Building a software product company in India: A Perspective – Sridhar Vembu
January 12, 2011 § Leave a comment
One of the challenges of being a software product company in India is that professional services companies dominate the IT landscape. This manifests itself in many ways; indeed it is fair to say that the very word ‘software’ in India has come to be associated with providing professional services. This services-oriented industry mindset means that a product company has to cross several hurdles, starting from recruiting people to creating efficient internal processes. What we have learned from over a decade of operating as a ‘different’ kind of company in India is to be patient and pave our own path.
Let’s consider recruitment first: it is hard for a company that is so different from the mainstream to stand out and attract the required talent pool. The prevailing mindset is that a product company is ‘risky’, while a services company has a safe predictable business model. Prospective candidates generally prefer companies that, from their perspective, have proven models and shy away from companies with perceived risky models. This has forced us to be very creative in our recruitment, something that I have written about extensively in our blogs. Indeed, the sheer necessity of finding good people to work for us forced us to explore alternative paths, and our entire culture has changed as a result. For instance, we tend not to focus much on academic credentials; instead we prefer to hire people who can get stuff done in the real world. It is amazing to see how well this principle works in practice, yet how hard it is to convince people, who haven’t experienced it, about this. People seem to be so overawed by fancy credentials that they just don’t bother to look at the data supporting their real world relevance. This is even truer in India than, for example, in the U.S.
Scond, we also noticed that most of college education really doesn’t add value; so we created our own alternative program and hired students directly out of high schools. This has worked significantly well for us. We would likely never have done these if the hiring environment weren’t difficult for us. Even after hiring, until new recruits get familiar with the company and become comfortable about our business, the lingering question is always there in their mind. Every new hire-orientation I attend will have at least a couple of people asking the questions ‘Why don’t we do services? Wouldn’t that be an easier business’. So often have I answered that question that I can now answer that question even in my sleep!
The challenges continue even when employees become experienced. The prevailing social norms in India value a (perceived) stable job in a prestigious company in matters like marriage. I have heard the lamentation (half in jest, often) numerous times from our eligible bachelors that being employed in a new kind of company like ours isn’t great for their marriage prospects – we have even considered going into the match-making business to counter this!
Let’s come to the more serious issue of internal processes. Professional services companies work on fairly rigorous processes that are tuned to their business model of ensuring absolute predictability and certainty in their schedules. Their clients are typically large corporations who value exactly that. But such predictability and certainty come at a cost: the system necessarily has to slow down the pace of development. I have written about this issue in detail elsewhere, but suffice it to say that software development schedules are inherently and intrinsically unpredictable. Any fairly complex software project has exploratory elements (‘cutting the path through an unknown jungle’ is the analogy I often use) and anything exploratory in nature just eludes precise scheduling. The practical way to religiously adhere to a schedule is to slow down the system sufficiently that the probability of a miss is kept low- even then, the schedule miss probability can never be reduced to zero. There are all manners of buffers built into a schedule. In a product company, such a mindset can be wasteful, and even prove fatal. It is telling that while professional services companies tout their process certifications such as CMM Level 5 and so on, product companies world wide do not rely on such certifications, but let their products do the talking. This is not to argue that all process is necessarily bad. A certain amount of discipline in development is mandatory in order to achieve anything productive. But in a product company, there is a real danger of going too far and adopt overly rigid processes. The prevailing services oriented mindset in India makes that danger all too real. A product development environment has to foster a certain amount of creative chaos or controlled anarchy. We have to combine creative freedom while having sufficient discipline so that we actually ship a quality product on a schedule and budget that allow us to be profitable. This balance is a lot harder to strike than it appears. Embracing uncertainty and risk is intrinsic to a product company so, in a sense, people who seek perfectly safe and predictable environments are right to stay away from it. There are numerous rewards to balance that risk, there is nothing more satisfying professionally than to ship a product that gets widespread customer acceptance.
Finally, product companies have another dimension that is usually not a big concern for services companies: marketing. In fact, marketing can be as important in the success of many products as the quality of R&D. The reason is that even the largest services companies do business with a fairly small number of global corporations (typically a few hundred) so that their sales teams can reach them easily. Product companies have to reach a much broader audience, so marketing is crucial. Yet, this is easy to overlook, particularly for people with an engineering background. I certainly have made my share of mistakes in this area, so I can speak from personal experience here. One of the hardest things about marketing is that it is so hard to measure its effectiveness. Internet traffic is easy enough to measure, but how do you measure brand perception or market credibility, which can be very important in understanding whether a product is successful or not? This lack of measurability trips up a person coming from a typical services project management background, where precise measurability is the gospel. Finally, it has taken us over 12 years at Zoho to get to where we are today. I am not very patient by nature, but we had no choice but to be patient as a company, as we figured out a lot of these things over time. We certainly didn’t have an appreciation of all this when we started. Over the years, we have matured as an organization, our products have evolved, and more than just execution of an idea it has been a voyage of adventure.