Fractal Analytics: India's Second-Largest Analytics Firm

It could be a scene out of sci-fi thriller Minority Report: Procter & Gamble (P&G) executives seated in an oval-shaped room at their headquarters in Cincinnati, Ohio, flanked by two concave 8-ft-by-30-ft screens that have analytics dashboards coursing with product, financial and commodity information. Only, here, Tom Cruise isn’t slicing and splicing data across a fluid display—a data analyst is.

It is no secret that P&G, which did $84 billion in sales last year, is a data-driven business. The world’s largest FMCG company touches four billion consumers every day, generating terabytes of data. Visualising numbers helps the country, category and brand heads identify products which are working, and areas where new opportunities lie, all in real-time. Called the ‘Business Sphere’, this model has become the stuff of corporate legend; the innovation even won P&G’s Filippo Passerini the Information Week ‘CIO of the Year Award’ in 2010.

But P&G did not single-handedly create this analytics approach. It had help from heavyweights like SAP and Cisco, as well as a made-in-India company called Fractal Analytics.

“P&G is the most data-driven company we know of. They use data everywhere,” says Srikanth Velamakanni, 39, co-founder and CEO of Fractal Analytics, a company considered a “strategic partner” by P&G. “Usually their strategic partners are 100x Fractal’s size, like Accenture or HP,” he points out. Fractal’s current revenues stand at Rs 120 crore. Velamakanni is visibly proud of his company’s work—after all, Business Sphere is now used in over 50 P&G offices worldwide and is on its way to getting patented.

Apart from such high-profile associations, what has put Fractal, India’s second-largest analytics company, in the limelight in recent times is a Rs 150-crore investment by Boston-based private equity fund TA Associates in June 2013. Naveen Wadhera, country head of TA Associates India, is bullish on the industry, saying data-driven decision-making has become a major tool for companies, even expanding its reach beyond just the large-sized firms.

Power of Persistence
Forbes India met Velamakanni and Fractal co-founder Pranay Agrawal, 38—the two remaining members of the original five that created the company—the day after their marathon five-hour meeting with the global head of insurance major AIG, who had flown down to Mumbai in his private jet. “The CEO of such a large company doesn’t visit an office like ours unless it is a high priority for him and analytics is a high priority for AIG,” says Velamakanni.  
 
The Fractal office in Andheri, Mumbai, is ideally suited for the business it is in. On entering, you quickly realise that it is a haven for nerds, number lovers and creative data crunchers.

“We are geeky enough to call our conference room the ‘Nash Equilibrium Room’,” says Velamakanni. (Nash Equilibrium is a solutions concept in game theory, named after mathematician John Nash.) And the geeks are now churning out algorithms and other data analytics tools that aid companies like P&G, Colgate, Kimberly-Clarke, VISA and Mastercard make smarter decisions.

Big-ticket partnerships didn’t fall into Fractal’s lap overnight, though. Quality execution of smaller projects helped the company get its foot in the door. It has come a long way since it took on a project with ICICI Bank, its first client, in 2001. Using mathematical models, Fractal built the first statistical consumer credit risk scorecard in India as a pilot for the bank.

Back then, however, the world wasn’t quite ready for data analytics. Later, in 2006, Fractal approached P&G with a proposal for a four-week research engagement worth $2,000. They spoke for almost two years with no success till 2008, when it finally got the green signal.

Impressed by the results, P&G considered Fractal—along with 25 other companies—as a potential global analytics partner. Two, including Fractal, were picked to pitch head-to-head. Fractal won.

It was a significant moment for the fledgling company. P&G wanted to project global demand: 15,000 country-category combinations that would have to be forecast and reforecast every quarter. It sent a two-person team to work with Fractal which deployed 10 employees for the project in 2008; now, the team has expanded to 200.

Then, in 2009, Fractal did a project for an American automotive lender which was trying to recover money from bad loans: It built a model that prioritised recoveries. “The interesting thing was that they were losing money because the dynamics were changing in the time between repossession of cars and recovery,” Velamakanni says. This was a tough one and required an all-nighter from the young team.

