Case Study

SMU805

SEMICON INDIA: DEMYSTIFYING WORKFORCE ANALYTICS (A)

It was May 2015 and Filip Daw, a human capital strategist with a leading consulting firm in India,

had just wrapped up a meeting with the C-suite leaders of Semicon India (SI), a subsidiary of

Semicon Inc., a global leader in the semiconductor industry. The American parent had its operations

spread across more than 70 locations in 19 countries, and supplied microcontrollers and embedded

processors to automotive, networking, industrial and consumer product industries.

The semiconductor industry comprised three key activities - design, fabrication and assembly - in the

development and manufacturing of semiconductor chips. Of these, design was the most skill-

intensive activity and determined a company’s time to market. Thus, the design centres constituted a core strategic unit of Semicon, and their productivity and turnaround capabilities directly impacted

its competitiveness as a supplier to customers in various end-user sectors.

In 2005, SI was established as a design centre, with an objective to leverage the low cost and easy

availability of talent in India. However, contrary to expectations, it suffered from high workforce

attrition that led to a decline in its productivity and not just lowered its competitive position in the

customers’ vendor selection process but also impeded it from responding to customers’ requests in time. Daw was engaged to help understand the reasons for its high attrition and provide insights to

address the issue.

SI’s Human Resource (HR) department held several opinions as to reasons for the attrition. But Daw

knew that as a strategist, his insights must be driven by data rather than opinions. Identification and

analysis of the appropriate data would not only help determine the validity of the opinions held by

HR, but also unearth reasons that were beyond the obvious. In addition, what was challenging at that

point in time was that the design function, unlike sales, had no direct measure of productivity. Daw

therefore wondered if there was a metric that could be used as a proxy to measure productivity and

track the impact of attrition.

Industry Overview

Market Size

In 2015, the global revenue of the semiconductor industry was US$342.7 billion, and was projected

This case was written by Dr Fermin Diez, Professor Tan Hwee Hoon and Lakshmi Appasamy at the Singapore Management University. The case was prepared solely to provide material for class discussion. The authors do not intend to illustrate either effective or ineffective handling of a managerial situation. The authors may have disguised certain names and other identifying information to protect confidentiality. The case is based on the article “Using an Analytical Approach to Increase Retention in High-Growth Countries: An Example in India”, by Fermin Diez, published in WorldatWork Journal, Second Quarter 2009, and has been developed with permission from WorldatWork [Contents © 2009. Reprinted with permission from WorldatWork. Content is licensed for use by purchaser only. No part of this article may be reproduced, excerpted or redistributed in any form without express written permission from WorldatWork. WorldatWork website: ww.worldatwork.org]

Copyright © 2020, Singapore Management University Version: 2020-05-18

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

to reach more than US$655.6 billion by 2025.1 The Asia Pacific region accounted for over 50% of

the worldwide semiconductor market. India, at US$14.5 billion in 2015, was estimated to grow at a

compound annual growth rate (CAGR) of 29.4% to reach US$52.6 billion by 2020.2

The prospects of semiconductor and electronics sectors were intertwined, as the growth in one fuelled

the growth in the other. Increase in consumption of electronics, primarily mobile computing and

mobile telephony, drove up the demand for semiconductors. Indian electronic devices consumption

was on the rise, and was expected to grow at a CAGR of 41% to reach US$400 billion by 2020 from

US$75 billion in 2015.3

Furthermore, with the increased sophistication of semiconductor chip design, new possibilities had

opened up in unconventional areas such as nanoscience, remote sensing, biotechnology and

photonics; such advancements led to an increased semiconductor content in electronic devices.

