Offshoring to India: A Case Study about Ketera Technologies

Offshoring software development and other work to India often fails or succeeds on the details of implementation — focusing on the right areas of effort, attracting the right kind of employees (and getting them to stay) and showing smarts about setting up operations and dealing with people. In this case study, the author shares the details that enabled his company — Ketera Technologies — to succeed with its offshoring efforts. Ketera started working with India in outsourcing in late 2002 and then transitioned to a wholly owned subsidiary starting in late 2004. This article is based on the author’s experiences as VP of Engineering for the company, which he has since left.

A Background on Ketera Technologies

Ketera helps companies cut purchasing costs, streamline procurement processes, achieve higher performance from suppliers and grow their business without the expense and overhead of traditional software applications.

Ketera provides its software as a service (SaaS, also called “on-demand” and “ASP” for application service provider). The company’s growing customer list is evidence that the days of “behind-the-firewall enterprise software” are coming to an end. Companies like Salesforce.com and NetSuite have educated the market in this model of providing enterprise-class business solutions in the CRM and e-verticals, and Ketera is extending this model to the spend management category.

Ketera is a private company headquartered in Santa Clara, CA, with about 150 employees worldwide. Investors include Kleiner Perkins Caufield & Byers, Foundation Capital, Integral Capital Partners, Emergence Capital Partners and American Express.

Ketera made a strategic decision to utilize India for all functional areas in the company. This was primarily to create a cost-efficient and nimble structure that would allow us to offer our services at a significantly lower price in the market with a considerably faster time to market.

In 2002 and 2003 the company made arrangements with three separate service providers in India to provide software development, client services, customer support and IT support. This was Phase 1 of offshoring, which I cover briefly.

In Apri1 2004 the company decided to create a wholly-owned subsidiary in India and to transition from all “outsource” to mostly “in-house-offshore” Ketera now has a wholly owned subsidiary in Bangalore with about 75 people onboard and still growing. The company continues to outsource a small portion of its work to the existing service providers and new providers for special needs. This is what I call Phase 2.

Phase 1: The Outsourcing Work in India

When I joined Ketera in March 2004, it already had outsourcing arrangements in place. My first task was to evaluate them and recommend improvements.

The functions that were outsourced included:

  • Software development for our core service.
  • Configuration and catalog management for our customers.
  • Customer support.
  • IT support for production hardware.

The total headcount in the outsourced operations peaked at 105 in June 2004. This worked reasonably well, but there were certain issues:

  • The overall productivity of the engineers was less than desired. This assessment was largely qualitative.
  • Most engineers provided by the vendors were junior, which probably contributed to my previously mentioned assessment. We provided a talent matrix, but our vendors took a longer time to recruit senior candidates and succeeded in attracting too few. We concluded that their working environment and salary structure made them less competitive in the market place.
  • We didn’t see any innovation coming from that team. I assessed that this is the nature of a project-oriented (rather than a product-oriented) culture. The project-oriented culture is the norm for most vendors in the business of offshore development.
  • The error rate was high in the client configuration team, causing significant customer satisfaction concerns for Ketera. High error rates also affected our time and costs incurred for product releases.
  • The attrition rate was high, and this led to disruption of the team and the whole product development process.
  • Customer support was of poor quality. Upon switching to another vendor that specialized in customer support, the quality improved significantly.
  • In software development the US staff was also a bottleneck. There were only three managers and one architect for about 80 engineers.
  • None of the Ketera executives had visited the vendors in India! As a result, there were significant communication gaps.

Phase 2: Transition to a New Offshoring Model

We considered the following offshoring models going forward:

Making the current relationships work.

We concluded that the culture of a project or service-oriented company doesn’t lend itself well to product innovation. We even considered creating an architecture team in the United States and utilizing India as an implementation team. But this went against our goal of reducing overall cost of operations.

Alternative Service Providers

We talked to several potential vendors. I found more companies that had “product” thinking and culture compared to what I had seen in the late 1990s when I set up the Informix India Development Center. Technological innovation is crucial to Ketera’s long-term success, and therefore we needed to have direct control of these resources to be able to attract, retain and motivate the team.