Fractal believes analytics is shaking up the insurance industry, among others. The traditional actuarial insurance sector has always been adept at understanding risk but the playing field is changing with new sources of data. People are re-evaluating actuarial tables because one’s own loss experience is not good enough—market experience matters. “If you don’t understand risk—and you price for average risk—only the bad customers take your policy,” Velamakanni says.

Wadhera also underscores the disruption caused by analytics. He says he became interested in the space when observing portfolio companies in the US, such as one which sold movie tickets online. It used buying pattern insights to pre-release tickets to high priority customers. It could then go back to film studios with crucial information on the demand for upcoming movies.

And that is how analytics works: By helping companies identify opportunities. “Capital One suddenly became the most effective credit card company [and business school example] in the US,” says Wadhera, “when it stopped charging fees and, instead, started to mine tremendous amounts of data to figure out where profits get generated, shifted focus to the most profitable customers and changed the paradigm for credit cards.”

He says GEICO (Government Employees Insurance Company) did the same for insurance. Retailers can use information from payment portals to understand who the most price-sensitive shoppers are and offer them discounts. “Particularly in consumer-driven or customer-end-point companies, you have people telling you something every day—why aren’t you using that signal to drive your pricing, strategy or timing?” Wadhera asks.

An Industry On the Cusp
People often only use experience as their basis for choice. But gut instinct supported by some data is even better, according to Fractal. “As human beings, we are poor judges of things like credit, for example. We are inconsistent, subjective and volatile. There is a balance one has to strike between algorithms and experience,” Velamakanni says. In some cases, he feels, data can replace that intuition altogether.

Agrawal sees outsourcing of high-end, big-ticket functions like analytics becoming the norm. He says a lack of supply in the analytics industry is making companies process data in-house. “[But if] you have an ad agency, an IT partner, a law firm, an audit firm, etc… why not an analytics firm? Companies will ask: Why are we doing this stuff ourselves?” he says, underlining the need for analytics firms.

The feeling is that the analytics industry is where the IT industry was 10-15 years ago: Those who wish to stay ahead must protect their market share or watch it get eaten up by the big dollars of large consulting firms.  

There are more than 200 analytics providers in India today; there were 60 when Fractal surveyed the market in 2007. Despite the growing numbers, the consensus is that the potential of this business is so high that dogfights over individual clients are not necessary: The real challenge is growing the industry from $4 million to $10 million in India.

With far higher revenues, Mu Sigma (over Rs 600 crore) is the clear market leader but second-placed Fractal claims to be growing faster. Other prominent players are Delhi-based AbsolutData, LatentView Analytics of Chennai and Analytics Quotient from Bangalore.

Fractal, like competitor Mu Sigma, has capitalised on a knowledge gap, says Wadhera. “Mu Sigma is a very good company,” he adds. However, no one has more consumer product companies as clients than Fractal, Wadhera says. Because they essentially trade in insights and privileged knowledge, Fractal, TA Associates and their clients are understandably guarded about exactly how analytics are put to use.  

Ramki Sreenivasan, former CEO of analytics firm Marketics, says both Mu Sigma and Fractal are pioneers in the space—and their biggest hurdle is non-consumption. The amount of data being generated far outpaces the extent to which it is being utilised by companies.

“Companies might be data-rich but they are insight-poor,” he says. Analytics helps convert data to insight if coupled with smart domain knowledge. Scaling up for the industry will mean assessing which large companies are receptive to data analytics and targeting them.

In Marketics, this would be referred to as ‘analytics maturity’—a critical parameter to evaluate client readiness for analytics. Output is measured by how much value you can create for the client and how fast. “Data has shelf-life,” says Sreenivasan. “Six-month-old data doesn’t matter when retailer stocks are going to run out tomorrow. You need stock levels now.”