Background of the Semiconductor Industry4

Production of the integrated chip, the most critical type of semiconductor, comprised three stages – pre-fabrication (design), fabrication (fab), and test and assembly. In the design phase, the circuit

boards had a series of abstract representations of details of transistors that would be interconnected

by a circuit of wires. During fabrication, the circuits on the chip, detailed in the design, were built on

the silicon wafer in successive layers, while the assembly phase involved cutting the wafer into

individual chips and packaging them in a protective shell. The design phase was skill-intensive and

required investments in expensive Electronic Design Automation (EDA) software. The fabrication

phase was capital-intensive but the worker skill requirement was lower than the design phase. While

the assembly phase was also capital intensive, it was not as high as the fabrication phase, and its

worker skill requirement too was relatively lower.

Prior to globalisation, the semiconductor industry in the developed countries of the West was

vertically integrated, as the companies owned and operated facilities for all stages of semiconductor

production. However, as the assembly phase involved less-skilled labour, offshoring it to low-cost

geographies presented an easy and attractive alternative, and in the 1960s, offshoring started. Soon

after, in the 1970s, the fabrication phase too was offshored, and finally in the 1980s, the design phase,

which required access to highly-skilled labour such as design engineers, began to be offshored as

well.

The merchant chip producers who sold chips to end-use manufacturers were the first to offshore

production, while the captive producers who produced the chips for use in their own products started

later. Eventually, with technology diffusion and foreign investment stimulating the emergence of

local companies in Asia, offshoring gave way to outsourcing. Outsourcing resulted in reduced

investments in capacity and reduced risks of overcapacity while increasing access to a variety of chip

types for the Western semiconductor companies. The industry thus consisted of three types of players

– Integrated Device Manufacturers (IDM), fabless companies and foundries. The IDMs manufactured and sold the chips; fabless companies designed and sold the chips but outsourced

1 Handel Jones, “Whitepaper: Semiconductor Industry from 2015 to 2025”, SEMI, 4 August 2015, https://www.semi.org/en/semiconductor-industry-2015-2025, accessed December 2019. 2 “Electronics”, Indian Brand Equity Foundation, 6 June 2017, https://www.ibef.org/download/Electronics-June-2017.pdf, accessed

December 2019. 3 Jaideep Shenoy, “Indian Electronics Market Expected to Reach $400 Billion by 2020: Study”, Times of India, 12 June 2017, https://timesofindia.indiatimes.com/business/india-business/indian-electronics-market-expected-to-reach-400-billion-by-2020-

study/articleshow/59108734.cms, accessed in December 2019. 4 Clair Brown, Greg Linden and Jeffrey Macher, “Offshoring in the Semiconductor Industry: A Historical Perspective”, Brookings Trade Forum, 2005, p. 279–333. JSTOR, www.jstor.org/stable/25058769, accessed in December 2019.

2/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

manufacturing to foundries, and foundries manufactured the chips that were designed and sold by

others.

Design Service5

The increased outsourcing of assembly and fabrication had resulted in ‘fabless’ model firms that offered design services on contract. On the other hand, the captive manufacturers and IDMs operated

their own exclusive offshore design centres in order to avoid IP exposure and to ensure performance

and reliability. Since designing required high skills, more engineers were engaged at this stage than

any other. Thus, offshoring and outsourcing was concentrated in Asian economies that had a

substantial pool of engineering talent. The complexity of the project, turnaround time and availability

of resources determined the size of the design teams in a company.

Designing involved three hierarchical stages – specification (which defined the behaviour, features and functions of the chip), logic design (an abstract representation of signal processing within the

chip) and physical design (translation of the abstract into an actual network of wires and devices on

silicon layers). A prototype thus developed was tested in a hardware simulation of the system in

which the chip would be used. The software that would get embedded in the chip was developed

alongside. EDA software was used to realise the design and to automate the later stages of designing

that required lower engineering input. As demand for the sophistication of performance, reliability,

features and functions of products in which the chips were to be embedded increased, a larger number

of transistors had to be packed on each line width of the chip. With the increase in complexity of the

chips, and the fact that software had to be co-developed for better integration with end systems, the

engineering hours went up.

