Analyze This: Seattle as the Epicenter of Big Data

It affects the foods we eat, the politicians we vote for and the roads we drive on. It has changed the amount of legroom were allowed on airplanes and the speed at which life-saving medicines are invented. Its a safeguard against terrorist attacks but finds time to recommend a new shirt we might like, given all it knows about us. Its the Big Data industry, and the Puget Sound region, with the nations largest concentration of software engineers, has a shot at becoming its epicenter.

Ask a hundred people in tech what Big Data means and their answers will contain a hundred different nuances. At its most basic level, the buzzword is fairly self-explanatory: Its the collecting of massive amounts of information and then using it to improve how businesses and organizations do nearly everything.
In terms of game-changing potential, industry evangelists compare it to the advent of the internet. They say leading businesses that dont use Big Data will be gone within a decade.

To some, this sort of hype seems so over the top that it must be excessive. But industry analysts say as this new trend gathers momentum, this region the home of such data powerhouses as Amazon and Microsoft, data analytics companies like Tableau Software and Context Relevant and data collectors like Zillow and Inrix is bound to benefit.

Large-scale data crunching has existed for decades to some extent. Scientists at Fred Hutchinson Cancer Research Center used it to gain the upper hand on disease. Firms on Wall Street use it to gain market advantage.

To illustrate how things have changed in recent years, though, imagine a traditional supercomputer. Supercomputers are too small when you need them and too big the rest of the time, says Matt Wood, general manager, data science, at Amazon Web Services. When youre not using them and they just sit there, you have to pay to cool them, fix parts when they break and all that. But when you want to run your clinical trial or your discovery process, you never have all the capacity youd like.

Thats not an issue anymore. Thanks to better computing and a growing tendency to collect data in the cloud, Wood says, The limits have melted away. By using cloud-based services, he adds, companies can get as much computing power as they need when they need it.

When amazon launched it first Amazon Web Services (AWS) product on March 14, 2006, the event didnt elicit much fanfare. However, many have since pointed to that date as the beginning of a seismic shift in the IT world, comparable to the iPhones arrival in the mobile market.

Describing AWSs first product, named S3, an Amazon news release said it would provide a simple web services interface that can be used to store and retrieve any amount of data, at any time, from anywhere on the web. And just like that, Wood and others say the cloud industry was effectively born in Seattle. It is now estimated to be a $100 billion market.

Many are familiar with the term cloud by now, understanding it as services that allow documents, applications, operating systems and much more to be retrieved anywhere via the web. Perhaps most important: The cloud can offer cheap, seemingly infinite data storage and computational power at a time when both are direly needed.

According to the research firm International Data Corporation (IDC), the amount of digital data in the world nearly quadrupled between 2010 and 2013. Last year, the amount was estimated to have reached roughly four zettabytes the equivalent of about a trillion DVDs and its growth accelerates more every day.

The Man at Microsoft. Mike Neil, general manager of Windows Azure.

That amount of data, properly crunched, offers an evolutionary leap in how we understand the world. Cloud systems provide a cost-effective way to store it, organize it and search it for answers. When youre working with a traditional infrastructure … your mindset starts to become limited by the resources available to you, says Wood. You start constraining the questions you ask. Thats whats really changed here. There are no longer [any] real constraints.

That reality has already paved the way for medicines to move past clinical trials faster, Wood notes, and for a startup like Airbnb, the vacation-rental website, to handle a worldwide logistical operation with only a handful of people.

The specific definition of the cloud is broad, encompassing a wide variety of services, platforms and infrastructure. Yet theres no disputing that many major players in the space are situated in the Puget Sound region, including Amazon Web Services, Microsoft Azure and F5 Networks, as well as major satellite offices of large players like Google and Facebook.

As data streams become oceans, a local ecosystem of companies is wading in to make sense of it. The same chemistry that occurs in Silicon Valley for consumer applications, its happening here for complex systems, says Mike Neil, general manager of Windows Azure, Microsofts cloud offering. That gravity well is forming here. … Its creating and attracting a lot of talent.

Every morning, UPS sends a set of directions to its drivers around the globe, mapping their specific routes for the day. For a delivery company, thats nothing out of the ordinary. UPSs instructions include one interesting feature, however. Drivers are instructed never to take any left turns unless absolutely necessary.

The policy may sound strange, evoking images of UPS trucks moving in circles all day. But by eliminating such issues as idling at traffic lights, the company says its policy saves tens of millions of dollars in fuel costs annually. According to Jim Bak, spokesman for Kirkland-based Inrix, this offbeat but lucrative insight resulted from his companys early data crunching.