Build Operate Transfer (BOT)

A spectrum of BOT models is available now. At one end of the spectrum is the model where a vendor in India can build and own everything, manage the operations and then transfer control to you. On the other end of the spectrum, the vendor in India can provide a legal entity and space for housing the center, provide the network infrastructure, and hardware, while the US company’s subsidiary hires the employees. Vendors are willing to create a dedicated center that looks, feels and functions like your own. On most of this model, there is no upfront investment required, and the overhead can be made completely variable. In our analysis, at a team size of more than 35, it was more economical for us to be on our own. We were planning to grow to 75 as quickly as practical, and hence we didn’t find the BOT model suitable for us.

Wholly-owned Subsidiary

We had the time to set up a strong center because our service providers agreed to continue working with us during that period and cooperate with our team for a smooth transition. Our executive management team believed that this model of establishing a subsidiary was the best strategy for us.

Functions Offshored

Software Development and Client Configuration

While staffing our own center, these functions were transferred from our primary service provider project by project. The first projects we chose were the ones with high attrition, and we asked our partner not to hire replacements. We increased the staff in the United States to four managers as well to create a better US to India ratio (1:10). This transition took nine months.

Operations IT Support

This was still being transferred at the time of my departure from the company.

Customer Support

We continued to outsource this, which, because it’s a 24×7 function, requires a different infrastructure. The company was considering moving it in-house in 2006.

Portions of Product Management

This was a new function. We recruited product managers in India to work closely with development managers in India as well as product managers in US.

Marketing

We also started moving certain marketing support functions.

Telemarketing, Sales Support and Telesales

The company expected to evaluate extending its US teams to India for these functions in 2006 — although these require a special new infrastructure, and it initially expected to outsource it until it achieved critical mass.

Back Office Finance

As our needs in US grew for financial administrators, we planned to add staff in India to augment the resources in US.

City Selection

When I set up the Informix India Development Center in 1995, I did an extensive study of many cities. For Ketera, I visited Ahmadabad, Bangalore, Bombay, Pune, Chennai/Madras, Delhi and Hyderabad. In each city I spoke with at least the following:

  • 2 companies with whom I would consider an outsourcing or BOT arrangement.
  • 2 real estate agents.
  • 1 recruiter.
  • 2 friends or acquaintances working at an executive level.

Based on these discussions I came up with the following city evaluation matrix.

 India City Selection Matrix

 

 Weighting

 Weighting

 MAA

 BLR

 PUN

 DEL

BOM

 HYD

 AMD

 Skills/Availability

 25

 

 

 

 

 

 

 

 

  Programmers & Managers

 

 10

 5

 8

 7

 5

 6

 4

 6

  Product Designers/Architects

 

 10

 5

 8

 7

 6

 7

 4

 5

  Overall maturity

 

 5

 5

 8

 6

 7

 8

 5

 5

 Retention (attrition)

 15

 

 

 

 

 

 

 

 

  Junior

 

 5

 7

 5

 6

 6

 8

 5

 8

  Senior

 

 10

 7

 7

 7

 7

 7

 5

 7

 Infrastructure

 30

 

 

 

 

 

 

 

 

  Supporting eco-system

 

 8

 5

 9

 5

 5

 7

 5

 6

  Cost

 

 8

 7

 6

 8

 6

 5

 8

 8

  Office premises

 

 2

 6

 8

 7

 6

 6

 6

 6

  Power

 

 2

 6

 6

 7

 4

 8

 6

 8

  Telecom/Datacom

 

 5

 7

 7

 7

 6

 8

 6

 7

  International travel

 

 5

 6

 7

 8

 8

 4

 4

 City Attraction

 10

 

 

 

 

 

 

 

 

  Relocation from elsewhere

 

 3

 4

 9

 7

 5

 5

 4

 4

  Quality of life

 

 3

 4

 8

 7

 6

 5

 6

 5

  A?????

 

 1

 6

 7

 8

 7

 6

 6

 6

  Housing

 

 1

 6

 7

 7

 6

 4

 7

 7

  Transport

 

 1

 5

 5

 7

 5

 5

 5

 7

  Climate

 

 1

 4

 9

 8

 5

 5

 4

 4

 Other (specific to situation)

 20

 

 

 

 

 

 

 

 

  Interaction with current vendor

 

 12

 9

 7

 4

 6

 6

 4

 4

  Transfer of staff

 

 8

 8

 6

 4

 3

 4

 4

 4

 

 

  100

 100

 632

 714

 624

 575

 638

 490

 571

 Weighted average

 

 

 6.3

7.1 

 6.2

 5.8

 6.4

 4.9

 5.7

Score from 1 (low) to 10 (High).
High score in every row is green.