Companies established offshore design centres to gain access to specialised skilled labour, establish

closer contact with customers, and reduce cost. While the centres brought significant cost savings,

problems such as communication and coordination issues between offshore teams and parent

company also increased. This was in addition to the managerial challenges of assigning projects,

assessing team productivity and individual performance. The challenges were aggravated with

increased training and monitoring costs due to lower quality and productivity of inexperienced

engineers. Despite these challenges, offshoring of design centres had increased driven by the cost

advantages and proximity to production facilities and high-growth markets. The dynamic nature of

the electronics industry demanded new models and improved variants at a fast pace. Original

Equipment Manufacturers (OEMs) and product companies not only faced declining margins, they

also were under pressure to shorten product development/innovation cycles. Hence, they had started

outsourcing design services to low-cost centres in Asia. The contract design service centres that

designed the entire chip or undertook specific functions of design projects, such as physical design

and logic verification, for captive manufacturers and system companies were growing, especially in

Japan, Taiwan, China and India.

India6

The growth of the semiconductor design service industry in India was driven by the mature software

sector, emergence of the local market for electronic devices, availability of an English-speaking

skilled workforce, particularly, ample supply of electronics engineering graduates, and an influx of

foreign investment. International companies had either set up their subsidiaries that functioned as

5 Ibid. 6 India Brand Equity Foundation, “Semiconductor: Market and Opportunities”, https://www.ibef.org/download/Semiconductors_220708.pdf, accessed in December 2019.

3/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

offshore design centres in the country or outsourced their design requirements to Indian companies.

There were 125 design companies in India, and of the 25 world’s top semiconductor companies, 19

had already established subsidiaries in India. Large Indian IT companies such as Wipro, Tata

Consultancy and HCL Technologies had also embarked on providing design services. This was in

addition to the emergence of many start-ups in the sector with the high penetration of technology and

increase in foreign investments. Furthermore, in February 2007, the Indian government had rolled

out a host of incentives for the industry in its first-ever exclusive policy for sharing of up to 20%

capital expenses in the case of companies in Special Economic Zones (SEZ) and 25% for those in

non-SEZ, and up to 26% equity participation. The policy was anticipated to attract more players and

investments into India.

Most design companies in India were also moving up the value chain. They had started to produce

high-value front-end design in addition to the design centres of multinational firms undertaking end-

to-end design development of entire chips. This had led to a need for highly-specialised skills across

a wide range of design-related activities which the local engineers lacked. Thus, they faced steep

learning curves, resulting in delayed projects.

The Indian design services companies billed their clients in one of three methods – time and material, fixed price or joint go-to-market. In the time and material method, billing was based on the actuals,

and in some cases was subjected to a ‘not-to-exceed’ clause. The fixed price method involved fixing the deliverables by calculating the engineer-months required to complete the project. In the joint go-

to-market method, the captive manufacturers engaged the third-party service providers for design,

and paid for it in the form of royalties from the sale of end products. Thus, productivity was critical

to ensure sustainable revenues and any delays in executing the projects would severely impede the

earnings of the design service companies.

In 2015, foreign companies accounted for a majority of India’s US$12.9 billion semiconductor

design industry's revenue. The design service industry was forecast to reach US$23-29 billion by

2020.7 A 2011 study by the Indian Semiconductor Association had estimated that the industry would

employ 230,000 people.8. The workforce comprised primarily graduates, of whom close to 60% had

only up to eight years of experience. Women accounted for 20% of the workforce; the embedded

software segment had a higher ratio of women than other segments.

Both public and private initiatives were underway to help develop the designing skills of the local

engineers as the more than 200,000 relevant engineering talent that the country produced annually

lacked the quality and level of skills required by the industry, with only 6% of this talent coming

from tier-one institutes of the country. The majority of the fresh relevant engineering graduates were

from the southern part of the country, as most of the universities offering relevant degrees remained

concentrated in the South. Notably, many of the experienced engineers, after a few years of working

and strengthening their basic skills, opted to enrol in higher education and specialisation courses as

it enhanced their domestic and overseas career prospects.