Inrix is an example of a company that focuses on one specific issue and tries to tackle it better than anyone else. In this case, its traffic. Half of the commercial fleet in the United States constantly delivers data to Inrix, Bak says, with GPS sensors documenting speed, heading and exact whereabouts. Along with data from road sensors, mobile-app users and partner car companies, Inrix analyzes more than a trillion data points a month, from more than 100 million vehicles around the world.
These data are crosschecked in real time against construction and event schedules, weather predictions and more, forming what are arguably the most predictive traffic models available. Bak says this traffic intelligence has saved Inrix clients huge sums of money and can improve construction planning and incident response times.

As Microsoft and Amazon conducted pioneering work in data analysis over the decades, staff would spin off into separate, data-intensive ventures, often putting down local roots. Inrix, which was founded by former Microsoft employees in 2004, is but one example among the first wave of Big Data firms stockpiling information in mind-boggling amounts, then using it to become experts in tightly defined areas.

Its not about the size of data, says Stan Humphries, chief economist at Seattle-based Zillow Inc., its the questions you can answer with it. Our mission is to take this sea of data and boil it down in a precise, useful way.

Zillow was founded in 2005 and has since amassed data on 110 million residential properties in the United States. Pulling data from public records and partner brokers, the company prides itself on offering the most accurate predictive modeling in real estate. When its IPO hit NASDAQ in 2011, that sort of expertise granted Zillow a $539 million valuation. Its market cap has since risen to $3.6 billion.

However, data analysis is becoming more sophisticated, Humphries observes, and torrents of new information are becoming available all the time. Despite Zillows success so far, the overall industrys potential has only been glimpsed.

Were at a new tier of evolution right now in what we do, Humphries says. Were just starting to mine the data on the 77 million unique users visiting the [Zillow] site each month, for example: what theyre looking at [online] and how they consume it. That will end up dwarfing what weve done with other data. … It can let us really look at the future of areas, to say, In three months, home sales will go up by 5 percent in this neighborhood. The possibilities seem limitless.

An avalanche of companies is catching on to datas value. Some have turned to acquiring specialist firms to stay ahead of the curve, as in agriculture giant Monsantos $1.1 billion purchase of weather predictor The Climate Corporation, which is based in San Francisco and has an office in Seattle.

Even more companies are attempting to become data experts themselves. In a race to prevent competitors from outsmarting them, firms across nearly every sector of the economy are amassing as much information as possible, often before they know what to do with it.

Derek Edwards, CEO of the Seattle data analytics company Globys, remembers when no one could even imagine this. Back when the only cloud was floating in the sky above, Edwards was trying to get mobile companies to share everything they knew about their customers so he might analyze it and improve their business. It took some hand holding.

Fifteen years ago, wed show up to customers and say, Heres what we can do if you give us the six billion transaction records of your customers and let us livestream them, explains Edwards. Their eyes would just gloss over.

Using that data, Edwards says Globys has helped mobile carriers understand what a customer has done and what theyre going to do, what they need and how social behaviors might influence it. That, he adds, can result in some extremely effective marketing.

The first stage of Big Datas evolution was persuading organizations to collect it, Edwards says. Next, it must be compiled via systems like cloud storage or the companys own servers. Then comes the hard part: effectively utilizing it.

Most companies are not doing Big Data right now but claim they are, says James Staten, a principal analyst at independent market research firm Forrester Research. Theyre collecting the data, but theres not much evidence theyre really using it.

Were only seeing 10 percent of the kind of analysis theyll be doing in the future.

The reason for this, Staten says, is that effective analysis remains out of reach for most companies because of shortfalls in both technology and acumen.

The Man at Tableau Software. Francois Ajenstat, Tableaus senior director of product management.

Seattle-based Tableau Software is attempting to bridge some of that gap. Tableau offers tools that create data visualization, allowing users to combine separate streams of data and look for hidden relationships and patterns. Its an approach that Senior Director of Product Management Francois Ajenstat says enables people to think with data in a new way. Data analysis shouldnt just be done by a few priests.

Wall Street certainly sees a strong case for Tableau. When the business went public in May 2013 (with the New York Stock Exchange symbol DATA), it started with a market cap of $2.8 billion. In roughly a year, it has risen to nearly $4.2 billion. Tableau allows users to ask visual questions of data but doesnt do the analytics itself. It therefore does not quench the demand for Big Datas priests, as Ajenstat calls them those people with deep analytical talent who can act as interpreters for all the raw information.

Seattle-based Context Relevant is quickly gaining a name for itself in this area, helping companies tease out insights and predictions from data, using customized algorithms and machine learning.

The push for companies to collect data has become an onslaught, says Context Relevant CEO Stephen Purpura. If youre a Fortune 200 company and not embracing data, youll be dead in 10 years. He points out that banks and other large companies are in an increasingly intense competition for the same limited pool of customers. The fight is over market share and companies that can use data effectively can capture an entire market, says Purpura. Amazon is one of the first examples of that.

Both Tableau Software and Context Relevant herald their approaches as the future of data analysis. Both can rattle off millions in savings or earnings theyve helped clients net, from stockbrokers to Seattle Childrens Hospital. And if theres one thing they and nearly every other company can agree on, its that while data collection is going into overdrive, the supply of people who can decipher it is nowhere near keeping up.

