The following interview is constructed from e-mails and skype calls from February to April 2012 with Arun Kumar of Kerika. See also this video: https://vimeo.com/24502212 and the Kerika blog for more info on Kerika.
Q: Can you talk a little bit about your background
I started off as a Physics major at IIT Delhi, a little too young to make sensible life decisions. After three years, I decided to pack it in and move to the US for a variety of reasons. I ended up at Washington State University where I graduated in Comp. Sci., took some grad courses in Comp. Sci. and business before dropping out of the MBA program to move to NJ to work as a UI developer at AT&T Bell Labs.
After a couple of promotions I found myself a pointy-haired boss at the ripe age of 27, managing a staff of ~40 — all of whom were older than me, which made for an ironic situation since I was supposed to be giving them career advice, so I quit and traveled for a while in India and Europe and then returned to the US to work at a tiny software company in Staten Island that was building state-of-the-art software for handling financial market data in digital form.
I was a manager of all trades: helping to build the business, design products, build teams, etc. I spent 2.5 years completely rebuilding the back-end and front-end systems a German news agency that required a monthly commute to Frankfurt, but that was one of my best projects ever because I had complete business and technology responsibility for a major account. I then spent some time managing the installation of the first digital trading floor in Argentina.
The company was sold to a Japanese conglomerate (CSK) which asked me to help them develop their financial markets business in Asia/Pacific. I spent a year living out of hotels in Tokyo, Hong Kong and Singapore before getting tired of that lifestyle and moving back to the US to become a product manager.
I joined Morgan Stanley in New York in 1995, helping to establish product management processes for their IT department. After a couple of years, I became an E-Business strategist. I moved to London in 1999 to get the European E-business initiatives into a higher gear. While there, the big achievement was putting together a €100MM joint venture with OMX from Stockholm to launch a new pan-European stock exchange (Jiway) that would lower the costs for cross-border trading by retail investors.
I helped create and launch Jiway from scratch in just 9 months. Morgan Stanley asked me to join Jiway after it was founded, to help them safeguard their interest in the JV. Spent most of my time establishing strategic partnerships with the leading banks and brokers in Europe and US to support the new exchange and did a lot of PR (press, radio, TV, conferences, etc. across Western and Eastern Europe).
I decided to move back to the US in 2002 to live in the Seattle area and pursue my own ideas away, from financial services and back in technology.
First attempt to create the Kerika business was 2002-2006. After failing to gain enough traction, I put the company in hibernation and spent a year as a contractor at Microsoft while deciding what to do next. I joined Onvia in Seattle as head of their program management in 2007, eventually ended up managing all their IT and setting up offshore teams, but left in 2010 and decided to restart Kerika, re-imagining and rebuilding the original vision as a Web App.
That’s about it. Side-note: while I was in London, I was asked to join the board of a Swedish company (Polopoly). I remained on the board after moving back to the US, until I resigned in 2005 after concluding the relationship had become too long-distance.
You can see all this at http://www.linkedin.com/in/forArunKumar
Q: Can you talk a little bit about what led you to found your company, what was the problem that motivated you?
There was a moment of epiphany: when I was at Morgan Stanley in the late 90s, I reserved a conference room (at random) for a meeting. Walking into the room I found a large, complex diagram and set of notes on the whiteboard. In one corner, was a reference to one of the products that I was responsible for. I was astonished, because this room was in a part of the company that I had never visited before, and I couldn’t work out the reference to my product. That got me thinking: “what if I could see the world through the eyes of the people who had been in that room before?”
I spent time thinking about this problem, but didn’t do very much. When I moved to London to build Jiway, the question of effective collaboration within distributed teams became an acute problem for me: I was trying to coordinate the efforts of people in New York, London and Stockholm; people working within Morgan Stanley, OMX and external organizations; people from different departments and backgrounds.
The tools we had were from Microsoft: Project, SharePoint, and Outlook (the lowest-common-denominator way of collaborating). I became convinced that there was a class of projects/work that was not supported by traditional approaches to collaboration, and this work was characterized by almost continuous changes in strategy and tactics, practically on a daily basis.
Traditional tools didn’t recognize the fluidity of knowledge-intensive work, where you learn about the project as you execute the project. So, pretty soon you find that instead of MS Project and SharePoint supporting your work, you spend most of your time feeding the beast: trying to keep these “tools” up to date on the shifting reality of your project.
Not handling change gracefully is one flaw, the other is that task management is considered to be the key element of project management, when increasingly it is content management that drives success. And increasingly, that content is sourced externally from the Web rather than generated internally by the team.
While I was working at Onvia, I realized this problem was getting worse because of some secular trends: teams were becoming increasingly distributed, with more work getting done offshore, and managing content, particularly the combination of Web and Office content, was becoming increasingly difficult.
