Thursday, 25 July 2013
Under the hood of Croatian, Filipino, Ukrainian, and Vietnamese in Google Voice Search
Although we’ve been working on speech recognition for several years, every new language requires our engineers and scientists to tackle unique challenges. Our most recent additions - Croatian, Filipino, Ukrainian, and Vietnamese - required creative solutions to reflect how each language is used across devices and in everyday conversations.
For example, since Vietnamese is a tonal language, we had to explore how to take tones into consideration. One simple technique is to model the tone and vowel combinations (tonemes) directly in our lexicons. This, however, has the side effect of a larger phonetic inventory. As a result we had to come up with special algorithms to handle the increased complexity. Additionally, Vietnamese is a heavily diacritized language, with tone markers on a majority of syllables. Since Google Search is very good at returning valid results even when diacritics are omitted, our Vietnamese users frequently omit the diacritics when typing their queries. This creates difficulties for the speech recognizer, which selects its vocabulary from typed queries. For this purpose, we created a special diacritic restoration algorithm which enables us to present properly formatted text to our users in the majority of cases.
Filipino also presented interesting challenges. Much like in other multilingual societies such as Hong Kong, India, South Africa, etc., Filipinos often mix several languages in their daily life. This is called code switching. Code switching complicates the design of pronunciation, language, and acoustic models. Speech scientists are effectively faced with a dilemma: should we build one system per language, or should we combine all languages into one?
In such situations we prefer to model the reality of daily language use in our speech recognizer design. If users mix several languages, our recognizers should do their best in modeling this behavior. Hence our Filipino voice search system, while mainly focused on the Filipino language, also allows users to mix in English terms.
The algorithms we’re using to model how speech sounds are spoken in each language make use of our distributed large-scale neural network learning infrastructure (yes, the same one that spontaneously discovered cats on YouTube!). By partitioning the gigantic parameter set of the model, and by evaluating each partition on a separate computation server, we’re able to achieve unprecedented levels of parallelism in training acoustic models.
The more people use Google speech recognition products, the more accurate the technology becomes. These new neural network technologies will help us bring you lots of improvements and many more languages in the future.
Wednesday, 17 July 2013
11 Billion Clues in 800 Million Documents: A Web Research Corpus Annotated with Freebase Concepts
“I assume that by knowing the truth you mean knowing things as they really are.”
- Plato
When you type in a search query -- perhaps Plato -- are you interested in the string of letters you typed? Or the concept or entity represented by that string? But knowing that the string represents something real and meaningful only gets you so far in computational linguistics or information retrieval -- you have to know what the string actually refers to. The Knowledge Graph and Freebase are databases of things, not strings, and references to them let you operate in the realm of concepts and entities rather than strings and n-grams.
We’ve previously released data to help with disambiguation and recently awarded $1.2M in research grants to work on related problems. Today we’re taking another step: releasing data consisting of nearly 800 million documents automatically annotated with over 11 billion references to Freebase entities.
These Freebase Annotations of the ClueWeb Corpora (FACC) consist of ClueWeb09 FACC and ClueWeb12 FACC. 11 billion phrases that refer to concepts and entities in Freebase were automatically labeled with their unique identifiers (Freebase MID’s). For example:
Since the annotation process was automatic, it likely made mistakes. We optimized for precision over recall, so the algorithm skipped a phrase if it wasn’t confident enough of the correct MID. If you prefer higher precision, we include confidence levels, so you can filter out lower confidence annotations that we did include.
Based on review of a sample of documents, we believe the precision is about 80-85%, and recall, which is inherently difficult to measure in situations like this, is in the range of 70-85%. Not every ClueWeb document is included in this corpus; documents in which we found no entities were excluded from the set. A document might be excluded because there were no entities to be found, because the entities in question weren’t in Freebase, or because none of the entities were resolved at a confidence level above the threshold.
The ClueWeb data is used in multiple TREC tracks. You may also be interested in our annotations of several TREC query sets, including those from the Million Query Track and Web Track.
If you would prefer a human-annotated set, you might want to look at the Wikilinks Corpus we released last year. Entities there were disambiguated by links to Wikipedia, inserted by the authors of the page, which is effectively a form of human annotation.
You can find more detail and download the data on the pages for the two sets: ClueWeb09 FACC and ClueWeb12 FACC. You can also subscribe to our data release mailing list to learn about releases as they happen.
Special thanks to Jamie Callan and Juan Caicedo Carvajal for their help throughout the annotation project.
Tuesday, 16 July 2013
New research from Google shows that 88% of the traffic generated by mobile search ads is not replaced by traffic originating from mobile organic search
Often times people are presented with two choices after making a search on their devices - they could either click on the organic results for their query, or on the ads that appear on the page. Website owners who want to build a strong online presence often wonder how to balance organic search and paid search ads in driving website traffic. But what happens when ads are paused? Would businesses see an increase in organic traffic that could make up for the loss in paid traffic? To answer these questions, we released a “Search Ads Pause” analysis in 2011 showing that 89% of traffic generated by search ads is not replaced by organic clicks.
As smartphones become increasingly important to consumers, we recently conducted the same studies for mobile devices to understand the role of mobile search ads in driving site traffic. From March 2012 - April 2013, we ran 327 unique studies across US-based mobile advertising accounts from 12 key industries.
We selected AdWords accounts that exhibited sharp changes in advertisers’ spending on mobile search (ad spend) and identified stable periods before the spend change (pre-period) and after the spend change (post-period). We observed the number of organic and paid clicks, and the number of times organic results appear on the first page of search results (impressions) during both the pre-period and post-period. Google then created a proprietary statistical model to predict what the number of organic and paid clicks would have been in the post-period had the ad spend not changed, and compared those figures to the actual number of clicks observed. We then were able to estimate what percentage of paid clicks are incremental, i.e. a visit to the advertiser’s site from an ad click would not have been replaced by a visit to the site from an organic click.
The final results showed that mobile search ads contribute to a very high proportion of incremental traffic to websites. On average, 88% of mobile paid clicks are lost and not recovered when a mobile search campaign is paused. This finding is consistently high across the 12 key industries, including automotive, travel, retail and more. The full study, including details around the methodology and findings, can be found in the paper ‘Incremental Clicks Impact of Mobile Search Advertising’.
Tuesday, 9 July 2013
Google Databoard: A new way to explore industry research
It’s important for people to stay up to date about the most recent research and insights related to their work or personal lives. But it can be difficult to keep up with all the new studies and updated data that’s out there. To make life a bit easier, we’re introducing a new take on how research can be presented. The Databoard for Research Insights enables people to explore and interact with some of Google’s recent research in a unique and immersive way. The Databoard uses responsive design to to offer an engaging experience across devices. Additionally, the tool is a new venture into data visualization and shareability with bite-sized charts and stats that can be shared with your friends or coworkers. The Databoard is currently home to several of Google’s market research studies for businesses, but we believe that this way of conveying data can work across all forms of research.
Here are some of the things that make the Databoard different from other ways research is released today:
Easy to use
All of the information in the Databoard is presented in a bite-sized way so that you can quickly find relevant information. You can explore an entire study or jump straight to the topics or data points you care about. The Databoard is also optimized for all devices so you can explore the research on your computer, tablet or smartphone.
Meant to be shared
Most people, when they find a compelling piece of data, want to share it! Whether it’s with a colleague, client, or a community on a blog or social network, compelling insights and data are meant to be shared. With the databoard, you can easily share individual charts and insights or collections of data with anyone through email or social networks, just look for the share button at the top of each chart or insight.
