do you know what you’re losing, what future movements (interaction) you’re changing?
Example: The amount of white space under the Google logo on a search page with results.
This test showed convincingly that less white space rather than more produces more usage and more revenue.
Example: Yellow or blue background for paid results.
Yellow produced both more satisfaction and more revenue.
Understand users better than they understand themselves.
“If you don’t listen to your customers, someone else will” - Sam Walton
“If I had asked people what they wanted they would have said faster horses.” - Henry Ford
The Number of Search Results Returned
Split AB testing was done with the “number of results” returned on a query. When given a choice between returning 10, 20, and 30 results; everyone said they would like to see 30 results returned. Soon after this change was made they noticed that the amount of searches decreased by 20% (this is bad).
They put on their thinking caps and tried to figure out why there would be less searches. The most obvious explanation was that, they’re receiving more results so there are fewer results to page through. After extensive research it was found that people were performing less searches because the page took longer to load; a variable that was not considered in their change. People really care about speed. They worked with this discovery and put their Google maps on a diet and noticed that they gained a 30% increase in people searching maps.
They also discovered that instant feedback allowed their users to learn faster. Instant feedback provides a steeper learning curve. Turn your users into experts as soon as possible.
Google Video vs. YouTube
It took Google 24-48 hrs to get your video uploaded. YouTube did it instantly. People preferred YouTube because of the experience.
The urgent can drown out the important. How can we take searchers and make them better. Could we do something that would teach them to become better searchers?
The learning curve on search is extremely fast. People start off by asking complete questions, they later learn to optimize their searches by using keywords. The learning curve is so short people become experts in minutes.
Add to the experience by making it didactic (instructive, art or science of teaching).
Think Long Term
Earlier we spoke about thinking short term. Short term allows you to see your daily routine and solve that problem or inconvenience you deal with all the time. Everyday tasks shouldn’t be hard, you shouldn’t even have to think about them — so if you are — maybe there’s something worth solving there.
Thinking long term is good because it allows you to consider how your world may change. You use a cell phone today, but a cell phone is only a means to an end … but what end; communication, entertainment, organization?
6 month thinking lead to a nicer cell phone, 2 year thinking lead to a more robust cell phone, 10 year thinking lead to the iPhone.
It’s obviously harder to predict the status quo the further you go out, but that’s what separates you from everyone else.
Google knows that in 10 years out they’ll have a bigger index, and there will be a dramatic context change. Google believes that search will be more personalized.
A simple search was first created, then an image search, then a blog search, then a news search, then a video search, then a map search. Now Google needs a search engine that searches their search engines. What’s always been wanted is the ability to search.
Creating what is Hard
Google has a secret, why are their products so amazing and why won’t many compete with them? It’s because they work on the impossible. They have fun gadgets and toys, but there is meaning to their madness. They build everyday applications that solve the harder problems of the world.
“A Healthy disrespect for the impossible” - Larry Page
????????The Impossible Search
Search is an impossible problem. But we work very hard for it every day. Dr means drive, doctor, and Dominican Republic … trying to disambiguate the search based off of 2 letters.
Getting to a 95% solution is a huge benefit.
GOOG-411
Google doesn’t know if they’ll ever make revenue from GOOG-411, they believe that’s helpful and a good idea overall that will help people in everyday tasks, but there’s more.
Behind this application lies advanced technology, voice recognition. This leads to bigger better things. Should your first iteration not work, there is always a new way to package it.
Breaking the Language Barrier
You search in Arabic, the query translates to English, finds the results, retranslates the results, then translates the pages as you click through them.
Be scrappy. Revel in constraints. The company is composed of many small teams, we’re constrained by time, people, etc …
Constraints help. Google translated their pages to 14 different countries. They went through a lot of trouble. Google sent this information to translators; it went through testing and “Q&A”. Google would update while information was getting translated, the deliverable translations were old and headaches followed soon thereafter.