At Onvia we used Project and SharePoint, again, and it was clear that their basic command-and-control model, which presumes that everything can be predetermined at the start of a project with relatively little discovery taking place during the project, were fundamentally unsuited for all sorts of knowledge-intensive work.
Q: How did you get started?
I started in 2002 doing a ton of market research on the problem: I build mockups and showed them to potential users in all sorts of companies and organizations — commercial, nonprofit, governmental — both in the US and Europe. That helped me evolve the basic vision quite a bit. (In retrospect, I did way too much market research and should have started building a usable product sooner). I built a small team and decided to build Kerika as a p2p Java app that used the JXTA open-source networking technology.
We built several versions and got it in the hands of a broad range of users, with a surprising number of them overseas.
BTW, I used my savings to fund the company and never sought external investors. Not sure that was the smartest move I made…
Q: Can you give me a brief overview of where the company is today?
I restarted the company in August 2010. I decided then that the old p2p model was made obsolete by the development of a broad range of Web services, so I decided to build a Web App the second time around. I hooked up with Jon Cohen who acted as CTO and worked on the back-end; on the front-end we ended up cycling through a number of people locally in the Seattle area before finally hooking up with a very talented group of developers in India that are now the core part of the dev team.
We got a first beta version out in March 2011, and started showing it to a few friendly users that were local. We got more people using by word-of-mouth, but we focused primarily on getting beta users in the Seattle area who would let us observe their use of the product and give us direct, extensive feedback. This was enormously helpful: we learned a ton about usability of the product, and we also learned which parts of the product were more useful than others. The product improved hugely in the April-Dec timeframe, to the point where we now view it as very much out of beta. Along the way we picked up users from a broad range of backgrounds: students using it for group projects, other entrepreneurs using it for their startups, people using it for political and advocacy purposes, folks in enterprises…
The basic product is of generic use, but we think the value proposition is strongest when distributed teams have to collaborate, because then the greatest risk of project failure comes from team members not being on the same page, all the time — and not because team members are inherently incompetent or lazy. We are now focusing our marketing focused around this single message: Kerika can eliminate (or at the very least mitigate) the substantial project risk that comes from people not being aligned on strategy, structure and content.
The company is self-funded again (i.e. “me-funded”), and will probably remain that way for a little while longer: I want to see more traction before I seek funding.
Q: What are the two or three things that you have been able to accomplish that you take the most pride in or satisfaction from?
Getting an offshore team working well has been a source of great satisfaction. That’s not easy, and it’s not accidental. I have learned a lot over the years about making cross-border teams work, and getting an offshore team working well takes a mix of cultural, technical and management skills. At this point now I feel I have one of the most productive teams I have ever had, and they are producing output of a higher quality than I could get from local talent.
I don’t know if Kerika is for everyone, but for some people it really does match their style of working. I love the moments when I demo the product to someone who has been looking for a visual collaboration tool for years, without even knowing that they were looking for such a tool, and their face lights up as they immediately “get it”.
Finally, there is the professional joy of being in a startup, where you have to wear so many hats: you have to take on all these different disciplines and jobs that would normally be done by some other department, if you were in a bigger company.
Oh, and getting a patent issued was a pretty big high…
Q: What has been the biggest surprise? What was one key assumption you made, perhaps even unconsciously, that has caused the most grief?
With the first version of Kerika (“K1″), the biggest blunder we made from a technology perspective was relying upon JXTA. It seemed perfect when we first came across it: a beautiful architecture for p2p networking that had been designed by Bill Joy himself, which was now an open source project. However, the project was never really “open” because Sun controlled it too much, and eventually smothered it. JXTA had a number of critical flaws that we assumed would go away over time, because they were all right there on the product roadmap, but the flaws remained and Kerika 1 was inherently unstable as a result. It was really heartbreaking to see people come on board, get all excited about the product, and then walk away in frustration when they realized the communication and collaboration wasn’t reliable.
The big lesson learned there was: if you are going to use an open-source technology, make sure that (a) the technology already has everything that you think you might need, so that if it never improves in the future you are still in good shape, and (b) the product has a vibrant opensource community with a lot of active contributors. Something that is open source may yet prove to be the most expensive decision you will ever make.
With K2, we may have made a strategic miscalculation in assuming the product will be used only by small professional services firms, because we are finding unexpected interest in larger organizations. That’s going to require some changes to our product roadmap.
Q: What development, event, or new understanding since you started has had the most impact on your original plan? How has your plan changed in response?
In Kerika 2, we have a feature that we originally considered to be considered relatively minor: you can place arbitrary text on a project page. The original purpose of this feature was to put labels on pages, but we happened to build it with a rich-text editor. This little feature turned out to have much more appeal than we had every anticipated.
