This article describes the progress made in establishing and running the India operations of Tensilica, Inc., a private, venture-funded startup providing microprocessor intellectual property for system-on-chip markets.
Last year, in another article, I discussed the beginning of our offshoring attempt. We had begun by partnering with an Indian design services company to establish an offshore development center under the build-operate-transfer (BOT) model. This article describes the steps taken since then to transfer to our own subsidiary, Tensilica Technologies India, located in Pune.
The article addresses the key reasons to establish our own subsidiary, the important steps needed to turn that into reality, the advantages we’ve seen and the challenges we face today.
Tensilica, a venture-funded startup based in Santa Clara, California, continues to do well in providing high-value products to the embedded systems market. Its Xtensa configurable and extensible processors and newer Diamond pre-configured processor cores are finding use in printers, mobile phones, digital cameras, audio and video applications, routers, modems and a host of other applications. The business and volume of products shipping with Tensilica cores is expanding at a rapid pace.
Tensilica began its India operations in partnership with eInfochips Pvt. Ltd. of Ahmedabad, India. This partnership began in November 2004 when we started an offshore design center based in Pune. The design center was set up under a BOT model, and our experience with offshoring was described in my previous article. This article describes the reasons to transfer and the challenges and opportunities that transfer now provides us.
Recap of Situation Pre-transfer
Let’s recap some points from last year’s article:
- Only R&D work was to be offshored initially (in other words, no customer support, no IT services, etc.)
- Do additional work there; in other words, supplement the work done in Santa Clara, not replace it.
- Do test and QA, but also development, in India
- Fundamental innovations will continue in the United States. We see this as imperatives due to tight coupling with other functions within the company and the unique skills in the existing team.
We faced several challenges in the first seven months of operation.
Building the Team
Recruiting in the high-demand, low-supply market of India was very hard. The BOT model and Tensilica’s lack of a name brand in the Indian market were even greater obstacles.
Management and Startup Effort
The large effort that is required to get an operation 12 hours away is very demanding and often frustrating.
The main challenge was the lack of success in quick startup. We had not, at that time, been able to get contribution and productivity at the pace we had expected. The smoothness of the transfer was also unknown.
Effect on the US Team
There was apprehension about how this new initiatives would work out on the part of the US team. They were unsure about the real productivity and savings from the operation and also about the impact of this initiative on the career growth potential for our US team.
Decision to Transfer
Around mid-2005, our Design Center in Pune had five employees. They were all working on the development and testing of the integrated development environment for our embedded microprocessor technology. We were finding it impossible to hire senior engineers and managers. Without senior, well-experienced engineers there it was very difficult to manage the team from the United States and achieve good productivity.
One of the clear reasons for this inability to hire senior employees was the fact that we were in a BOT model. The more experienced people in the Indian labor market were far more interested in working directly for a US company or a product company. They were not interested in working for an operation that was, in effect, a services company. We interviewed several folks, but they invariably ended with an acknowledgement on their part that while the work and technology that Tensilica was offering was great and interesting, they preferred to wait until we had our own subsidiary.
The BOT model had successfully enabled us to get going at a low cost and provided us good experience working and recruiting in India. This experience gave our management team confidence to continue our efforts in India with an even greater commitment. With this as a background, we decided to first hire our managing director and then work through him to separate from our Indian partner.
Our managing director, a Pune native, came with a background in embedded systems with over 25 years of work experience in various service and product companies in the electronics and software field in India. He was also a key member of a couple of startup companies and has vast knowledge and contacts in the field in India. We recruited him with help from an executive search firm located in the Silicon Valley specializing in recruiting high-level executives in India.
Once our MD was hired, he confirmed and, in fact, accelerated the separation.
Due to the relationship we had with our partner, the clear evidence we could show about the recruiting process and the fact that we had always had open and good communication with them, it was not difficult to separate. We established our subsidiary, Tensilica Technologies India Private Ltd. (TTI), on November 9, 2005, and moved into our own premises on January 1, 2006.
The legal process to set up a subsidiary and to officially separate was not particularly hard or onerous. The hardest part was finding suitable office space in Pune that fit our budget. We were surprised to learn that the space cost more than in Silicon Valley and had a huge deposit (10 months of rent!).
Accomplishments and Challenges Continue
We started 2006 with seven employees and ended with 16 employees and two consultants. The scope of work has expanded from just IDE development to DSP codec development and verification of our instruction extension (TIE) compiler. We have hired several very senior technical experts and managers. Some of these senior hires have advanced degrees from excellent US or Indian universities and substantial experience working in the United States.
