Tasktop Finds Path to Profits, Via a More Efficient Interface Inspired by Brain Science

For Mik Kersten, it all started when he saw Maria Klawe speak at the University of British Columbia. It was the mid-1990s, and Klawe, a distinguished mathematician and computer scientist—now the president of Harvey Mudd College and recently appointed to Microsoft’s board of directors—was giving a lecture to students and faculty. “She talked about her hippie days traveling in India, and it convinced me to switch to computer science,” Kersten says.

Kersten was an undergrad at UBC studying anthropology. Today, he is the co-founder and CEO of Tasktop Technologies, a Vancouver, BC-based startup that is working to reinvent user interfaces for software developers and other knowledge workers so they can be much more productive. It is one of those quiet Northwest success stories you probably haven’t heard much about yet, but you will—Tasktop is profitable, and has recently signed a number of important deals with the likes of IBM and Microsoft.

The company’s basic idea is to organize work around tasks, instead of files, folders, or Web pages. Kersten’s “task-focused interface” builds tools and information around the specific task you are trying to accomplish—writing code to import digital media into a library, say, or analyzing trends in a database. Tasktop’s software automatically gathers screenshots, notes, e-mails, and other information related to the task at hand and puts it on your desktop in a single handy spot for reference. If you come back to the task an hour later, or a week later, your desktop is returned to where you left off.

It’s a far cry from the way most people work on tasks today, using tools that are glorified Windows Explorer or Mac Finder applications, or Outlook or Google search tools that make you scroll through tons of results, Kersten says. As a software engineer himself, he had felt quite a bit of personal pain. “I was getting bad RSI [repetitive strain injury] in my forearms,” he says. “I was spending more time looking for the information I needed to write code than actual coding.”

Kersten’s early career path took him to Palo Alto Research Center (formerly Xerox PARC) in Silicon Valley, where he worked on user interfaces until 2003. There, he was exposed to a technology called “degree of interest trees.” This is a type of interface that lets you navigate large, branching structures of information. The amount of detail displayed is based on your level of interest in each item, so you don’t get swamped with lots of information about low-priority matters. As Kersten explains, this “makes it easier for programmers to work with very complex systems”—like having to refer to millions of lines of code, or search through 100,000 files. “Programmers get completely overloaded with information,” he says. “It’s extremely difficult to find what they’re looking for.”

After a six-month stint at Bellevue, WA-based Intentional Software (billionaire Charles Simonyi’s company), Kersten decided to quit industry to do fundamental research on how to improve user interfaces. He went back to graduate school at UBC to do his Ph.D. with Gail Murphy, a professor of computer science and an expert on software development tools. “We have to rethink the way tools work,” Kersten says of his ideas at the time.

So he and Murphy did a series of studies on how software engineers get work done, and how different kinds of interfaces affect their productivity. “We experimented with realigning tools around the way developers work,” Kersten says. “We made tasks the primary unit of interaction, and then made all things related to that task show up automatically…Everything you navigate becomes part of that task.”

This focus on tasks, rather than on files or Web pages, had a big effect on programmers at IBM, where Kersten did his tests. His conclusion was based on comparing how much the workers edit, versus how much they click and search. Kersten made his task-based tool available as open source, and in 2005, about 100 people signed up to try it and report their results. Kersten was thrilled to see a statistically significant increase in their productivity when they used the new tool, which he reported in a series of research papers. “I figured this was the fastest way of getting my Ph.D.,” he says.

Kersten finished his degree at the end of 2006. The day after he got his thesis signed off, in January 2007, he and Gail Murphy founded Tasktop. The company had paying customers right away, and has been able to bootstrap on its revenues without taking outside funding. It sells its software and services to small developers and enterprise customers (mostly software companies). Kersten says the firm has always had positive cash flow and has been profitable from its second year. In the meantime, it has grown to a dozen employees plus contractors.

The next step, Kersten says, is to bring the task-focused interface to big software companies—and then to a broader customer base of project managers and other knowledge workers, by integrating the tools with Gmail and Outlook. “We’ve met our first goal, to integrate all the various sources of tasks and documents for developers,” he says. “The eventual goal is to become the task-focused desktop.”

That will take Tasktop into a more crowded space, as there are lots of companies working on ways to improve project management, collaborative software, desktop search, and e-mail management. Some Seattle-area companies in these areas include Smartsheet, Gist, and LiquidPlanner; and then there are big-company efforts to unify Web-based communications, like Google Wave. But Kersten maintains that his company is truly differentiated because of how it layers its software on top of whatever tasks customers need to perform. “We’re the glue,” he says. “We don’t have a direct competitor on the task-focused side.”

And of course, that is still Tasktop’s main thrust—and its key selling point. Kersten says that software developers who use the company’s open-source tool, called Eclipse Mylyn, are typically twice as productive as before. There is a learning curve to using the interface, he admits, but eventually you get “one click multitasking.” Instead of having to open and close windows and files every time you switch tasks (which Kersten says developers do every 11 minutes), the Tasktop tool keeps track of where you are in each task and shows you just what you need.

There is some pretty strong neuroscience behind all this. Dig a little deeper, and Kersten will tell you about the difference between semantic memory (facts and categories) and episodic memory (times, places, context). “For all knowledge workers, semantic memory is completely overloaded,” he says. “What we [Tasktop] do is use episodic memory. We weigh everything by how often you access it. What you see on-screen represents your memory of the task. We make the user interface line up with how your brain works.”

When I talked with Kersten a few weeks ago, it sounded as though big companies like Microsoft were starting to wake up to what Tasktop is doing. Indeed, just last week, Tasktop announced it has formed a partnership with Microsoft to make Eclipse Mylyn work with Windows 7. And Tasktop has had an ongoing partnership with IBM’s Rational Software division, to help improve business management and software engineering for companies.

Besides having been originally inspired by Microsoft board member Maria Klawe, Kersten has another Seattle connection. One of Tasktop’s board members is Neelan Choksi, the co-founder of Lexcycle (maker of Stanza, the e-book iPhone app), who now works at Amazon. Kersten met Choksi at a conference in India a couple years ago, and Choksi’s company at the time, SpringSource, bought Tasktop’s product. Choksi then joined Tasktop’s board in the summer of 2008 and has played an important advisory role ever since, Kersten says.

It will be interesting to see whether Tasktop really has a product for the masses—helping not just software developers to boost their productivity, but most everyone who taps at a keyboard much of the day. The Tasktop tool for developers gets about a million downloads per month, and Kersten says the task-focused interface problem is “mostly solved” for developers. “We’re figuring out how to apply it to the mainstream market,” he says. “We’re working on it, and working with partners.”

Gregory T. Huang is Xconomy's Editor in chief. E-mail him at gthuang [at] xconomy.com. Follow @gthuang

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

Comments are closed.