It took me a while to understand what this is all about, but now I realize that formatted blocks of text (containing lists, tables, pictures, etc.) are what I would call “glass-box” documents: you can see what’s inside without opening/launching that document. There’s an 80/20 rule in effect that much of what people need can be better served with glass-box documents than with black-box documents. In other words, our “little” feature is actually more important to most users, more of the time, than our “big” feature of offering seamless integration with Google Docs.
On the business front, we originally planned to market the product exclusively to small professional services firms. But we are getting unsolicited interest from enterprises, and supporting enterprises will require some significant changes in our product roadmap. We are still grappling with this issue.
Q: Do you have some key lessons learned?
- From Kerika 1: don’t spend too much time on market research. After some point, you are not discovering anything new; you are just hearing the same points being rephrased in different ways. Move faster into building your first couple of versions.
- From Kerika 1: be very careful about what open-source products or libraries you incorporate into your own product. If the feature is really important, it’s not free.
- From Kerika 1 & 2: watch users where possible; don’t rely upon them to tell you what they are having difficulties with. People often don’t articulate problems if they think they will look stupid in doing so, and sometimes people don’t even realize what problems they are having. With face-to-face contact and conversation you can find out what people want to achieve, which is often different from what they are complaining about.
- From Kerika 1 & 2: your users will use your product in ways you never considered. That’s a good thing. When you are trying to find your product-market fit, remember that you can’t push on a string: you need to find a use case where someone is pulling on the other end. And if that particular use case wasn’t the one that you envisioned originally, that’s an opportunity not a problem.
- From Kerika 1 & 2: stay true to your vision, which should transcend details of implementation, strategy or tactics. Kerika’s vision is about getting people to understand their project’s current state, structure and trajectory. In terms of path taken to realize the vision, be very flexible: desktop app, Web app, mobile app — these are all different embodiments of the vision, they aren’t necessarily the vision itself.
- From Kerika 1 & 2: your family are fellow-entrepreneurs. If they are not 100% behind your idea of being an entrepreneur, you need to exit before you damage your family.
- From Kerika 2: build your value hypothesis, your discovery hypothesis and your growth hypothesis carefully by identifying all your leap-of-faith assumptions. Cross-check them for coherence. You will be surprised at how the corollaries of your various leap-of-faith assumptions can end up clashing against each other. Once you do that, you will be further surprised to learn that what your startup needs to succeed is not what you read about on TechCrunch, or what you think is “conventional wisdom”.
- From Kerika 1 & 2: a programmer without an understanding of computer science will build you a prototype very quickly, and possibly cheaply, but will not be able to build you a scalable, robust product. Algorithms, data structures, etc. are all still relevant topics, no matter how high-level your scripting language.
- From Kerika 1 & 2: you will almost never fire someone too soon.
- Overall : I have become much more aware of the need to get all the details right. Concepts are great, but execution is what matters, and execution is just plain laborious most of the time. I have come to understand that there are no instant successes: every successful company has a revisionist history that makes its founders look unusually brilliant. I have realized that you can fail by misfortune, but are unlikely to succeed by chance.
Q: What key skill or experience did you lack when you started that has caused you the most problem?
I knew about marketing in general, but not enough about online marketing, or marketing Web Apps or SaaS. And much of what I had learned about marketing tactics was irrelevant when it came to SaaS. That said, a lot of the basic concepts about positioning and competitive analysis are still relevant.
I had a ton of business experience, but very little of that was in sales itself. Where I had done sales before, it was at a business development level: e.g. building strategic partnerships. I really wish I had more sales experience itself: the hard-core stuff about cold calls, follow-ups, etc.
Q: You seem focused on global teams and multi-firm collaboration, what do you see coming?
I think the key phrase is “incessant collaboration”. That’s because any knowledge-intensive work (software, management, marketing, design…) is based more upon discovery during the execution phase of a project, and less on perfect planning at the start of the project.
Our project tools need to embrace change, because resisting change, or treating it always as a danger to your project model, is simply an unrealistic option. Consider SharePoint for example: when you create a new SharePoint page, you have to decide what kind of page it is — a document library, a workspace, etc. Once you decide, you are stuck with that decision. If your project changes radically — and it surely will — e.g. the entire project gets subsumed by some larger project, or some hitherto small piece (sub-project) grows to swallow it’s parent project, you are pretty much screwed: there’s no easy way to reconfigure your SharePoint pages to reflect the new reality. And the reality changes every day, because every day you discover something new: you learn more about a particular requirement, you learn more about your competitors, you learn more about your customer needs, something you thought would be easy turns out to be particularly hard or vice-versa.