Create a cohesive story
Most research studies set out to answer a specific question, like how people use their smartphones in stores, or how a specific type of consumer shops. This means that businesses need to look across multiple pieces of research to craft a comprehensive business or marketing strategy. With this in mind, the Databoard lets you curate a customized infographic out of the charts or data points you find important across multiple Google research studies. Creating an infographic is quick and easy, and you can share the finished product with your friends or colleagues.
The databoard is currently home to six research studies including The New Multi-screen World, Mobile In-store shopper research and Mobile search moments. New studies will be added frequently. To get started creating your own infographic, visit the Databoard now.
Sunday, 7 July 2013
Google Logic: Why Google Does the Things it Does
“What does Google want?”
A favorite pastime among people who watch the tech industry is trying to figure out why Google does things. The Verge was downright plaintive about it the other day (link), and I get the question frequently from financial analysts and reporters. But the topic also comes up regularly in conversations with my Silicon Valley friends.
It’s a puzzle because Google doesn’t seem to respond to the rules and logic used by the rest of the business world. It passes up what look like obvious opportunities, invests heavily in things that look like black holes, and proudly announces product cancellations that the rest of us would view as an embarrassment. Google’s behavior drives customers and partners nuts, but is especially troubling to financial analysts who have to tell people whether or not to buy Google’s stock. Every time Google has a less than stellar quarter, the issue surges up again.
As I wrote recently when discussing Dell (link), it’s a mistake to assume there’s a logical reason for everything a company does. Sometimes managers act out of fear or ignorance or just plain stupidity, and trying to retrofit logic onto their actions is as pointless as a primitive shaman using goat entrails to explain a volcano.
But in Google’s case, I think its actions do make sense – even the deeply weird stuff like the purchase of Motorola. The issue, I believe, is that Google follows a different set of rules than most other companies. Apple uses “Think Different” as its slogan, but in many ways Google is the company that truly thinks differently. It’s not just marching to a different drummer; sometimes I think it hears an entirely different orchestra.
Google’s orchestra is unique because of three factors: corporate culture, governance, and personal politics. Let’s start with the culture.
Google culture: You are what you do
The strategic thinking of most companies is shaped by the way they do business. For example, a farmer thinks in terms of annual seasons and crops; everything revolves around that yearly cycle. Manufacturing companies, the traditional foundation of a 20th century economy, plan in terms of big projects that take a long time to implement and require a lot of preparation. If you’re building a car or a plane or even a smartphone, you have to plan its features well in advance, drive hardware and software to completion at the same time, and arrange manufacturing and distribution long before you actually build anything. The companies that build complex physical things naturally plan their products in terms of lifecycles lasting at least 12 to 24 months, and sometimes much longer.
That long planning cycle dominated big companies in the 20th century, and was driven into all our heads through generations of business books and business school classes. It’s how most of our brains were formatted.
An internet company, like Google, works at a fundamentally different pace. Web software changes continuously. You don’t plan it rigidly; you evolve it day by day in response to the behavior of customers. The faster and more flexibly you evolve, the more successful your products will be.
This evolutionary approach, and the Agile design processes that support it, is built into the fiber and psyche of web companies. They don’t think in terms of long-term detailed plans; they think in terms of stimulus and response.
This is a dramatic change in the history of business. In the past, the nimble companies were always the little ones. The larger your company, the more it valued planning and the long-term view. Google is one of the first very large tech companies ever to pride itself on rapid response rather than rigid planning.
On top of this quick-turn bias there’s the cultural training of Google’s senior management. Most big companies end up being run by professional managers who came up through business school or finance, where they get trained in the rhythms and personality of traditional big business. They learn a shared vocabulary and set of values that are very familiar and comfortable to investors. By contrast, Google is completely controlled by engineering PhDs. They speak the language of science rather than business, and they’re contemptuous of the vague directional platitudes and reassuring noises made by modern finance and marketing.
I think most reporters and analysts don’t understand how fundamentally different the engineering mindset is from traditional business thinking. It’s a very distinct paradigm, unfamiliar to most people who haven’t studied science (link).
One key element of the engineering mindset is the use of scientific method: you encourage a Darwinian marketplace of ideas, you test those ideas through controlled experiments, and you make decisions based on experimental data.
In its behavior and vocabulary, Google oozes scientific method. A couple of times recently I’ve heard Google executives say in public, “if you can’t measure it, you can’t improve it” (link). It's an old quote, dating back at least to Lord Kelvin in the 1800s. It's also a subtle twist on the traditional mantra used in web design: “that which you measure, you can improve.” The web design version says you should measure everything you can; the Google executive version implies that nothing really matters unless you can measure it.
That’s a very scientific, rational point of view, but I couldn’t help thinking that if you had said something like that to Steve Jobs, he would have taken your head off with a dull knife. The whole idea of vision at a place like Apple is that you pursue things you can’t fully quantify or measure; that great product design is an art, and the most important changes are the ones you intuit rather than prove in advance.
But engineers are trained not to act on intuition. You are allowed to have intuition, of course, but you use it to make hypotheses, which you then test. You act on the results of those tests.
There have been other big companies run by engineers, of course. HP in its glory days was a great example. But those companies were almost always wedded to traditional long-term planning cycles. What makes Google unusual is its combination of an engineer’s love of scientific method with the web’s rapid iterative development. Put those two characteristics together, and Google often behaves like a big bundle of short-term science experiments.
Why did you kill my favorite product? Take Google’s bizarre practice of publicly killing products. To most companies, killing a product is a shameful thing. It disappoints customers, and it hurts your own ego because it’s an admission that you failed. Most companies hide their product cancellations: they try to disguise them as a “reallocation” or “new focus” or some other doublespeak.
Google does the exact opposite – a couple of times a year it trumpets to the world that it’s terminating products and services that millions of people love and rely on. Google isn’t merely up front about these cancellations; it’s downright cheerful, as if turning off Google Reader or Google Desktop is an accomplishment to be proud of.
And to Google, maybe it is. If you look at the world through the eyes of the scientific method, every Google project is an experiment, and experiments must be periodically reviewed. When an experiment is completed, you either choose to follow up on it, or you terminate it and move on to something else. A scientist doesn’t get emotional about this; it’s the way the system works, and everyone knows that it’s all for the best.
By announcing its terminated experiments, I think Google isn’t admitting failure, it’s proudly demonstrating that scientific principles are in use. I think Google’s management views the cancellations as proof that it’s being focused and logical.
Google management: Who’s in charge here?
The second unusual aspect of Google is its ownership structure. Never forget: Google is not really a public company. Sure, it has stock and all the other attributes of a normal public company, but 56.7% of Google’s voting shares are held by cofounders Sergey Brin and Larry Page (link). As long as they remain friends, they can do whatever they want with the company, and they cannot be fired.
I don’t have a problem with that. Google has always been up front about it, and besides I’ve seen many large public companies manage themselves into ruin in pursuit of quarterly returns. It’s refreshing to see a big company that doesn’t enslave itself to the quarterly report. As Page put it in 2004, “by investing in Google, you are placing an unusual long term bet on the team, especially Sergey and me” (link).