The depth of work, productivity, quality and direct contribution to Tensilica’s business success have all increased substantially in the past year. For example, the DSP team in TTI has added to our ability to deliver a larger number and more complex audio codecs to our customers. They have also become the primary contact with our other DSP partners, receiving, accepting and enhancing their deliverables. We have also substantially increased our capacity to do TIE compiler verification and several enhancements and bug fixes have been made in the IDE to make it more usable and robust. This work by TTI has been released to customers and is helping make our product offerings richer, more usable and robust.
To improve the understanding of each other’s culture, work habits and better teamwork, we have organized many trips between Silicon Valley and India. These have used time and employee bandwidth. Despite these efforts, not everything has gone exactly as planned.
Let me address each of the points I mentioned earlier as challenges in 2005 and compare them with where we are today.
Building the Team
As I mentioned, we have clearly succeeded in attracting, hiring and retaining several technical managers, senior and junior engineers. But the ramp has been slow and painful. We have not achieved our goal of growing to 25 employees by year-end. The sourcing of qualified candidates continues to be difficult. There is a dearth of well-qualified engineers in India and the competition for them is intense. That, combined with our insistence on quality over quantity, a somewhat unknown brand (though visibility continues to improve) and location (Pune vs. Bangalore, which is the center of the IC design industry in India. The choice of Pune is serving us very well in terms of retention and stability), but our success rate in bringing on new engineers has not been stellar.
It’s also clear that we have found it easier to hire into the DSP group and the Verification group than into the IDE group. Moreover, our attempt to staff a Customer Engineering and Support organization has so far been totally unsuccessful. The technical skill combination of processor knowledge and GUI experience that we need for the IDE group is almost non-existent in India. Similarly, the CE and Support group candidates with the desire and ability to interact directly with customers at a deep technical level combined with hardware and software skills for the embedded space have been difficult to secure.
Management and Startup Efforts
As the team there has grown and local team managers and leaders are in place, the management effort from the United States has declined. However, in order to continue the close cooperation and joint development work that our technology requires, it is still imperative that we continue to do the following:
Weekly or More Frequent Conference Calls
These are generally at least an hour and often much longer. They are also generally at an inconvenient time for one of the parties. This, of course, is in addition to daily email exchanges.
Nothing substitutes for working in the same location. We continue to visit each other’s sites and spend a couple of weeks to a couple of months there. This helps not only in technical training, but also in imparting our culture and values, that we hope to recreate at TTI, Pune. New managers and technical leaders must visit the United States for an extended period for technical training and to experience and live our culture.
Of course, any time a new group begins in Pune, the startup effort is still quite intense. There is much more involvement of managers and others from the United States to help in getting the new group trained and effective. Due to somewhat better support now available in Pune and due to having a larger group, this startup effort is also easier.
This model wasn’t particularly successful for us in the long run. It did help us get started in India with a very low initial investment. We would have found it almost impossible to convince the management team to start off directly with a subsidiary because of the large investment required to get off the ground, and we had no experience in working in and with India. That risk was too high.
Once the Design Center operation had begun, we had some experience and contacts and were far more confident of building an engineering team within a subsidiary. For this type of operation we wanted to establish — a permanent facility with an excellent engineering team, building products there and complementing the team in the United States — a full-fledged subsidiary is the way to go. It allows one to build one’s own team and instill in it the culture and values that are required for success.
Effect on the US Team
This is a work in progress. There is certainly appreciation on the part of several US team members of the direct benefit the company has due to the India team. The aspect of productivity of the team and total savings is not yet fully obvious to everyone. In order to get there, the overall productivity of the Indian team should increase further. In addition, the Indian team needs to be more confident and proactive on engineering and product issues. I believe this is a matter of time and additional experience with Tensilica technology.
There are now examples where additional responsibility and career growth has occurred in the United States, due to opportunities in working with the Indian team. But there is still sensitivity to growth in India vs. here in the United States.
No Quick Solutions
The establishment of our own subsidiary in India has benefited Tensilica. The four major challenges we had earlier have been addressed to varying degrees. Recruiting continues to be the greatest hurdle, and we don’t see a quick solution on the horizon. It will be some time before India graduates many more well-qualified engineers, and then acquire the level of experience that we require. Even with this hurdle, the subsidiary is bringing business benefits and resource flexibility to the company. There are high expectations for the India operations, and given the people we have hired, I’m sure they’re up to the challenges.
Tensilica and Tensilica Technologies India Private Ltd.
“Build Operate Transfer: Tensilica’s Experience in Offshoring Engineering to India”