That gives a big advantage to companies like Microsoft and Amazon, which not only sell IT services but are also experienced in collecting and analyzing data about their customers. As the largest online retailer in the world, Amazon collects vast amounts of information about the shopping habits of its nearly 200 million customers, giving it a distinct advantage in knowing what people like and encouraging them to buy more. Some analysts say Amazons new Fire TVwhich allows users to stream online video over their televisionswas driven in part by its desire to know more about what customers are doing when theyre not shopping on Amazon.com.

More data means more potential insights, Microsoft CEO Satya Nadella noted in a recent blog. The opportunity we have in this new world is to find a way of catalyzing this data exhaust from ubiquitous computing and converting it into fuel for ambient intelligence. Microsoft, for example, is working with Accenture to help downtown Seattle buildings cut energy costs by up to 25 percent by mining data on their usage of lighting, heating and ventilation and predicting when to dial up and down the systems. The company is offering tools to make it easier to pull such insights from data.

But both Microsoft and Amazon recognize that better use of data will require more fundamental research. Two years ago, the University of Washington sought to recruit as faculty four of the nations leading minds in data science and machine learning. Competition was fierce from such schools as the Massachusetts Institute of Technology and Stanford, which also offer highly regarded education in data-related fields.

However, the UW had a unique advantage. Local (data and cloud) companies stepped up and offered tremendous help, says Ed Lazowska, director of the UWs eScience Institute, which recently received a major grant to encourage the use of Big Data analysis in university research. Amazon.com provided two $1 million endowed professorships to help recruit two of these stars, and CEO Jeff Bezos met with them personally when they visited Seattle. In the end, all four professors joined the UW staff.

More recently, during the recruitment of another faculty candidate, it was discovered that the spouse of the prospective professor would need to find a job in the legal profession. In the space of an hour, we got [the law firm] Perkins Coie and the chief legal people at Amazon, Microsoft and Tableau [Software] to offer to meet with her, Lazowska says.

Big Data is an industry in its infancy. Creating local expertise, therefore, is integral to cementing a foothold. Innovation and technical expertise are traditional strengths for the area, Lazowska says, but if theyre not cultivated, theyll disappear. The willingness of local companies to help, he says, is what makes us strong.

Every field is transitioning from data poor to data rich right now, Lazowska adds. You can make the argument that the Puget Sound [area] is in the catbird seat, both in the cloud and data analytics. Adds Microsoft Windows Azure General Manager Mike Neil, The same chemistry that occurs in Silicon Valley for consumer applications is happening here for complex systems.

Forrester Researchs Staten says that things that with the industry still in its early stages, and Silicon Valley being home to such powerful players as Google, the Bay Area could still outflank Seattle. But Seattle holds some strategic positions.

Most cite the location of Microsoft and Amazon first and foremost, as well as the ecosystem of data-minded engineers and developers they created. Staten adds an unlikely accomplice: Boeing. The aerospace company has long used data to construct better airplanes, boost assembly line efficiencies and gauge how little foot room passengers would actually tolerate.

Context Relevants Purpura says the mentality of Puget Sound tech workers lends itself to the long-term work it takes to build up data expertise. In the [Silicon] Valley people are expected to change jobs every few months, and they do, he says. Seattle has the engineering talent to build complicated systems. In many ways, its more attractive than anywhere else in world in that. … Engineers here are also more loyal. We can see plenty of things we want to do over the next decade and we want people who will stick with us.

Proximity to the clouds biggest players has value beyond an expanding talent pool. Both Zillow and Inrix use Amazon Web Services in some of their operations, and having them local and being able to meet face to face is certainly helpful for coordination, Zillows Humphries says.

The most compelling case for the region is broader. In some ways, the rise of Big Data mirrors the computer industry itself. Computers once took up entire rooms and were usable only by experts. Now theyre a fraction of the size, easy to use and immensely more powerful. That evolution principally took place in the Bay Area.

But as computing moves to cloud-based systems that can access and analyze data more cost effectively, Seattle is emerging as a powerhouse that can hold its own. Its seeding dozens of players in the cloud and Big Data space and creating an important concentration of talent and enterprise.

From an economists standpoint, capabilities tend to reinforce each other, says Humphries. Working in cloud computing and Big Data takes specialized skills. People who have them are being hired here. Thats important.

Humphries offers up Dalton, Georgia, as an analogy. Not long ago the majority of the global supply of carpets was created there. Thats because, during a pivotal time, the town attracted a critical mass of skilled workers in the field and became the best place to manufacture carpets.

Even today, says Humphries, Anyone who wants to go into the carpet business, they go to Dalton. Thats where they can find people who know how to do it. Skill sets become self-reinforcing and [in Big Data] people here [in Seattle] are skilled in some pretty complex things.