How long term is that bet? I’m not sure Google’s senior management even thinks in terms of annual returns, let alone quarterly. Brin and Page are both about 40 years old as of 2013. They have a life expectancy of about 38 more years, to about 2050, and I have no reason to think that they plan to work anywhere else in their lives. So I think Google’s planning horizon goes to at least the year 2050. Page himself likes to talk about his 50-year planning horizon, so he may well be thinking out to the 2060s.
To put that in context, some scientists predict that we’ll achieve superhuman machine intelligence well before 2050 (link). I’m not endorsing that timeline, by the way; I think it may be optimistic. But my point is, Google could be planning almost anything.
Combine the first two unique things about Google and you get an interesting picture. Most companies have a long, detailed planning cycle in pursuit of quarterly goals. That often makes them very predictable. It also makes it hard for them to get anything done – when your planning cycle is longer than your goal cycle, you’ll often change goals faster than you can achieve any of them.
Google does just the opposite. It has a short, unpredictable planning cycle in pursuit of very long-term objectives. It’s likely to pursue those objectives relentlessly, but its near term actions will look random, because they’re just Darwinian experiments along the way.
In other words, there is probably a method to Google’s madness, but they’re not going to tell you what it is.
But there’s one more factor about Google that we need to consider: it’s run by human beings. Larry Page is not Spock. No matter how logical and dispassionate he tries to be, he and the rest of Google’s managers have psychological needs and reactions that they cannot transcend. That means Google has corporate politics.
Google politics: The coming-out party of Larry Page
I don’t think you can fully explain Google’s behavior over the last several years without looking at the relationship between its CEOs during that time, Eric Schmidt and Larry Page. Google’s first CEO, in its very early days, was Page. Investors convinced Page and Brin that they needed to bring in professional management to organize the company. Reluctantly they agreed, and supposedly Steve Jobs was at the top of their wish list. That raises some fascinating what-if scenarios, but Jobs was already occupied, and eventually they settled on Eric Schmidt, formerly of Sun.
A video of Page from 2000 gives an interesting insight into his thinking at the time. It was recorded a year before Schmidt joined Google. A nonprofit called the Academy of Achievement recorded video interviews with Page and Brin. The videos are a fascinating window into the early thinking of both men. In one clip, Page is asked about the challenges of being a CEO at age 27 (link). He replies:
So Page acknowledged his need for tutoring in management, but at the same time he went out of his way to call himself a visionary. I haven’t met Larry Page, but there’s one thing I know for sure: anyone who calls himself a visionary at age 27 does not lack for confidence.
Schmidt arrived soon after, and for the next ten years Page served a kind of management apprenticeship under him. I don’t want to overstate Schmidt’s role; even then, Page and Brin had control of the company, and could have ousted Schmidt if they really wanted to. But even if Page agreed that working for Schmidt was necessary, it can’t have been easy.
Early in Schmidt’s tenure, he and Page appeared together to address students at Stanford. The session was recorded on video, and Stanford posted it online here (link). The whole video is worth watching, but the segment I’ve embedded below is especially interesting because it shows the sometimes awkward interaction between Schmidt and Page.
Schmidt is the more articulate of the two. He interrupts to preface things before Page can make a comment, and sometimes comes back afterward to put a different spin on something Page said. In this clip, watch Page’s face when Schmidt interrupts him to deliver the punchline at the end. You should judge it for yourself, but to me Schmidt and Page look like one of those married couples who value each other but also get on each-other’s nerves.
No matter how much Page appreciated Schmidt’s wisdom, no matter how fruitful their collaboration, it can’t have been easy for Page to be mentored like this for ten years. If I were in his shoes, I’d have compiled a long list of things I wanted to change as soon as I was in charge.
That time came in 2011, when Page returned as CEO and Schmidt was kicked upstairs to be Google’s Chairman and chief explainer (link).
Page acted quickly, reorganizing the company and accelerating the termination of projects (link). I think that helped reinforce the use of the scientific method. It also helped Page assert his authority.
Then Page bought Motorola Mobility for over $12 billion. I don’t think you can understand the Motorola deal without taking into account the management change at Google. It was Page’s first major business deal as CEO, a chance to finally spread his wings and put his distinctive stamp on the company. Any human being with Page’s experience and ego would want to do something like that. So I believe ego played a role in the Motorola deal. But I don’t think that was the only motivation.
My take on why Google bought Motorola
Remember Google’s business situation in 2011. It still had huge economic resources, but it was no longer the dynamic new kid in the industry. That crown had fallen to Facebook, which was growing like a weed and which was not Google’s friend. At the time, Google was kicking itself for failing to recognize the threat earlier, and for responding to it so ineptly. I’m sure Page was adamant that he didn’t want to repeat that mistake.
Like social networking, mobile was a critical growth area for Google. The threat in mobile was Apple, which was doing a great job of integrating hardware and software to produce superior products. Many people at the time felt Google was destined to play second fiddle to Apple in mobile forever.
Then the opportunity came along to buy Motorola. Here’s how I think that parsed to Google:
—If people are right about Apple’s power in system design, we may need to move much more aggressively into mobile hardware than we have to date. If that happens, owning Motorola gives us a head start.
—Even if we don’t end up needing Motorola’s hardware business, we’ll learn an enormous amount from managing the company. Those skills and insights will help us manage our other hardware licensees.
—We’re going to pay a bunch of money for the patents anyway, so why not buy the whole thing? We might end up writing off most of the purchase, but who cares about annual returns? It’s better to have a bad year than take the risk of being blind-sided the way we were by Facebook.
I think the Motorola deal wasn’t just about the patents or about making a profit in device sales. It was about buying insurance against a surprise from mobile device manufacturers, especially Apple. If you think of Google as a company that sets long-term objectives and then runs experiments in pursuit of them, the Motorola deal is just an unusually large experiment along the road to mobile.
Add to that chain of logic Page’s natural desire to exercise his new powers, and the Motorola deal starts to look very understandable to me.
So was the deal worth the money? It’s too early to tell, but I doubt Larry Page is even asking that question. As long as Google learns from the purchase and doesn’t get blindsided in hardware, the deal served its purpose.
What happens next?
If you’re an investor, you should expect more off-the-wall acquisitions and product cancellations from Google. They’re built into the system. But I think Google’s unusual culture and management structure give it some other fairly predictable weaknesses. Those are potential opportunities for competitors, vulnerabilities for Google to guard against, and issues for investors to consider.
Weakness #1: Wandering vision. Google’s iterative development approach is very effective for pursuing a long-term goal when the company has a clear idea of its destination. The company’s development of self-driving cars is a good example: by relentlessly testing and tweaking the design, they’ve made much more progress than I believed was possible. Like most people in Silicon Valley, I’ve had the experience of driving on the freeway alongside those Google cars, and it’s very impressive (except for the fact that they adhere rigidly to the speed limit, but that’s a subject for a different post).
Google is much less effective when its original goal in a market changes. Because of its quick-reaction nature, Google frequently launches projects that seem very important at the time, but later turn out to be not so critical after all. The market evolves, priorities change, maybe a competitor becomes less prominent. When that happens, the Google projects are in danger of cancellation, and nobody likes working on a canceled project. So the teams frequently start iterating on their goals the same way they would on their features. Usually they end up chasing the latest trendy issue in search of a revenue stream and continued existence.
That’s usually the road to hell. Once a project starts changing goals, it’s almost impossible to diagnose the cause of any problems it has with market acceptance. Did we choose the wrong goal, or did we execute poorly? It’s usually impossible to tell.