The following anecdotes support some of the above ratings:

When I talked to a few recruiters in Bangalore and described Ketera’ s business, they immediately rattled off names of companies they could target. In Chennai, the recruiters asked me, “Which companies should we go after?”

In Pune, I found the availability of infrastructure support shallow. If a company like Veritas or Infosys initiated a large building project, they could easily use up all available resources such as building contractors, carpenters, etc. of high quality.

Generally, I find it challenging that most companies in India say, “Yes,” to almost everything. To my pleasant surprise, I found service providers in Bangalore who could articulate what they were good at and what they were not.

When I visited a company in Hyderabad to evaluate the local talent, their “product” team dialed in from Bangalore to make a presentation!

At a professional network event organized by The Indus Entrepreneurs (TiE) in Bangalore, 15% of the people I met had worked outside India (mostly SiliconValley) for at least two years.

Facilitators/Service Providers

From my Informix experience, I knew that if we weren’t careful, we could end up spending too much management time in administrative work, dealing with Software Technology Park of India (STPI), Reserve Bank of India (RBI), Customs, etc. We chose the strategy of hiring service providers to do anything that wasn’t core to establishing our operation, namely hiring and then managing the right employees and transferring the processes from our service providers. Specifically, except for choosing the right employees and managing them, we wanted to outsource everything. Some of the facilitators and service providers that helped us on the way (or are still helping us) are as follows:

VC Resources

Our lead investor, KPCB, was very helpful in introducing us to a variety of providers in India: investment firms in India, software companies, call centers, legal firms, accounting and administrative services, real estate agents, BOT providers, etc.

Legal

We chose to work with Thakker & Thakker, which took care of all the corporate legal matters in India without requiring a lot of management time.

Finance & Administration

We retained Ujwal Management Services, a Bangalore firm, to take care of all of non-core functions such as accounting, payroll, taxes, administration, site selection, lease negotiation, project management, recruitment coordination and travel management. The Ujwal/Thakker & Thakker duo allowed us to put our management focus on our core business: recruiting the right engineers, training them, and taking over responsibility of our products and other functions.

Executive Recruiter

Global Executive Talent’s founder, Anu Parthasarathy, helped us find the General Manager for the center.

Other Recruiters

We started with several recruiters, but then narrowed it down to a few. One of the recruiters has a full-time employee at our location to speed up coordination of all interviews.

Project Managers

We retained a project manager (design architect and general contractor) to manage the tenant improvement in our facility in Bangalore.

Very importantly, all of these types of facilitators are now available in India, allowing a company like Ketera to concentrate on organizing its operation and hiring the key persons to ensure success.

1st Hire: General Manager

We needed a person who had experience working in a Silicon Valley start-up and in India, so that he or she could create a similar work culture with the appropriate local variations suited for India.

The initial job description emphasized a lot of experience managing engineering teams. However, as we began talking to candidates, it became clear that we needed someone to create a positive buzz about Ketera in Bangalore for us to attract topnotch candidates. We finally chose a GM with considerable marketing experience who has been able to use his marketing skills effectively in making Ketera known in the Bangalore software development community. High-caliber candidates now began approaching Ketera.

Temporary and Permanent Site

We found a temporary working place immediately for about 20 engineers through personal contacts. Business process outsourcing service provider e4e was kind enough to sub-lease space to us for three to four months.

Within two months we found and signed a lease for a permanent place. Our primary criterion was the ability to create a Class A working environment.

Bangalore has been going through a building boom. However most of the high quality buildings are less than three years old, and therefore there is almost no inventory of ready-to-use space that someone else is vacating. Thus, we took a “cold-shell” space and managed the full interior work. Most builders are willing to do the tenant improvements for a bigger rent and deposit. In our analysis the breakeven point for recovering our up-front investment was about a year. In order to preserve our capital, in cases where the recovery period was longer than two years, we chose to rent furniture and some equipment.

We managed the project in two parts. The first part, with a capacity of 48 seats, was finished in six weeks. The second part, with a capacity of 62 seats, was completed in another three weeks. The center was officially launched on December 8, 2004.

