Understanding attrition in analytics sector in India

If you have worked with a software team based out of India, or know someone did, you would be aware that the technology sector in India has a high rate of employee attrition or churn. This is estimated to be as high as 30 %.

As someone who changed jobs a decade ago before I turned entrepreneur I try and understand what makes young people in this decade change jobs. Here are some of the few reasons I think they go through ( I am trying to think like the employee not the company)-

  1. Bad Leadership – No it’s not money, but the lack of a good working professional relationship with your manager that makes you quit job.  You can blame the young person for jumping for a 40% hike, but the only reason people stay on is they like their boss and like working under him. More investment in leadership training (and not just games /team building junkets /offsites) can be a solution
  2. Stagnation in Learning Opportunities– Young people realize they lack experience and are willing to take a hit to learn more. Their ambition is thwarted by situations that force them to do routine jobs (without proper explanation of why it is important) and lack of credible road map to guide their career expectations, and ambitions. More investment in digital training can be a solution.
  3. Unfair working times– If working conditions demand that you ask team members to stay in office more than 9 hours you need to explain properly why they cant work from home using remote desktop and internet. Employees that help their companies live their office actually sidestep this issue quite effectively.
  4. Lack of Stock Options– Salary is paid and consumed. Stock options make people hang on. Stock options are comparitevly less in India though Infosys has been an admirable company in it. More intelligent and fair to both sides agreeement on stock can be a solution
  5. Lack of penalties for attrition on manager– Managers get by with excuses on blaming attrition on employee, industry,salary but not themself. Surprise attrition is unforgivable. Linking vraiable compensation of team lead and manager will force them to develop a working relationship with team member.  Using analytics to predict attrition and using variable incentives to reward low attrition team managers can be a solution
  6. Cynicism on company processes– The failure of HR in India to stand up and confront line managers has prevented them from acting as a safety release for pent up steam in employees. Culture eats strategy for lunch and if your best people keep leaving at 30% your strategy goes for a toss. A segmented approach at attrition can help
  7. Delinking performance with attitude issues-  A company culture that obligates employees to ask and be responsible for issues in working conditions makes their manager link attitutde and perfromance unfairly.  BetterHR metrics and buddy/mentor arrangements can be a solution. This can also be reinforced by a matrix reporting structure.

Politics is the enemy of analytics. Indian culture is political per se, so making young people more professional and responsible in long term is the job of senior management nut just middle level team leads.


Seasons to turn

Some people like to take and take. They will demand more vacation time, more pay, more stock options. They usually end up whining in a negative manner having failed to achieve success in lets take and take attitude.

Some people pretend they like to give and give, though they really are sub consciously taking more than they are giving. Hypocrisy with attitude can work in sales and marketing but not in real life.

To build a unicorn statup you need ninja hackers. One of the best lines in the famous hacker attitude is – attitude is no substitute for competence

Sometimes we need to turn our past habits of smug success to hungry learning and probing. There are seasons to turn in every startup and you better change your attitude before your attitude hurts your baby unicorn.


Brief History of Analytics in India

Business Analytics as it used to be called before it became branded as data science has two decade long history in India. It’s roots were in outsourcing, as associate divisions of large Information Technology firms

The setting up of business analytics in the 1990s was done by two centres- McKinsey Knowledge Centre (1998) ( by the global McKinsey consulting company) and GE Analytics ( by GE corporation).

GE Analytics (1998) was a separate company headed by an XIMB Professor Shrikant Dash https://www.linkedin.com/in/sdash1 which was then merged with GECIS to be a division called ACOE analytics centre of excellence.  In 2005 GECIS was spun off by GE to become GENPACT. ( The author worked briefly for Dash, GE , GECIS  )

Among the people helping set up McKinsey Knowledge center were Neeraj Bhargava and his team. Neeraj later became CEO at WNS Global (2002) and helped take WNS to it’s IPO. Assorted McKinsey alumni helped to boot up WNS especially in it’s analytics division called Knowledge Services. ( The author worked briefly for WNS Knowledge Services)

Business research was primarily pioneered by Evalueserve founded in 2000  by IBM as well as McKinsey alumni. They also may have been responsible for popularizing the term KPO ( Knowledge Process Outsourcing) as a differential brand from BPO ( Business Process Outsourcing). Evaluserve expanded from Business research to financial research, market research but was not so succesful in efforts to establish itself as a leading analytics player.

From Captive Centres to Third Party Offshoring was the next big shift in outsourcing that affected analytics. In the meantime American Express had the largest captive centre (for analytics primarily at parent company).