To put it in scientific terms, it’s like running an experiment in which you have several independent variables. Good luck interpreting your results.
Google Docs is a great example. It was launched to undercut Microsoft’s Office franchise. Over time as Microsoft became weaker, that was no longer a compelling reason for existence, and Docs was merged into Drive and repurposed as a competitor to the newly-trendy Dropbox. Feature evolution in the core applications moved at a crawl.
Now there are two new challenges to Drive/Docs: Apple is turning iWork into a cross-platform web app, and Flickr has upped the stakes in the free storage race to a terabyte (yes, I know Flickr is photos only, but you don’t really think Yahoo will stop there, do you?) Which threat will the Drive team respond to? I don’t know, but because of the way they’ve been wandering there’s a very good chance they’ll end up below critical mass against all of their chosen competitors.
Weakness #2: Poor external communication. Scientists aren’t generally knows as great public communicators, and there’s a reason for that. PR is the art of telling a story in a way that people are open to hearing. To the scientific mindset, that comes across like dishonesty and manipulation. A scientist wants people to believe things because they make logical sense, not because their emotions are engaged.
Adding to that challenge, Google is very bad at anticipating how people and companies will react to its initiatives. Time and again, Google has taken actions that it tried earnestly to explain logically, and been surprised and hurt when people didn’t understand. I think Google views itself as a highly principled company pursuing the good of humanity; it expects people to give it the benefit of the doubt when there’s confusion, and to understand the good intent behind its actions. Google’s management doesn’t seem to understand that a hyper-rich company whose founders have private jumbo jets is automatically an object of jealousy and suspicion. Or if they do understand it, they aren’t willing to take the steps necessary to counter it.
One prominent example of Google’s communication problem was book digitization. Google was trying to make out-of-print books more available to the public, a noble goal by almost anyone’s standards. But Google handled the process so clumsily and arrogantly that it frightened authors into allying with publishers, an outcome equivalent to getting wild cats and dogs to sit down together for tea.
A second example was the backlash from the purchase of Motorola. It’s hard to overstate what a profound shock the Motorola deal was to Google’s Android licensees. Before the deal, the handset companies and operators viewed Google as a benign giant who could be trusted to champion mobile data without preying on its licensees. After the deal, they viewed Google as a villain little different from Microsoft.
The irony of the deal is that the threat from Apple has receded somewhat, so the Motorola experiment probably wasn’t needed. The rising challenge to Google now is that an increasingly feisty Samsung has too much market power in the Android space, and there’s a rising Amazon-inspired movement to fork Android and take control of it away from Google. The Motorola acquisition made companies like Samsung much more likely to cooperate with a non-Google OS. In trying to prevent a Facebook-style breakout in mobile, Google actually weakened its position in the mobile market.
Even casual public comments can create trouble for Google. In response to a question at the Google IO conference in 2013, Larry Page said of Oracle: “We’ve had a difficult relationship with Oracle.... money is probably more important to them than having any kind of collaboration.” (link)
There are several problems with this statement. First, if you want a cooperative relationship with Oracle, calling them a bunch of greedy bastards isn’t the way to get it. Second, public companies are supposed to put making money ahead of collaboration. That’s what their shareholders expect. This is a good example of how Google’s thinking is out of step with typical corporate governance.
The third problem is that Page’s comments came across to some people as hypocrisy:
Om Malik: “I think Larry (and all other technology industry leaders) should actually practice what they preach.” (link)
Slate: “Page criticized Microsoft for treating Google as a rival, blasted Oracle for caring too much about money, and then whined about everyone being so negative. Heck, if it weren’t for those other companies standing in the way, Google would have probably already solved world hunger. Well, except for all the laws and bureaucrats and journalists who are also standing in the way.” (link)
John Gruber: “Google is a hyper-competitive company, and they repeatedly enter markets that already exist and crush competitors. Nothing wrong with that. That’s how capitalism is supposed to work, and Google’s successes are admirable. But there’s nothing stupid about seeing Google being pitted “versus” other companies. They want everything; their ambition is boundless.” (link)
Gruber’s comments show the trouble that Google gets itself into when poor communication combines with its wandering product goals. Google doesn’t see itself as a predator eating tech startups, but when its internal projects start iterating on their goals, they inevitably target successful startups because that seems like the logical thing to do. The behavior is a natural outcome of the way the company works. Larry Page says he’s all about cooperation and I think he means it, but his product teams relentlessly stalk the latest hot startup. The result is a company that talks like a charitable foundation but acts like a pack of wolves.
No wonder he gets labeled a hypocrite.
Google’s trouble communicating its own intentions, and the mismatch between its words and behavior, becomes a serious problem whenever the company has to deal with big political or PR battles. Google’s competitors are often better at courting public opinion, and that opinion often drives the outcome of political processes. If you want an example, watch Google struggle with European Union regulators.
Weakness #3: Science vs. art in product management. Google’s strength in science and quick response makes it very fast at incrementally improving the performance and reliability of its products. But that same process makes it almost impossible for Google to lead in features or product ideas that can’t be proved or verified through research. That’s why Google struggles in user experience, creating new product categories, and fitting its products to the latent needs of users: all of those are intuition-led activities in which it’s very hard to prove ahead of time what’s right or wrong. Even if there are people within Google who have extraordinary taste and vision, it’s very hard for them to drive action because their ideas can’t pass the science-style review process that Google uses for decision-making.
That puts Google at a disadvantage when competing with vision-led companies. The most obvious example of this is Google vs. Apple. When Apple is implementing its strategy properly, it comes up with new product categories faster than Google can co-opt them, and executes them with more taste and usability. As long as Apple can keep moving the bar, Google is forced to play catch-up to Apple’s leadership.
(The big question post-Steve is whether Apple can continue to move the bar. But that’s another topic for a separate article.)
The exception to normal Google decision-making is the special projects run by Sergey Brin. In those projects, Google chooses a few long-term product goals that can’t necessarily be justified logically, but that look possible and would have a big impact if they succeeded. It’s a logical way for an analytical company to try to inject some vision into its business.
What we don’t know yet about those special projects is whether Google can apply the smaller dashes of intuition that are needed throughout the development process to pioneer a new product category. The iPod wasn’t just a good idea, it was a long series of clever decisions that Apple made in the design of the device, software, store, and ecosystem. They all fit together to make a great music management system. Can Google make a similar series of great, coordinated decisions to create a compelling user need for Glass, or will its glasses just be a technophile toy? I don’t think we’ve seen the answer yet. Until we do, there’s a strong danger that Google is just doing the advanced R&D that some other company will use to make a successful wearable computing device.
Should Google try to change?
Every successful company has weaknesses. The strengths that make it powerful always create corresponding blind spots and vulnerabilities. Google’s strengths are unusually well suited to its core business of search advertising. The Internet is so big that you have to use some sort of algorithmic process to organize it, and it takes a vast series of logical experiments to gradually tune search results and the delivery of advertising around them.
The question for investors is if or when Google will run out of room to grow in the search advertising market. At that time, to maintain its growth (and stock value), it’ll need other substantial sources of profit. Can Google find other businesses in which its analytical, experimental culture will produce winners? Or can it adapt its culture to the needs of other markets?
So far, the signs aren’t promising. Google is very good at giving away technology (Android, for example), but not very effective at making large amounts of money from it. Google’s product experiments have produced many failures and a few popular services, but very little in terms of major incremental profit. In fact, some financial analysts refer to two Googles – the search engine company that makes all the profit, and the other Google that sucks away some of that profit.