A Journey To Remember
Velamakanni remembers well the day Fractal got going: Five IIM-A graduates met on January 2, 2000, in his one-bedroom apartment in Takshshila, in Andheri East, a northern Mumbai suburb, with an open playground as a conference room.

Each founder brought Rs 2 lakh to the table. The original plan was to create a dotcom to support online purchases, but in six to eight months they figured it wasn’t going anywhere. For the first year, they didn’t draw a salary; they raised $600,000 from angel investor Gulu Mirchandani of MIRC Electronics who still owns a large stake.

The next round of capital infusion only happened in 2013, with employees and founders still owning over 40 percent. “When we started, we wanted to help consumers make better decisions. Then we decided to help companies,” says Agrawal.

The journey wasn’t easy. It took them till 2005 to become a $1 million company. By the end of 2007, eight years into business, Fractal was just 100 people and $4 million in revenues—modest by most measures. That year, two of the five founders, RK Reddy and Pradeep Suryanarayan, left the company on bad terms after an election to make Velamakanni CEO didn’t go down well with the partners. The split was well publicised in the media much to the exasperation of the remaining founders. Fractal lost 25 percent of its team and 25 percent of its revenues; when General Motors went under, Fractal lost 10 percent of its billings in one shot.

But then things changed. Revenues are up 3x from 2011. “Revenues grew by 35 percent in 2009-10, 45 percent in 2010-11, 55 percent in 2011-12 and 70 percent in 2012-13. Fractal is growing at 80 percent today,” says Velamakanni. He says Fractal has been profitable every year apart from 2009 and Ebidta (earnings before interest, taxes, depreciation and amortisation) levels are at roughly 20 percent currently. The company has 650 people and is hiring aggressively. Though Nirmal Palaparthi, the fifth co-founder, left on amicable terms last year, the turnaround has been remarkable.

What changed? Assisted by the resurgence of the US economy, strong Fortune 500 balance-sheets and the emergence of big data analytics, Velamakanni and Agrawal moved to New York and Fractal shifted focus from smaller Indian assignments to bigger ticket American/global clients.

Of and by the people
“I’ve been friends with Pranay longer than I have known my wife,” says Velamakanni; they were both recruited by ICICI Bank during campus placements, and worked on the first CDOs (collateralised debt obligation) in India in 1999. Of the pair, Velamakanni talks faster and rattles off numbers and figures whereas Agrawal has a more measured way of speaking and tends to give a macro perspective. In that sense, they complement each other well.

A young family in New York means that Agrawal, whose official title is executive vice president, spends far less time in India than his co-founder. He works closely with TA Associates and leverages their 40-plus years of experience in the US markets to deepen ties with existing clients and acquire new ones.

There are plans to double the team to 1,200 people this year and grow to 5,000 in the next three to four. Acquiring and retaining talent will be one of the biggest challenges as financial services and IT and consulting giants look to attract the best. Velamakanni says the free meals, childcare services, transport facilities and employee stock options Fractal provides has helped keep attrition at 14 percent, down from 16 percent two years ago.

It was a ‘people’ issue that caused the founding team to unravel in 2007. Each member was running a piece of the business independently, but investors and the board wanted a more formal structure. Velamakanni, who was elected CEO in 2006, says the split happened for two reasons: First, the others didn’t agree with the nature of the new system; second—the real cause, according to him—they were all too geographically separated with two in Singapore, one in Mumbai and two in New York. “Where you stand is a function of where you sit,” he says.
 
If their revenues are anything to go by, the company has grown stronger after the split. Agrawal summarises why: First, a formal structure fell into place and the business changed from partnership to corporation, and from that emerged speed in action and clarity of leadership; second, they started inducting senior talent from the industry and third, they had been spending energy serving only domestic clients and now got more focussed internationally.

In 2007, the team had spent five days rebuilding the company values during an off-site immediately after the split. Of the 20 senior management team members who got together, 17 are still around. Perhaps the people have spoken.