GENPACT and WNS were and remained leaders in BPO but were joined by EXL (https://en.wikipedia.org/wiki/EXL )  in adding Analytics as an added service to its offshoring portfolio.

In June 2006: EXL acquired Inductis, a Analytics firm that had been a pioneer till then in being a pure play analytics company. There were other smaller companies that got similarly acquired ( Adventity by Sutherland Global , Marketics by WNS, marketRX by Cognizant, Symphony by GENPACT)

The leading IT and software services companies in INDIA like Infosys, TCS and in particular WIPRO did start their own analytics divisions but they were primarily aimed at cross selling analytics to existing clients.

GENPACT was listed in 2007 (https://en.wikipedia.org/wiki/Genpact) and its revenue from analytics is not known though it is generally considered to be one of the premium priced analytics company. As of 2012 it’s analytics revenue were said to be $250 million.

The confused strategy, premium pricing, cross-selling outsourcing with analytics  of all these players led to a gap that was exploited by Mu SIgma. Founded by Dhiraj Rajaram in 2004, Mu Sigma (https://en.wikipedia.org/wiki/Mu_Sigma_Inc. ) expanded rapidly to be the first pure play analytics company above 5000 employees. Dhiraj was brilliantly assisted by his wife Ambika, who played a role similar to what Sudha Murthy did at Infosys.Ambiga continues to works for Mu Sigma and has played multiple roles. This probably makes her the most influential  woman in data science as a service in India ( and given’s Mu Sigma size, probably the world )

( The author briefly met them when Mu-Sigma was 200 member and then 600 member strong)

In February 2013, Mu Sigma received an investment of $45 million from MasterCard, which placed the company over the $1 billion unicorn milestone. It also marked a vindication in the model of treating analytics separately and distinct from offshoring services

Fractal Analytics was founded in 2000 and managed to survive both the acquisition binge by the big players as well as the cyclical turns in the offshoring business. It is another pure play analytics company with both a long history, a comparatively early focus on pure play analytics as distinct from outsourcing and comparatively recent focus on venture fund-raising efforts. Fractal continues to grow rapidly, and has attracted multiple rounds of investment.

(The author has briefly met and interacted with the CEO of Fractal, where he also once trained in R)

In 2014, the people calling in analytics industry in India started calling themselves data scientists and they all lived happily ever after 😉 .*

( This is a brief article- it is a continuous work in progress that is trying to chronicle how an industry grew and flourished)

*the Big Data as well as Data Science movements have led to tremendous growth opportunities as well as risks for Indian analytics companies who now have to compete with Silicon Valley startups for analytics  as a product than as a commoditized service.

Dead Heroes make great T Shirts

Some of my heroes whom I wear on my T Shirts

1 Kurt Cobain – marrying the wrong woman can lead to a hole in your head

2 John Lennon – marrying the wrong woman can break up your band no matter how big. Imagine.

3 Bob Marley – Too many Drugs will kill you everytime

4 Freddy Mercury – Too much love will kill you everytime

5 Gandhi – Peace and Non Violence. Forever

6 Steve Jobs – I can be a jerk and get way with it if I am an iGenius

7 Jesus Christ –  Even Jesus can get angry

Jesus entered the temple courts and drove out all who were buying and selling there. He overturned the tables of the money changers

Data Science to kill and Data science to sell

My computer science professor at University of Tennesse tried to teach me genetic algorithms. I managed to score an A thanks to the brilliance of my team member who created a genetic algorithm using C. For some reason the loveable professor had something against C ++. Real men talk to the metal using C.

Anyways the professor had this theory to say about computer science. He said computer science had two principal sources of revenue.  One is increase kill ratio. The other is to make people sing your tune.

Department of Defense consumes the maximum revenue a nation produces globally across most nations and it funds the maximum data science atleast in the USA.  Making people click you ad is the other source that data science gets funded for. Apparently there is history involved here.  The Italian navy funded Galileo’s telescope ( but stood aside when the Pope disagreed). DARPA funded the internet. Insert here- standard joke on Al Gore inventing the internet.

Principal sources of data for this signal intelligence (sigint) fueled data science comes from telecom, websites and sensors. Principal sources of analysis is actually a huge amount of reports (written on conversations with humint) that are better of  mathematically text mined but are in practice dissected one by one.

Real intelligence begins when no guns are fired. The cold war on the internet is a real threat to free internet and cyberwar is just a way to combine kill ratios for servers to data sourced from ads.

In the meantime, I prefer the maths of Good Will Hunting than the maths of Patriot Games. I prefer the music of Bourne again.