It’s easy for someone like me to say that Google should change its culture to give it a better chance of success in other markets, but in the real world those culture-changing experiments often fail catastrophically. You end up destroying the source of your previous success, without successfully transitioning to a new winning culture. In that vein, I worry that even the Motorola deal is a risk for Google, as it brought into the company a huge number of employees trained in a very different, famously dysfunctional culture.
For now, the search business is so strong that I don’t think Google is likely to make major changes in the way it works. Companies rarely change until they have to. Until and unless that happens, Google is likely to continue its scientific management, and competitors are likely to continue countering it through vision, public communication, and product management.
If you’re a Google investor, I think the situation is still the same as it was at Google's IPO: You’ve made an unusual long-term bet on Page and Brin and their scientific approach to running a tech company. It’s quirky and it’s different from the way most other companies operate, but it does make its own logical sense, if you look at the world through the eyes of an engineer.
A favorite pastime among people who watch the tech industry is trying to figure out why Google does things. The Verge was downright plaintive about it the other day (link), and I get the question frequently from financial analysts and reporters. But the topic also comes up regularly in conversations with my Silicon Valley friends.
It’s a puzzle because Google doesn’t seem to respond to the rules and logic used by the rest of the business world. It passes up what look like obvious opportunities, invests heavily in things that look like black holes, and proudly announces product cancellations that the rest of us would view as an embarrassment. Google’s behavior drives customers and partners nuts, but is especially troubling to financial analysts who have to tell people whether or not to buy Google’s stock. Every time Google has a less than stellar quarter, the issue surges up again.
As I wrote recently when discussing Dell (link), it’s a mistake to assume there’s a logical reason for everything a company does. Sometimes managers act out of fear or ignorance or just plain stupidity, and trying to retrofit logic onto their actions is as pointless as a primitive shaman using goat entrails to explain a volcano.
But in Google’s case, I think its actions do make sense – even the deeply weird stuff like the purchase of Motorola. The issue, I believe, is that Google follows a different set of rules than most other companies. Apple uses “Think Different” as its slogan, but in many ways Google is the company that truly thinks differently. It’s not just marching to a different drummer; sometimes I think it hears an entirely different orchestra.
Google’s orchestra is unique because of three factors: corporate culture, governance, and personal politics. Let’s start with the culture.
Google culture: You are what you do
The strategic thinking of most companies is shaped by the way they do business. For example, a farmer thinks in terms of annual seasons and crops; everything revolves around that yearly cycle. Manufacturing companies, the traditional foundation of a 20th century economy, plan in terms of big projects that take a long time to implement and require a lot of preparation. If you’re building a car or a plane or even a smartphone, you have to plan its features well in advance, drive hardware and software to completion at the same time, and arrange manufacturing and distribution long before you actually build anything. The companies that build complex physical things naturally plan their products in terms of lifecycles lasting at least 12 to 24 months, and sometimes much longer.
That long planning cycle dominated big companies in the 20th century, and was driven into all our heads through generations of business books and business school classes. It’s how most of our brains were formatted.
An internet company, like Google, works at a fundamentally different pace. Web software changes continuously. You don’t plan it rigidly; you evolve it day by day in response to the behavior of customers. The faster and more flexibly you evolve, the more successful your products will be.
This evolutionary approach, and the Agile design processes that support it, is built into the fiber and psyche of web companies. They don’t think in terms of long-term detailed plans; they think in terms of stimulus and response.
This is a dramatic change in the history of business. In the past, the nimble companies were always the little ones. The larger your company, the more it valued planning and the long-term view. Google is one of the first very large tech companies ever to pride itself on rapid response rather than rigid planning.
On top of this quick-turn bias there’s the cultural training of Google’s senior management. Most big companies end up being run by professional managers who came up through business school or finance, where they get trained in the rhythms and personality of traditional big business. They learn a shared vocabulary and set of values that are very familiar and comfortable to investors. By contrast, Google is completely controlled by engineering PhDs. They speak the language of science rather than business, and they’re contemptuous of the vague directional platitudes and reassuring noises made by modern finance and marketing.
I think most reporters and analysts don’t understand how fundamentally different the engineering mindset is from traditional business thinking. It’s a very distinct paradigm, unfamiliar to most people who haven’t studied science (link).
One key element of the engineering mindset is the use of scientific method: you encourage a Darwinian marketplace of ideas, you test those ideas through controlled experiments, and you make decisions based on experimental data.
In its behavior and vocabulary, Google oozes scientific method. A couple of times recently I’ve heard Google executives say in public, “if you can’t measure it, you can’t improve it” (link). It's an old quote, dating back at least to Lord Kelvin in the 1800s. It's also a subtle twist on the traditional mantra used in web design: “that which you measure, you can improve.” The web design version says you should measure everything you can; the Google executive version implies that nothing really matters unless you can measure it.
That’s a very scientific, rational point of view, but I couldn’t help thinking that if you had said something like that to Steve Jobs, he would have taken your head off with a dull knife. The whole idea of vision at a place like Apple is that you pursue things you can’t fully quantify or measure; that great product design is an art, and the most important changes are the ones you intuit rather than prove in advance.
But engineers are trained not to act on intuition. You are allowed to have intuition, of course, but you use it to make hypotheses, which you then test. You act on the results of those tests.
There have been other big companies run by engineers, of course. HP in its glory days was a great example. But those companies were almost always wedded to traditional long-term planning cycles. What makes Google unusual is its combination of an engineer’s love of scientific method with the web’s rapid iterative development. Put those two characteristics together, and Google often behaves like a big bundle of short-term science experiments.
Why did you kill my favorite product? Take Google’s bizarre practice of publicly killing products. To most companies, killing a product is a shameful thing. It disappoints customers, and it hurts your own ego because it’s an admission that you failed. Most companies hide their product cancellations: they try to disguise them as a “reallocation” or “new focus” or some other doublespeak.
Google does the exact opposite – a couple of times a year it trumpets to the world that it’s terminating products and services that millions of people love and rely on. Google isn’t merely up front about these cancellations; it’s downright cheerful, as if turning off Google Reader or Google Desktop is an accomplishment to be proud of.
And to Google, maybe it is. If you look at the world through the eyes of the scientific method, every Google project is an experiment, and experiments must be periodically reviewed. When an experiment is completed, you either choose to follow up on it, or you terminate it and move on to something else. A scientist doesn’t get emotional about this; it’s the way the system works, and everyone knows that it’s all for the best.
By announcing its terminated experiments, I think Google isn’t admitting failure, it’s proudly demonstrating that scientific principles are in use. I think Google’s management views the cancellations as proof that it’s being focused and logical.
Google management: Who’s in charge here?
The second unusual aspect of Google is its ownership structure. Never forget: Google is not really a public company. Sure, it has stock and all the other attributes of a normal public company, but 56.7% of Google’s voting shares are held by cofounders Sergey Brin and Larry Page (link). As long as they remain friends, they can do whatever they want with the company, and they cannot be fired.
I don’t have a problem with that. Google has always been up front about it, and besides I’ve seen many large public companies manage themselves into ruin in pursuit of quarterly returns. It’s refreshing to see a big company that doesn’t enslave itself to the quarterly report. As Page put it in 2004, “by investing in Google, you are placing an unusual long term bet on the team, especially Sergey and me” (link).
