(This is a live blog of Suhas Kelkar’s talk at the SEAP Breakfast Meet. Suhas talked about his experience of building an incubator at BMC Software.)

Background

• Suhas joined BMC Pune and was given the job of creating an innovation incubator within the company.
• This was the second attempt at creating an incubator in BMC. A previous attempt had failed spectacularly. The previous one had been started with great fanfare, with a 100 people team, and over time, it went down to 80, to 60 to nothing. With this history Suhas started his incubator with zero employees, and minimal fanfare.

On Innovation

• Suhas defines innovation as “Ideas to Cash.” This is important. The focus on cash, i.e. revenues, was an important difference between this incubator and the previous incubator, and also other research labs in companies around the world. Invention for the sake of invention, research for the sake of research is something that they definitely did not want to do. The wanted to ensure that everything they do has a direct or indirect revenue impact upside for BMC.
• There actually exists a document called “The Oslo Manual” which is a set of guidelines for how to do innovation. It is a free PDF that anyone can download, and Suhas recommends that to anyone interested in innovation.
• The Oslo Manual points out that innovation can happen in 4 different areas Product, Process, Marketing and Organization. Suhas adds a 5th category of innovation: “User Experience”

The BMC Incubator

• Why does a product company need an incubator? Product teams get bogged down by tactical improvements for existing customers, and the larger vision (beyond 12-24 months) does not get attention. Startups innovate all the time, and BMC does buy innovative companies, but then integrating them into the company is a huge overhead, and fraught with risks. It would be much more efficient to do innovation in-house if it could be made to work
• The incubator is a separate team who can focus on these issues. It is a small team (about 25 people) compared to the 200 people in just one of the product groups that BMC has. And these 25 people try various different innovative ideas, 9 of 10 of which are bound to fail. But even that failure adds value because that means there are 9 things that the 200 people product team does not have to try out – hence they’re shielded from dead ends and unproductive explorations.
• The mandate for the new incubator (partially based on lessons learnt from the failure of the previous incubator was):
  • Don’t alienate the product teams – you’ll never succeed without their help and blessings
  • Understand the base products thoroughly. Superficial understanding of the issues, toy applications, will not earn the respect of the product team
  • Frequent communication with the business teams
  • File many patents

• The Process

  • The incubator takes inputs from:
    • The office of the CTO, which strategizes and puts together a vision. Before the incubator team, the office of the CTO would hand over the long term vision and strategy to the product team, which was ill equipped to handle it. Now, the incubator fits this gap
    • Product Business Units
    • Customers and Partners
    • Academia (what is the latest in research)
  • The idea backlog is looked at by these three teams:
    • The governance team which meets once in 6 months
    • The alignment team which meets once a quarter
    • The execution team which meets once a month
  • The output of the incubator are:
    • White Papers
    • Prototypes
    • Delivery
    • Patents
    • Innovation Culture
• Challenges for an Incubator
  • How to measure innovation? Number of patents is not a good enough metric.
  • Motivation: the motivation for the incubator and the people on the team must come from within. Creating the motivation, and staying motivated, in the face of 9 failures out of every 10 ideas tried, is difficult.
  • Difficult skill set: the team needs people who are smart, intelligent architects, but also hands on developers, with ability to switch context frequently, understanding the overall BMC vision, ability to sell/market ideas internally, and most importantly they need to be technology as well as business savvy. Finding people like this is a tall order.
• The incubator only does small projects. There are two kinds of projects: “research” and “prototype”.
  • Research projects which are just 1 person 1 week, where that person is supposed to study something for a week and come back with a report.
  • Prototype projects are just 2 or 3 people working for a maximum of 2 months to build a prototype – not necessarily a shipping product. The prototype should prove or disprove some specific hypothesis, and there is a tricky balance to be made in deciding which parts of the prototype will be “real” and which parts will be simple mocked up.

Future Directions

• From technology incubation, they want to move to co-innovation, where they work in conjunction with product teams, and customers to innovate.
• After that they would even like to do business incubation – where the product team is not interested in looking at an adjacent business, in which case the incubator would like to have the ability to go after that market themselves.
• The Indian IT industry, from humble beginnings, is moving up the value chain.
  • First we were doing cost arbitrage (1990s)
  • Now we have process maturity (2000s)
  • The next step would be to get product ownership, and product management here (2010s)
  • Finally, in the 2020s we’ll be able to do innovation, incubation, entrepreneurship
  • The bottomline is that Indian IT industry should be focusing on taking on more and more Product Management responsibilities

Questions from Audience

• Q: The incubator needs people who understand the current products thoroughly. Which means that you need to steal the stars from each product team, because you cannot really hire from outside. And obviously the product team is not willing to give up their stars. How do you solve this problem?
  • A: In general, trying to get stars from the product teams is not possible. You wont get them, and you sour the relationship with the product teams. Instead, what works is to hire the smartest outside people you can hire and then make them learn the product. These people are then teamed up with the right people in the product team during the ‘learning’ process. The learning process is still a bit ad hoc and we haven’t yet formalized it, but at the very least it involves doing some work hands-on.
• Q: What do you answer when a product team asks what is the value you are adding?
  • A: We constantly worry about the value we are adding, and we keep pro-actively stay in touch with the product teams and constantly keep reminding them of the value we add. If it ever happens that a product team asks what value you are bringing, you are already too late
• Q: How are you engaging in academia, and what else would you like to do?
  • A: Currently, we get interns from academia. This allows them to do look at projects that would not get “approved” as regular projects, because “it’s just an intern project.”
• Q: Customers are in the US. Product Managers are in the US. And you cannot innovate unless you understand customers and have close ties with the Product Managers. How do you do that sitting in India?
  • The head of the incubator must be the ultimate product manager, and more. First, s/he must have almost as much understanding of the market and the customers as the product manager of the actual product. In addition s/he must have a vision beyond just what customers want, so that they are able to generate innovative ideas. Successful engagement and understanding of Product Management is key to success of an incubator.
• Q: How do you ensure that the output of an incubator prototype is actually accepted by a product team, and how does the process work
  • All prototype projects require buy-in from the product team and other stakeholders, agreeing tentatively that if the prototype is successful, the product team will actually put that project onto the release schedule. Once the prototype is completed, it is incorporated into the release schedule, and the 2/3 people who worked on the prototype transition into the product team temporarily.