Later on she spoke with a friend who owned a website; this small website was owned and worked on by 6 other employees. They managed to translate their site to 50 different languages. It turns out that fans enjoyed the site so much; they would offer to translate pages.
Google Health
Google Health is once again trying to work on a serious problem, giving people the ability to keep their own information. It took me (Marissa Mayer) 4 doctors to figure out if I was due for a vaccination because of a trip that I was taking.
It’s doubtful that allowing people to keep their own medical information will solve the problems of the medical history world, but we’re moving in that direction in a healthy disrespect for the impossible.
Imagination is a Muscle
Google Ride Finder
Maybe if we got enough of these we could define traffic patterns. Looking at these GPS signals, we wanted to know how closely we could get to the real thing. How live can the search be how we can best represent the information?
The purpose is not really to move forward with this but if we figure out the technology behind this we might be able to use it for a better purpose.
Brainstorm on Hypothetical Project
An example for working out the brain would be building a bridge between 2 buildings. How big would it be, what do we use, how high should it be. Something that currently has no practical use, but it gets the mind going
20 Percent Time
Let people work on what they want to work on, it’s where the best products come out from.
The overall approach is solving our own problems to the best of our ability then releasing that technology to everyone else. If it helped us it usually always helps everyone else.
Where Google gets most of it’s Input
The majority of user feedback used by Google comes from emails, statistics, and split AB testing.
Takeaways
Think about the everyday - Look at your everyday, if something in your everyday is requesting too much of your undivided attention, this might be something worth working on
Keeping it simple - Simple helps solve one definitive problem, solving one problem at a time and iterating on this solution seems to work
The Science of User Interaction Design - Know your users better than they know themselves, this sometimes means stepping away from the computer and putting yourself in their lives.
Think long term - Building for the next 6 months is great, but it would be best if you could think for a further future.
Create what is hard - Work on something meaningful, something hard even if it’s hidden in a not so meaningful disguise.
Imagination is a Muscle - Remember to spend time playing with what you’re passionate about. If you don’t know what you’re passionate about, you might want to start with that first.
June 1st, 2008 at 09:37pm
Eloy Zuniga Jr.
I’m sitting here in a bean bag chair at the Google conference. I have to admit the amount of attention paid to the atmosphere does deserve an applaud. I’m staring at a herd of nerds in front of me as they attempt to understand the fundamentals of fooseball. I think they may be over thinking the concept as they discuss the algorithms of the game.
They’ve also got pool tables. The majority of the people here range from their 30’s - 45’s. So as always it’s nice to feel like the youngen in the group. I’m appreciating this now while I still can. Soon I will be among the middle age and then what will I write about?
I keep comparing the atmosphere here to that of SXSW Austin. The environment (once again) awesome, but the majority of people here, not so great. Developers, programmers, people who do nerd stuff tend to form a type of god complex that makes them unapproachable. As if being social is a quality not worth investing in.
Now as I write these words you do have to keep in mind that I’m commenting on the majority. There are a recognizable number of people here doing their best to make friends.
SXSW Austin had very interactive people (for the nerds out there, pun intended). I think it was a little more lively/friendly because the people at SXSW Austin was made up of PR representatives and Marketing people. Those humans that are naturally friendly.
The sessions here can be compared to a day of Early Morning Sunday Church. You know you’re suppose to be there, and you know that the lesson is important, but it’s boring! Most of the lessons here can be self taught by picking up a book and spending a couple of minutes each day on me-time.
This does not mean that I haven’t learned anything. It’s only a comparison between conferences.
The next session is at 3p, I’ve decided to devote the rest of my efforts to the labs … that’s where you get hands-on experience with the Google developers. It’s awesome to learn conceptual ideas via a speaker, but actual programming lessons … not so much.
I’m going now. I’m either going to A run to the next session and try to get a good seat or B approach the people playing a game a foosball and watch them scatter as I make uncomfortable remarks such as hello.
May 28th, 2008 at 03:55pm
Eloy Zuniga Jr.