How long term is that bet? I’m not sure Google’s senior management even thinks in terms of annual returns, let alone quarterly. Brin and Page are both about 40 years old as of 2013. They have a life expectancy of about 38 more years, to about 2050, and I have no reason to think that they plan to work anywhere else in their lives. So I think Google’s planning horizon goes to at least the year 2050. Page himself likes to talk about his 50-year planning horizon, so he may well be thinking out to the 2060s.
To put that in context, some scientists predict that we’ll achieve superhuman machine intelligence well before 2050 (link). I’m not endorsing that timeline, by the way; I think it may be optimistic. But my point is, Google could be planning almost anything.
Combine the first two unique things about Google and you get an interesting picture. Most companies have a long, detailed planning cycle in pursuit of quarterly goals. That often makes them very predictable. It also makes it hard for them to get anything done – when your planning cycle is longer than your goal cycle, you’ll often change goals faster than you can achieve any of them.
Google does just the opposite. It has a short, unpredictable planning cycle in pursuit of very long-term objectives. It’s likely to pursue those objectives relentlessly, but its near term actions will look random, because they’re just Darwinian experiments along the way.
In other words, there is probably a method to Google’s madness, but they’re not going to tell you what it is.
But there’s one more factor about Google that we need to consider: it’s run by human beings. Larry Page is not Spock. No matter how logical and dispassionate he tries to be, he and the rest of Google’s managers have psychological needs and reactions that they cannot transcend. That means Google has corporate politics.
Google politics: The coming-out party of Larry Page
I don’t think you can fully explain Google’s behavior over the last several years without looking at the relationship between its CEOs during that time, Eric Schmidt and Larry Page. Google’s first CEO, in its very early days, was Page. Investors convinced Page and Brin that they needed to bring in professional management to organize the company. Reluctantly they agreed, and supposedly Steve Jobs was at the top of their wish list. That raises some fascinating what-if scenarios, but Jobs was already occupied, and eventually they settled on Eric Schmidt, formerly of Sun.
A video of Page from 2000 gives an interesting insight into his thinking at the time. It was recorded a year before Schmidt joined Google. A nonprofit called the Academy of Achievement recorded video interviews with Page and Brin. The videos are a fascinating window into the early thinking of both men. In one clip, Page is asked about the challenges of being a CEO at age 27 (link). He replies:
"If you manage people for 20 years, or something like that, you pick up things. So I certainly lack experience there, and that's an issue. But I sort of make up for that, I think, in terms of understanding where things are going to go, having a vision about the future, and really understanding the industry I am in, and what the company does."
So Page acknowledged his need for tutoring in management, but at the same time he went out of his way to call himself a visionary. I haven’t met Larry Page, but there’s one thing I know for sure: anyone who calls himself a visionary at age 27 does not lack for confidence.
Schmidt arrived soon after, and for the next ten years Page served a kind of management apprenticeship under him. I don’t want to overstate Schmidt’s role; even then, Page and Brin had control of the company, and could have ousted Schmidt if they really wanted to. But even if Page agreed that working for Schmidt was necessary, it can’t have been easy.
Early in Schmidt’s tenure, he and Page appeared together to address students at Stanford. The session was recorded on video, and Stanford posted it online here (link). The whole video is worth watching, but the segment I’ve embedded below is especially interesting because it shows the sometimes awkward interaction between Schmidt and Page.
Schmidt is the more articulate of the two. He interrupts to preface things before Page can make a comment, and sometimes comes back afterward to put a different spin on something Page said. In this clip, watch Page’s face when Schmidt interrupts him to deliver the punchline at the end. You should judge it for yourself, but to me Schmidt and Page look like one of those married couples who value each other but also get on each-other’s nerves.
No matter how much Page appreciated Schmidt’s wisdom, no matter how fruitful their collaboration, it can’t have been easy for Page to be mentored like this for ten years. If I were in his shoes, I’d have compiled a long list of things I wanted to change as soon as I was in charge.
That time came in 2011, when Page returned as CEO and Schmidt was kicked upstairs to be Google’s Chairman and chief explainer (link).
Page acted quickly, reorganizing the company and accelerating the termination of projects (link). I think that helped reinforce the use of the scientific method. It also helped Page assert his authority.
Then Page bought Motorola Mobility for over $12 billion. I don’t think you can understand the Motorola deal without taking into account the management change at Google. It was Page’s first major business deal as CEO, a chance to finally spread his wings and put his distinctive stamp on the company. Any human being with Page’s experience and ego would want to do something like that. So I believe ego played a role in the Motorola deal. But I don’t think that was the only motivation.
My take on why Google bought Motorola
Remember Google’s business situation in 2011. It still had huge economic resources, but it was no longer the dynamic new kid in the industry. That crown had fallen to Facebook, which was growing like a weed and which was not Google’s friend. At the time, Google was kicking itself for failing to recognize the threat earlier, and for responding to it so ineptly. I’m sure Page was adamant that he didn’t want to repeat that mistake.
Like social networking, mobile was a critical growth area for Google. The threat in mobile was Apple, which was doing a great job of integrating hardware and software to produce superior products. Many people at the time felt Google was destined to play second fiddle to Apple in mobile forever.
Then the opportunity came along to buy Motorola. Here’s how I think that parsed to Google:
—If people are right about Apple’s power in system design, we may need to move much more aggressively into mobile hardware than we have to date. If that happens, owning Motorola gives us a head start.
—Even if we don’t end up needing Motorola’s hardware business, we’ll learn an enormous amount from managing the company. Those skills and insights will help us manage our other hardware licensees.
—We’re going to pay a bunch of money for the patents anyway, so why not buy the whole thing? We might end up writing off most of the purchase, but who cares about annual returns? It’s better to have a bad year than take the risk of being blind-sided the way we were by Facebook.
I think the Motorola deal wasn’t just about the patents or about making a profit in device sales. It was about buying insurance against a surprise from mobile device manufacturers, especially Apple. If you think of Google as a company that sets long-term objectives and then runs experiments in pursuit of them, the Motorola deal is just an unusually large experiment along the road to mobile.
Add to that chain of logic Page’s natural desire to exercise his new powers, and the Motorola deal starts to look very understandable to me.
So was the deal worth the money? It’s too early to tell, but I doubt Larry Page is even asking that question. As long as Google learns from the purchase and doesn’t get blindsided in hardware, the deal served its purpose.
What happens next?
If you’re an investor, you should expect more off-the-wall acquisitions and product cancellations from Google. They’re built into the system. But I think Google’s unusual culture and management structure give it some other fairly predictable weaknesses. Those are potential opportunities for competitors, vulnerabilities for Google to guard against, and issues for investors to consider.
Weakness #1: Wandering vision. Google’s iterative development approach is very effective for pursuing a long-term goal when the company has a clear idea of its destination. The company’s development of self-driving cars is a good example: by relentlessly testing and tweaking the design, they’ve made much more progress than I believed was possible. Like most people in Silicon Valley, I’ve had the experience of driving on the freeway alongside those Google cars, and it’s very impressive (except for the fact that they adhere rigidly to the speed limit, but that’s a subject for a different post).
Google is much less effective when its original goal in a market changes. Because of its quick-reaction nature, Google frequently launches projects that seem very important at the time, but later turn out to be not so critical after all. The market evolves, priorities change, maybe a competitor becomes less prominent. When that happens, the Google projects are in danger of cancellation, and nobody likes working on a canceled project. So the teams frequently start iterating on their goals the same way they would on their features. Usually they end up chasing the latest trendy issue in search of a revenue stream and continued existence.
