TopCoder Steps into Void
Originally post in the Hartford Business Journal on April 29, 2013 by Sujata Srinivasan.
Spending cuts are poking a hole in the public sector’s ability to buy goods and services. But they’re also creating big opportunity for the big thinkers at TopCoder Inc.
The Glastonbury technology innovator reports an increase in business volume from government agencies for the use of its digital open innovation platform. In fact, tightening budgets may well be one of the reasons why government organizations are looking outside the traditional technology-sourcing model.
For instance, when NASA wanted to determine an optimal position for its solar panel collectors on the International Space Station, it posted a “TopCoder challenge” on NASA Tournament Lab, an online competitive space created by NASA, Harvard University and TopCoder. The goal was to generate as much power as possible during complex orbital positions so that astronauts had more time to conduct experiments. Announcing the competition, an animated robot makes a moving pitch, saying more time on the Space Station meant more time for science. That’s a clarion call to the TopCoder community, whose members wear their love for STEM — science, technology, engineering, math — on their sleeve.
The competition ran for just two weeks but 500 contestants submitted 2,000 solutions. Winners walked away with prize money totaling $30,000, a fraction of what technology consultants or solar energy experts would have charged.
But what about the quality?
“Off the charts. Statistically, you will see that a diverse group of people always come up with better answers than domain experts,” said Jack Hughes, founder and chairman, TopCoder. “It’s not just about numbers, it’s the reach. It’s really using the Internet to find the right person with the right skillset at the right time. Having said that, people inside the domain are very, very good at formulating the problem. But they are not good at solving them.”
This is the foundation of Top Coder’s business model. It’s based on the wisdom of crowds and the reach of the Internet to find 24/7 talent and solve problems in a record period of time in a cost-effective manner.
TopCoder members move on from one competition to the next, so there is no long-term relationship with the client.
Yet Hughes believes that this is the model for the future. Clients will want to avoid the hassle of locating, hiring and managing external contractors and will receive quality work due to peer evaluation by the community. All work submitted goes through a rigorous scoring mechanism and is reviewed by the client and a panel of judges. Sensitive data is withheld for security purposes, with contestants working on mock numbers.
Founded in 2000, TopCoder is Hughes’ second company. He previously sold Tallan Inc. and invested his savings in the current venture, which operates under the software as a service (SaaS) revenue model. Users pay a fee to access the community and the company earns a percentage of competition prize money. Clients have awarded $50 million in various contests to date.
Hughes declined to comment on the company’s financials, only saying: “We’re a tens of millions of dollars revenue company.”
Members are treated like sports stars and ranked on a baseball-like roster. They talk tech and more tech, review each other’s output and meet up at TopCoder Open, an annual event that Hughes says is March Madness for software developers.
Egor (his TopCoder handle), an algorithmist from Russia and a member since 2005, has earned $19,724 in prize money. But beyond that, the competitions are a matter of pride. “Member of the world’s largest global competitive community,” states Gevak (TopCoder handle), also from Russia, on his statistics page.
Nick Donofrio, a TopCoder board member who is also on the boards of Bank of New York Mellon, Advanced Micro Devices and Liberty Mutual, said the cultural shift brought about by social networking is leading to a change in how companies find answers to their technology problems.
“TopCoder is all about enabling innovation through its open, collaborative, multidisciplinary and global environment. Research suggests this is the enabling environment for real innovation. Issues of Big Data are manifold — volume, velocity, variety and veracity. What better way to understand and appreciate the bona fides of your data than to ask the community who likely created it?” he asks.
Case in point, Harvard Medical School enlisted TopCoder for a Big Data biomedics solution. The task: To accurately and more quickly calculate the edit distance (which determines how close two strings are) between a query DNA string and the original DNA string to aid immunologists in better understanding the immune system.
The best-known existing solution called MegaBLAST processed 100,000 sequences to a high degree of accuracy, but required 2,000 seconds to complete the process. For a full year, Harvard dedicated resources and came up with an improved outcome, reducing the computational time to 400 seconds. At the TopCoder contest, 122 members submitted algorithms and the winning solution offered a high degree of accuracy, while reducing the execution time to just over 16 seconds. The cost to Harvard was $6,000 in prize money.
“A two-week competition led to code that was just as good but almost three orders of magnitude faster for a few thousand dollars. Hard to imagine beating that,” Ramy Arnaout, associate director of clinical microbiology at Beth Israel Deaconess Medical Center, wrote on the TopCoder client site.
Big Data, the industry buzzword, is a key space for the company. Last year, TopCoder hosted a Big Data Challenge on NASA Tournament Lab, inviting the community to come up with ideas for mobile applications that integrate data across multiple U.S. government agencies in health, energy and earth sciences.
Co-sponsored by the U.S. Department of Energy and the National Science Foundation, the contest tasked competitors to imagine tools and techniques that find new value hidden in government information domains.
Competitors were then asked to describe how the solution might be shared across agencies — a huge cost-savings for the public sector.
Big Data is an area that everybody wants to get into. For instance last year, NewVantage Partners (NVP), a Boston-headquartered IT consultancy, surveyed more than 50 technology executives from Fortune 1000 companies and large federal agencies. In Connecticut, NVP interviewed C-level executives and function heads at General Electric, Pitney Bowes, Thomson Reuters, Aetna, Cigna, Conning Asset Management, The Hartford and Travelers, as part of its nationwide survey.
The results showed that the primary reason organizations are investing in Big Data is to improve analytic capabilities and make smarter business decisions.
“Unlike firms like Google, Amazon, Facebook and Yahoo! that were born into the Big Data era, established Fortune 1000 companies have vast troves of information accumulated over years that are maintained on legacy systems that are not integrated,” said Randy Bean, managing partner, NVP. And there lies the big opportunity.
The challenge, said Bean, is that more information is being generated each year than in the previous decades combined. So it’s a big haystack. Giants like IBM, Intel, Oracle, Microsoft and a clutch of smaller and leaner IT firms are offering A-to-Z capabilities for clients.
But Hughes is of the opinion that much of unstructured Big Data “is garbage.” He said it’s being pushed by big vendors and clients are driven toward big storage, big machines and big consulting contracts.
“Fast data is what’s really coming,” he added.
For example, a current TopCoder tech contest is for atrocity prevention — a project commissioned by USAID and Humanity United.
This challenge is to create a model to help identify risk factors in vulnerable parts of the world. Contestants wade through the sea of Big Data in the public domain and come up with ideas and construct probability models that can predict and even prevent crisis in record time.
Ethan Zuckerman, a competition judge and senior researcher at Harvard University, said contests are a good way to ensure that USAID and Humanity United know about which projects on atrocity prevention are currently underway globally and come up with a map of the field.
“TopCoder is a way to get a broader range of people involved in thinking about these issues and, potentially, a way to encounter novel solutions that come from innovative thinking outside the conventional human rights and technology community,” he said.
The biggest challenge of tapping into a global pool of people outside the domain? Ensuring that participants have a good understanding of on-the-ground challenges the contest is addressing.
“My guess is that some of the best contributions might come from people in the open innovation community working with organizations that understand crisis situations on the ground,” Zuckerman said.
To that end, TopCoder’s value may well be in bringing disparate groups together. The challenges are sustaining quality, low-cost and low turn times for projects; ensuring that business intelligence is embedded in the technology aspect; and building long-term relationships with clients.