Recruiting Challenge

Recruiting will always be the primary challenge in India. We were told that it would be very difficult for a small, “no-name” company to compete against the Indian and other international giants in Bangalore. Most people advised us not to establish our facility in Bangalore. However, we operated on the following logic. Even in Silicon Valley, there are people who will only work for a large public company and others who will only work for early stage start ups. If the pool is large enough, there are enough people who want to work in a start-up environment. In Bangalore, there’s also a growing group of engineers who want to work in a “product” environment rather than a “service” environment. With this thinking, we took the following actions that resulted in success in recruiting:

  • We used KPCB and then-partner Vinod Khosla’s name extensively (he has since gone on to found Khosla Ventures). We had been told that Silicon Valley VCs were not well known. What we found instead was that our best candidates were quite familiar with the elite Silicon Valley venture capital firms — and this turned into one of our filtering-in criteria.
  • We advertised a motorbike give-away for candidates applying through our website. This was a front page story in the local newspapers. Very low cost publicity!
  • We sponsored and organized public events with organizations such as TiE to make ourselves visible. During our launch, we organized a panel discussion at TiE with prominent industry names (Venkat Panchapakesan, CEO of Yahoo! India; Rajesh Jain, Managing Director, Netcore Solutions; and Steve Savignano, Chairman, Ketera Technologies). Although you could argue that our candidate pool doesn’t show up at such events, their advisors do. It is my observation that in India every one gets lot of advice — solicited and unsolicited.
  • We used recruiters to acquire the initial set of senior engineers. Once we had the critical number, employee referrals became more effective.
  • We actively encouraged employee referrals. We provided a small reward to every employee who brought a qualified lead. We gave a much larger reward if the candidate was hired and further bonuses for multiple candidates.
  • We streamlined our recruiting process. Once someone was invited for an interview, our rule was to complete the interview process in two days and have an offer made the next day, including approval from our US-based CEO. I had learned from my Informix days that making an offer quickly had a dramatically positive effect on acceptance rates; we achieved a more than a 66% acceptance rate.
  • We followed up diligently after an offer was made. In Bangalore today there’s a high rate of “no shows” — candidates who don’t join the company even after signing an acceptance letter. Our dropout rate is now 1 in 3. The industry average seems to be 1 in 2.
  • We tracked and graphed our recruiting and the conversion rates every step of the way to optimize the process.
  • Finally, we were also lucky. A few companies in related domains were downsizing, and we picked up several good candidates from there. We also took in a few key individuals returning form Silicon Valley to India or who had returned recently and were looking for a product company like ours.

We were able to hire 8 to 10 high quality individuals per month and grow the center to 75 in about nine months of active hiring.

Operating Challenges

Communication, Communication, and Communication

The primary challenge for the center became the lack of informal communication we experienced in our Silicon Valley operation. We missed the informal hallway and coffee station side chats. We missed going to a whiteboard and brainstorming an idea. To counter those obstacles, we took the following measures:

  • We recruited product mangers In India early on. This facilitated efficient communication with engineers in India. The US product managers now focus on market requirements and those in India work on the details.
  • We had an executive address the India-all-hands every other week.
  • There was at least one executive from the US traveling to India every other month.
  • We pushed the use of intranet and email discussion forums diligently across the company.
  • The hallway discussion and whiteboard brainstorming is now happening — but in India.

Attrition

It’s no secret: Attrition of staff is high in India. Some companies have reported attrition of about 35% a year in Bangalore. During my work with Ketera in India, our attrition rate was low — but it had only been about 20 months since I hired my first employee in India, and the majority of the staff had been with us for less than a year. We determined that the following would keep attrition low:

  • Challenging work.
  • A great working environment; a place where it’s fun to work.
  • Open and frequent communications with all levels of management.
  • Fair treatment.
  • Opportunity for US travel.
  • Continuing improvement of technical and soft skills.

The attrition, however, will never be zero. We simply needed to incorporate that in our plan.

Changes in Work Style for US Employees

Most of us in the United States found ourselves on the phone or in chat sessions with India for two to three hours three to four nights per week. Our leads and managers in India also ended up spending similar hours a night talking to someone in the United States. This changed our familial interactions significantly. Although I encouraged my staff to go home early, they all seemed to spend their regular hours in the offices as well. Would this cause burnout or is this just a growing pain?

Measurements

Setting metrics for performance in a software development environment is extremely difficult. What can you measure? Lines of codes per person? Number of bugs in a