That’s usually the road to hell. Once a project starts changing goals, it’s almost impossible to diagnose the cause of any problems it has with market acceptance. Did we choose the wrong goal, or did we execute poorly? It’s usually impossible to tell.
To put it in scientific terms, it’s like running an experiment in which you have several independent variables. Good luck interpreting your results.
Google Docs is a great example. It was launched to undercut Microsoft’s Office franchise. Over time as Microsoft became weaker, that was no longer a compelling reason for existence, and Docs was merged into Drive and repurposed as a competitor to the newly-trendy Dropbox. Feature evolution in the core applications moved at a crawl.
Now there are two new challenges to Drive/Docs: Apple is turning iWork into a cross-platform web app, and Flickr has upped the stakes in the free storage race to a terabyte (yes, I know Flickr is photos only, but you don’t really think Yahoo will stop there, do you?) Which threat will the Drive team respond to? I don’t know, but because of the way they’ve been wandering there’s a very good chance they’ll end up below critical mass against all of their chosen competitors.
Weakness #2: Poor external communication. Scientists aren’t generally knows as great public communicators, and there’s a reason for that. PR is the art of telling a story in a way that people are open to hearing. To the scientific mindset, that comes across like dishonesty and manipulation. A scientist wants people to believe things because they make logical sense, not because their emotions are engaged.
Adding to that challenge, Google is very bad at anticipating how people and companies will react to its initiatives. Time and again, Google has taken actions that it tried earnestly to explain logically, and been surprised and hurt when people didn’t understand. I think Google views itself as a highly principled company pursuing the good of humanity; it expects people to give it the benefit of the doubt when there’s confusion, and to understand the good intent behind its actions. Google’s management doesn’t seem to understand that a hyper-rich company whose founders have private jumbo jets is automatically an object of jealousy and suspicion. Or if they do understand it, they aren’t willing to take the steps necessary to counter it.
One prominent example of Google’s communication problem was book digitization. Google was trying to make out-of-print books more available to the public, a noble goal by almost anyone’s standards. But Google handled the process so clumsily and arrogantly that it frightened authors into allying with publishers, an outcome equivalent to getting wild cats and dogs to sit down together for tea.
A second example was the backlash from the purchase of Motorola. It’s hard to overstate what a profound shock the Motorola deal was to Google’s Android licensees. Before the deal, the handset companies and operators viewed Google as a benign giant who could be trusted to champion mobile data without preying on its licensees. After the deal, they viewed Google as a villain little different from Microsoft.
The irony of the deal is that the threat from Apple has receded somewhat, so the Motorola experiment probably wasn’t needed. The rising challenge to Google now is that an increasingly feisty Samsung has too much market power in the Android space, and there’s a rising Amazon-inspired movement to fork Android and take control of it away from Google. The Motorola acquisition made companies like Samsung much more likely to cooperate with a non-Google OS. In trying to prevent a Facebook-style breakout in mobile, Google actually weakened its position in the mobile market.
Even casual public comments can create trouble for Google. In response to a question at the Google IO conference in 2013, Larry Page said of Oracle: “We’ve had a difficult relationship with Oracle.... money is probably more important to them than having any kind of collaboration.” (link)
There are several problems with this statement. First, if you want a cooperative relationship with Oracle, calling them a bunch of greedy bastards isn’t the way to get it. Second, public companies are supposed to put making money ahead of collaboration. That’s what their shareholders expect. This is a good example of how Google’s thinking is out of step with typical corporate governance.
The third problem is that Page’s comments came across to some people as hypocrisy:
Om Malik: “I think Larry (and all other technology industry leaders) should actually practice what they preach.” (link)
Slate: “Page criticized Microsoft for treating Google as a rival, blasted Oracle for caring too much about money, and then whined about everyone being so negative. Heck, if it weren’t for those other companies standing in the way, Google would have probably already solved world hunger. Well, except for all the laws and bureaucrats and journalists who are also standing in the way.” (link)
John Gruber: “Google is a hyper-competitive company, and they repeatedly enter markets that already exist and crush competitors. Nothing wrong with that. That’s how capitalism is supposed to work, and Google’s successes are admirable. But there’s nothing stupid about seeing Google being pitted “versus” other companies. They want everything; their ambition is boundless.” (link)
Gruber’s comments show the trouble that Google gets itself into when poor communication combines with its wandering product goals. Google doesn’t see itself as a predator eating tech startups, but when its internal projects start iterating on their goals, they inevitably target successful startups because that seems like the logical thing to do. The behavior is a natural outcome of the way the company works. Larry Page says he’s all about cooperation and I think he means it, but his product teams relentlessly stalk the latest hot startup. The result is a company that talks like a charitable foundation but acts like a pack of wolves.
No wonder he gets labeled a hypocrite.
Google’s trouble communicating its own intentions, and the mismatch between its words and behavior, becomes a serious problem whenever the company has to deal with big political or PR battles. Google’s competitors are often better at courting public opinion, and that opinion often drives the outcome of political processes. If you want an example, watch Google struggle with European Union regulators.
Weakness #3: Science vs. art in product management. Google’s strength in science and quick response makes it very fast at incrementally improving the performance and reliability of its products. But that same process makes it almost impossible for Google to lead in features or product ideas that can’t be proved or verified through research. That’s why Google struggles in user experience, creating new product categories, and fitting its products to the latent needs of users: all of those are intuition-led activities in which it’s very hard to prove ahead of time what’s right or wrong. Even if there are people within Google who have extraordinary taste and vision, it’s very hard for them to drive action because their ideas can’t pass the science-style review process that Google uses for decision-making.
That puts Google at a disadvantage when competing with vision-led companies. The most obvious example of this is Google vs. Apple. When Apple is implementing its strategy properly, it comes up with new product categories faster than Google can co-opt them, and executes them with more taste and usability. As long as Apple can keep moving the bar, Google is forced to play catch-up to Apple’s leadership.
(The big question post-Steve is whether Apple can continue to move the bar. But that’s another topic for a separate article.)
The exception to normal Google decision-making is the special projects run by Sergey Brin. In those projects, Google chooses a few long-term product goals that can’t necessarily be justified logically, but that look possible and would have a big impact if they succeeded. It’s a logical way for an analytical company to try to inject some vision into its business.
What we don’t know yet about those special projects is whether Google can apply the smaller dashes of intuition that are needed throughout the development process to pioneer a new product category. The iPod wasn’t just a good idea, it was a long series of clever decisions that Apple made in the design of the device, software, store, and ecosystem. They all fit together to make a great music management system. Can Google make a similar series of great, coordinated decisions to create a compelling user need for Glass, or will its glasses just be a technophile toy? I don’t think we’ve seen the answer yet. Until we do, there’s a strong danger that Google is just doing the advanced R&D that some other company will use to make a successful wearable computing device.
Should Google try to change?
Every successful company has weaknesses. The strengths that make it powerful always create corresponding blind spots and vulnerabilities. Google’s strengths are unusually well suited to its core business of search advertising. The Internet is so big that you have to use some sort of algorithmic process to organize it, and it takes a vast series of logical experiments to gradually tune search results and the delivery of advertising around them.