Commenting on the talent scenario, Daw said,

Despite all the efforts in the past few years to address the skills gap, access to high-quality

engineers remains a challenge, and experienced engineers are in high demand. In order to retain

7 “Indian Electronics & Semiconductor Association, Indian ESDM Industry Update: An IESA-EY Report”, February 2017,

https://meity.gov.in/writereaddata/files/Executive%20Summary%20of%20the%20IESA%20EY%20Report.pdf, accessed December

2019. 8 Indian Semiconductor Industry Association, “Study on Semiconductor Design, Embedded Software and Services Industry”, April 2011, https://meity.gov.in/writereaddata/files/Semiconductor06April11_020511.pdf, accessed in December 2019.

4/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

talent amidst high competition, pay-based solutions are deployed by the players in the industry,

and senior management appears to think that SI should adopt the same strategies, although from

my perspective, these strategies are not sustainable in the long run.

Semicon India

Company Overview

In 2005, Semicon had established operations in India with the aim of developing a Centre of

Excellence (CoE) in System on Chip (SoC) integration and IP core 9 design. The American parent,

Semicon Inc, with a revenue of US$26 billion in 2014, was one of the world’s largest semiconductor companies. Its research and development, manufacturing and sales operations were spread across the

globe. It designed and manufactured embedded semiconductors for the automotive, consumer,

industrial, networking and wireless markets. SI operated two design facilities – one at its headquarters at Noida (adjoining India’s capital, New Delhi) and the second at Bangalore, and developed software and hardware designs for the wireless, networking and automotive markets. By 2015, the company

had expanded extensively, and in the past two years alone, it had invested over US$50 million and

filed over 30 patents.

Workforce Overview

Design projects were undertaken by composite teams that comprised members with diverse skills,

and each project, depending on its complexity, took from a few months up to a year to complete. By

2015, the Noida facility, a R&D centre, employed around 1,000 engineers who worked on advanced

technologies to develop digital and mixed-signal IPs and SoCs. The Bangalore facility, a software

centre, employed about 600 engineers who undertook the development of software platforms,

connectivity blocks, protocol stacks, OS porting and validation. SI planned to expand its operations

across the country, and to that end, the company needed to increase its design team by another 1,500

engineers over the next two years.

The skill level of entry-level engineers was diverse as they came from educational institutions of

varied standing. SI believed that compared to recruits from tier-two institutes, recruits from tier-one

institutes had the requisite soft skills along with better quality design skills. As an equal opportunity

employer, SI had a relatively substantial share of women in its workforce. India had a significant

number of women graduating each year as IT engineers; hence it was easy to recruit women into the

embedded software designing segment. However, the HR team believed that the attrition rate among

women was generally higher than men. Notably, the Bangalore facility, the software centre, had more

women than the Noida facility.

Workforce Management

Due to the extreme shortage of talent in the industry, there was stiff competition among companies

to attract and recruit talent. Poaching was rampant and the management and the HR team at SI

believed that most of its turnover was caused by higher pay offered by the competitors to lure its

experienced engineers. The demand for talent was not limited within the country but was widespread

in the region that was buzzing with outsourced and offshored semiconductor design and

9 A reusable unit of logic, cell or integrated chip design for which the design developer owns the intellectual property rights, which can

be licensed to other parties. IP cores are generally used as building blocks of larger software or hardware designs.

5/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

manufacturing activities. The HR team at SI strived to keep its median pay in line with that of the

industry's median by closely watching salary movements in the market.

Appraisal of employees was based on forced curve performance management philosophy, and

accordingly the employees were rated on a 1 to 5 scale by their managers (refer to Exhibit 1 for SI’s Performance Rating Scale). It was a practice at SI to conduct a survey among its employees to score

their team managers based on a set of leadership competencies. The score of each manager reflected

their competence as measured by the employees in their team.