But SharePoint doesn’t embrace change: rather, it actively resists it. And it doesn’t support easy management of Web-sourced content, which is increasingly more important than internally-generated content. These days, the smartest guys in the room are the best curators of Web content, not the best writers of original content.
MS Project’s even worse: it’s basic metaphor is for engineering projects, like building a bridge where the science is exact and you really cannot experiment in mid-stream, but it is used for knowledge-intensive work where your goal is not as precise as “build a bridge across this river at this precise location”, but something vaguer like “connect these two systems/organizations/communities in a way that would achieve this other goal, while discovering and dealing with constraints on the problem as you move along a path that you will have to map out on your own”.
Global teams are much harder to synch up because there are fewer overlapping timezones where you can even communicate on Skype (for example) on a regular basis, and sharing a mental model of the project’s structure, strategy and content gets much harder with cultural and linguistic barriers.
That’s our core focus: we actually believe that asynchronous collaboration is far more important than synchronous collaboration. In other words, being able to catch up on my India team’s work when I wake up in the morning is much more important than being able to chat with them in the evening on Skype, because, by definition, my Skype calls will be limited in duration in comparison to the total workday for me and my India teams.
So, if you are trying to synch up when the other person is not online, what are your options? You have (a) text: someone writes a long email, (b) linear organization tools, e.g. use Google Docs and share a document dump, (c) hierarchical organization tools, e.g. try to synch based upon a common taxonomy of collections/folders. Text/email is the worse of all options because the thread is lost very rapidly when there is more than one person involved in the conversation. I can have a very long email thread with you as long as it is just the two of us, but the moment a third person joins the conversation and hits “Reply All”, we are all screwed because there is no telling where that person’s ideas will veer.
Linear organization, which I call the document dump because that’s the basic view offered by docs.google.com, places all the cognitive burden back on the user: here are all the leaves, go ahead and assemble the tree on your own.
Hierarchical organization is better only when the taxonomy is enforced with an iron fist: precise naming conventions, very hard rules about where individual items may be placed, etc. There’s a whole class of document management tools (e.g. Documentum) that are in that business, and a whole class of users, e.g. people working on a new plane at Boeing, for whom this taxonomy works very well. But for a bunch of other people, it places a set of unreasonable constraints on the user community. The hierarchies don’t work because people don’t follow the rules, and people don’t follow the rules because the rules are viewed as peripheral to their core jobs of creating ideas and content.
Kerika believes the visual metaphor offers the least cognitive load, because it is based upon human biology: as babies, we learn to recognize faces and things long before we know their names. When you look at a Kerika page, you can very easily grasp its organization and flow. Many users have shared their private projects with me, so I have seen a wide range of styles in operation. Some people are very hierarchical, and treat Kerika as a way of creating visual folders, others put everything on one big page and rely upon clustering (i.e. spatial organization) to lay out their thoughts. The interesting thing is that it doesn’t matter what style they use: when I see their page I can, in just seconds, understand what their various projects are about. And the visual metaphor then becomes a way of accessing their content. Which means that when I get their content, I get it in a specific, useful, visual context.
When you have multi-firm collaboration, you are dealing with people from different corporate cultures, even if they are all from the same social culture. So you need some very basic, biology-based metaphor that everyone can understand. If plane designers from Boeing are going to collaborate with a PR agency when their plane is ready for launch, you are going to have a massive clash of styles that will be frustrating for all concerned. That’s where visual collaboration helps.
Q: How has your perspective changed on the available of Internet connectivity and the possibility for pervasive connectivity?
When I built Kerika 1, one of the reasons I decided to use p2p networking was that I didn’t believe there would be pervasive Internet connectivity soon enough, and I wanted to support offline usage because I envisioned my core user as someone on the road (or a plane) with a laptop.
With Kerika 2, we are not even trying to implement offline usage, because I believe there will be relatively few instances when you truly will be without Internet connectivity, and when that does happen (e.g. you are trapped in a car in a blizzard in the mountains), getting access to your Kerika projects probably won’t be your biggest problem. It’s like Apple dropping support for floppy disks before any other manufacturer, and then dropping support for CD drives: they are pretty good at skating to where the puck is going to be (although I think they misfired with Firewire).
(I think Google Gears is an example of a solution for a problem that is already disappearing.)
If you have vastly increased communication possibilities, are able to stay online more often (or, at least, whenever you want to), then the signal:noise ratio starts to become a bigger problem. There’s a ton of content available, a ton of new content getting generated, how do you sort the wheat from the chaff? I think new business models that address the signal:noise ratio will become very important in the future. I don’t see enough of that today: every startup is creating its own version of signal and assuming that it isn’t adding to the noise. (Someone could probably say that about Kerika, too ;-) )
Q: Thanks for your time