The question for investors is if or when Google will run out of room to grow in the search advertising market. At that time, to maintain its growth (and stock value), it’ll need other substantial sources of profit. Can Google find other businesses in which its analytical, experimental culture will produce winners? Or can it adapt its culture to the needs of other markets?
So far, the signs aren’t promising. Google is very good at giving away technology (Android, for example), but not very effective at making large amounts of money from it. Google’s product experiments have produced many failures and a few popular services, but very little in terms of major incremental profit. In fact, some financial analysts refer to two Googles – the search engine company that makes all the profit, and the other Google that sucks away some of that profit.
It’s easy for someone like me to say that Google should change its culture to give it a better chance of success in other markets, but in the real world those culture-changing experiments often fail catastrophically. You end up destroying the source of your previous success, without successfully transitioning to a new winning culture. In that vein, I worry that even the Motorola deal is a risk for Google, as it brought into the company a huge number of employees trained in a very different, famously dysfunctional culture.
For now, the search business is so strong that I don’t think Google is likely to make major changes in the way it works. Companies rarely change until they have to. Until and unless that happens, Google is likely to continue its scientific management, and competitors are likely to continue countering it through vision, public communication, and product management.
If you’re a Google investor, I think the situation is still the same as it was at Google's IPO: You’ve made an unusual long-term bet on Page and Brin and their scientific approach to running a tech company. It’s quirky and it’s different from the way most other companies operate, but it does make its own logical sense, if you look at the world through the eyes of an engineer.
Wednesday, 3 July 2013
Conference Report: USENIX Annual Technical Conference (ATC) 2013
This year marks Google’s eleventh consecutive year as a sponsor of the USENIX Annual Technical Conference (ATC), just one of the co-located events at USENIX Federated Conference Week (FCW), which combines numerous conferences and workshops covering fields such as Autonomic Computing, Feedback Computing and much more in an intensive week of research, trends, and community interaction.
ATC provides a broad forum for computing systems research with an emphasis on implementations and experimental results. In addition to the Googlers presenting publications, we had two members on the program committee of ATC and several keynote speakers, invited speakers, panelists, committee members, and participants at the other co-located events at FCW.
In the paper Janus: Optimal Flash Provisioning for Cloud Storage Workloads, Googler Christoph Albrecht and co-authors demonstrated a system that allows users to make informed flash memory provisioning and partitioning decisions in cloud-scale distributed file systems that include both flash storage and disk tiers. As flash memory is still expensive, it is best to use it only for workloads that can make good use of it. Janus creates long term workload characterizations based on RPC samples and file age metadata. It uses these workload characterizations to formulate and solve an optimization problem that maximizes the reads sent to the flash tier. Based on evaluations from workloads using Janus, in use at Google for the past 6 months, the authors conclude that the recommendation system is quite effective, with flash hit rates using the optimized recommendations 47-76% higher than the option of using the flash as an unpartitioned tier.
In packetdrill: Scriptable Network Stack Testing, from Sockets to Packets, Google’s Neal Cardwell and co-authors showcased a portable, open-source scripting tool that enables testing the correctness and performance of network protocols. Despite their importance in modern computer systems, network protocols often undergo only ad hoc testing before their deployment, in large part due to their complexity. Furthermore, new algorithms have unforeseen interactions with other features, so testing has only become more daunting as TCP has evolved. The packetdrill tool was instrumental in the development of three new features for Linux TCP—Early Retransmit, Fast Open, and Loss Probes—and allowed the authors to find and fix 10 bugs in Linux. Furthermore, the team uses packetdrill in all phases of the development process for the kernel used in one of the world’s largest Linux installations. In the hope that sharing packetdrill with the community will make the process of improving Internet protocols an easier one, the source code and test scripts for packetdrill have been made freely available.
There were also additional refereed publications with Google co-authors at some of the co-located events at FCW, notably NicPic: Scalable and Accurate End-Host Rate Limiting, which outlines a system which enables accurate network traffic scheduling in a scalable fashion, and AGILE: Elastic Distributed Resource Scaling for Infrastructure-as-a-Service, a system that efficiently handles dynamic application workloads, reducing both penalties and user dissatisfaction.
Google is proud to support the academic community through conference participation and sponsorship. In particular, we are happy to mention one of the other interesting papers from this year’s USENIX FCW, co-authored by former Google PhD fellowship recipient Ashok Anand, MiG: Efficient Migration of Desktop VM Using Semantic Compression.
USENIX is a supporter of open access, so the papers and videos from the talks are available on the conference website.
Tuesday, 2 July 2013
Natural Language Understanding-focused awards announced
Some of the biggest challenges for the scientific community today involve understanding the principles and mechanisms that underlie natural language use on the Web. An example of long-standing problem is language ambiguity; when somebody types the word “Rio” in a query do they mean the city, a movie, a casino, or something else? Understanding the difference can be crucial to help users get the answer they are looking for. In the past few years, a significant effort in industry and academia has focused on disambiguating language with respect to Web-scale knowledge repositories such as Wikipedia and Freebase. These resources are used primarily as canonical, although incomplete, collections of “entities”. As entities are often connected in multiple ways, e.g., explicitly via hyperlinks and implicitly via factual information, such resources can be naturally thought of as (knowledge) graphs. This work has provided the first breakthroughs towards anchoring language in the Web to interpretable, albeit initially shallow, semantic representations. Google has brought the vision of semantic search directly to millions of users via the adoption of the Knowledge Graph. This massive change to search technology has also been called a shift “from strings to things”.
Understanding natural language is at the core of Google's work to help people get the information they need as quickly and easily as possible. At Google we work hard to advance the state of the art in natural language processing, to improve the understanding of fundamental principles, and to solve the algorithmic and engineering challenges to make these technologies part of everyday life. Language is inherently productive; an infinite number of meaningful new expressions can be formed by combining the meaning of their components systematically. The logical next step is the semantic modeling of structured meaningful expressions -- in other words, “what is said” about entities. We envision that knowledge graphs will support the next leap forward in language understanding towards scalable compositional analyses, by providing a universe of entities, facts and relations upon which semantic composition operations can be designed and implemented.
So we’ve just awarded over $1.2 million to support several natural language understanding research awards given to university research groups doing work in this area. Research topics range from semantic parsing to statistical models of life stories and novel compositional inference and representation approaches to modeling relations and events in the Knowledge Graph.
These awards went to researchers in nine universities and institutions worldwide, selected after a rigorous internal review:
- Mark Johnson and Lan Du (Macquarie University) and Wray Buntine (NICTA) for “Generative models of Life Stories”
- Percy Liang and Christopher Manning (Stanford University) for “Tensor Factorizing Knowledge Graphs”
- Sebastian Riedel (University College London) and Andrew McCallum (University of Massachusetts, Amherst) for “Populating a Knowledge Base of Compositional Universal Schema”
- Ivan Titov (University of Amsterdam) for “Learning to Reason by Exploiting Grounded Text Collections”
- Hans Uszkoreit (Saarland University and DFKI), Feiyu Xu (DFKI and Saarland University) and Roberto Navigli (Sapienza University of Rome) for “Language Understanding cum Knowledge Yield”
- Luke Zettlemoyer (University of Washington) for “Weakly Supervised Learning for Semantic Parsing with Knowledge Graphs”
We believe the results will be broadly useful to product development and will further scientific research. We look forward to working with these researchers, and we hope we will jointly push the frontier of natural language understanding research to the next level.
Subscribe to:
Posts (Atom)