In high growth markets such as India, there prevailed a dilemma over the frequency of promotions – whether to promote employees who were not ready for higher responsibilities and then train them

along the way, or to wait until the employees were ready for promotion against the risk of losing

them to competition. SI had a career plan in place for its employees. The career plan established the

number of years between promotions. In addition, the company was required to promote no less than

15% of its employees at each level to maintain healthy levels of ‘build versus buy’ talent policies. As for pay progression within and across levels, it observed the market pricing and adjusted its

midpoints.

Trial of Turnover

Historical High

Despite intense competition for design talent in the Indian market, SI had managed to retain its

workforce turnover rate at par with the industry average. According to its HR team, employees who

stayed for at least three years tended to remain with the company for a long time. SI primarily

deployed pay-related solutions to retain its employees by benchmarking competitors and matching

retention packages with industry trends. However, despite keeping up with the market averages,

turnover was high in some segments of the individual contributor (IC) population – the entry-level analysts IC-1, and the next two levels - IC-2 (engineers) and IC-3 (senior engineers). At the IC-2

level, it was losing nearly as many people as it was hiring at entry-level (IC1), and it was hiring

nearly as many people at IC-3 as it was losing at the IC-2 level.

In 2015, SI’s turnover rate exceeded its historical average of 20%, alarming the management. If left unchecked, the high turnover rate had the potential of impacting productivity and revenues of the

company. Daw summed up the challenge ahead of SI thus,

To ensure long-term sustainability and to strengthen its competitive position, it was not enough

to keep up with the industry’s average attrition rate that stood at 20%, or to solely rely on pay- based solutions for retention. SI had to differentiate itself, add value and lower its turnover rate

below the industry average.

Betting on Analytics

While it was able to attract the right talent amidst a severe talent crunch, SI was unable to retain them.

Daw knew that although a zero turnover was unhealthy as well as unattainable, SI’s HR policies

would have to control the direction and the level of existing turnover such that the turnover rate

among the lowest performers was at the expected level, and the turnover rate among the highest

performers was brought to as close to zero as possible. According to Daw, this would warrant a

thorough data-driven analysis of the company's recruitment, development and retention policies, and

how the policies influenced the turnover drivers.

6/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.

SMU-20-0011(A) Semicon India: Demystifying Workforce Analytics

Daw quickly got to the task at hand and listed the questions that needed to be addressed: What were

the potential drivers of turnover? Were the opinions held by the HR team valid? What were the

hypotheses to be drawn to test the impact of the drivers of turnover at SI? What data would he require

to analyse the hypotheses, and how would he source the data? Was there a means to extend the

responsibility of retention beyond the HR department and make team managers accountable? If yes,

how could he prove his claim of managers’ accountability towards retention?

Daw remained undaunted by the extensive volume of data that would be generated by the 1,600

strong workforce of SI. He was aware that if he got the fundamentals such as the hypotheses right,

the insights from the data analysis would help in identifying retention solutions, and hence support

SI’s business strategy. However, first and foremost, a proxy metric confirming the link between

productivity and retention had to be identified.

EXHIBIT 1: SEMICON’S PERFORMANCE RATING SCALE

Rating Definition Highlights

1 Overall performance, including both

results and behaviours, represents the

highest level of contribution.

• • •

•

Exceeds all expectations Difficult rating to achieve year over year Highest opportunity for career progression and development Eligible for significant rewards

2 Overall performance, including both

results and behaviours, represents a

high level of contribution.

•

•

•

Meets all expectations and may exceed some

Opportunity for career progression and development Eligible for competitive rewards

3 Overall performance, including both

results and behaviours, needs to

improve to increase overall

contribution.

• •

•

Meets most expectations Opportunity for career progression and development targeted at improving performance Eligible for limited rewards

4 Overall performance, including both

results and behaviours, does not meet

the expected level of contribution.

•

•

•

•

Does not meet expectations. Performance needs to improve Identify and document actions to improve Development opportunities targeted at improving performance in current assignment Not eligible for rewards

5 Less than six months active.

Source: Internal data from the company

7/7

For the exclusive use of J. Giesbrecht, 2021.

This document is authorized for use only by Joanne Giesbrecht in 2021.