Long before he became a professor of information and operations management at the Wharton School of the University of Pennsylvania, Ravi Aron worked in the product development group of Citicorp in New Delhi. Memorably, his boss at the time told him to forget what he had learned about marketing at the Indian Institute of Management, Bangalore because in a highly service intensive industry like ours, marketing basically means information management. Aron realized he was right: Almost everyone at Citicorp was busy collecting, analyzing, aggregating and disseminating information. I realized that information was our stock-in-trade our input and our output, Aron recalls. I began to look for emerging themes, and realized that service essentially means collecting information about the context of the buyer.
That realization, coupled with a desire to understand it more deeply, prompted Aron who has worked as a consultant in Malaysia to pursue a Ph.D. degree at New York University s Stern Business School. Since joining Wharton in 1999 his research has focused on various aspects of information management, including the pricing of information-rich services and the efficiency of business-to-business markets and online auctions.
These days Aron, in collaboration with Jitendra Singh, a professor of management at Wharton, is studying the outsourcing of business processes such as call centers from the U.S. and other Western countries to India and other parts of the world. Despite the general slowdown that dogs the tech industry, business process outsourcing (or BPO) continues to be regarded as a hot market. Gartner, a tech consulting firm, reckons that the global BPO market will be worth more than $240 billion by 2005. In a recent conversation with Knowledge@Wharton, Aron discussed his findings about the potential and risks of this emerging phenomenon. Wharton’s Mack Center for Technological Innovation and the Wharton e-Business Initiative (WeBI) have helped support this research.
Knowledge@Wharton: You have been studying the outsourcing of services such as call centers to India. What have you found so far?
Aron: I am studying the supply chain of expertise; call centers are one aspect of that phenomenon. A call center can be driven by the fact that the cost of human labor intervention is much lower in India than in the U.S. and other Western countries. The orders of magnitude are one-eighth, one-tenth or one-twelfth, depending on the kind of work. But call centers represent just the most visible component. In three to four years, you will see that the supply chain of expertise that connects the West to India, the Philippines, Brazil or Malaysia will not be about call centers but about other things.
Knowledge@Wharton: What do you mean by the supply chain of expertise?
Aron: Many things that are done inside firms are labor intensive. I m not referring to physical assembly-line labor; I mean the kind of work where human beings have to intervene in a process and use decision-making skills such as interpretation, validation, translation, transliteration or transformation. For example, when a dealer is given an 8.5% rebate, does that mean 8.5% of the package price or the promotional price? Computers can t make such decisions. Computers are good at sorting and manipulating large data sets and executing instructions. But when it comes to making judgment calls, you need people. Human beings are good at making meaning out of one medium and translating it into terms that computers can understand. Such processes are called back-office processes, and they are rapidly migrating from the U.S. and other Western countries to India, Singapore, Brazil and other parts of the world.
Consider some examples. In the call-center industry, companies like Daksh (which has the Amazon account), Spectramind (which works for Dell), and Tracmail provide call-center support and telemarketing services to clients. But that is just the beginning. British Airways has moved its back offices to India, as has the World Bank. At these offices people have to look at vouchers and other documents, make decisions about what they mean, and convert information into a format that can be aggregated or manipulated by computer which is typical of back-office work. From a research standpoint, what is interesting is that things that were once internal to a firm are no longer so. The boundary of the firm in some ways is being compressed, while in other ways it is being expanded.
Knowledge@Wharton: How does this happen?
Aron: When Dell tells Spectramind to handle certain services in exchange for a fee, it means what was once being handled inside the firm has now become a market transaction. As a result, the boundary of the firm is compacted. But a different phenomenon also is underway. An extended organizational form is emerging in which firms relinquish control in return for monitoring.
For instance, let s say I have a company and I move my back office to Chennai. I stop exercising control; I don t tell an employee, I want you to be in the office at 8:30 a.m. But I do look at the number of errors per 1,000 documents processed and control that quality. In addition, I randomly sample phone calls to monitor the quality of the marketing. Consider DirectTV, whose call center has gone live in New Delhi. If someone were to do telemarketing for me from that call center, I would not tell them how to do their jobs, but we would agree on the outcomes and I would pay them depending on the number of deals they closed for me. So the extended organizational form is one in which you relinquish local control it is another firm that actually hires the employees but you monitor those employees.
This is a very different model from traditional supply chains. A company like Johnson Controls might supply automobile components to GM, but there s no question of control or monitoring by GM these employees are Johnson Controls employees.
There s a big difference between the supply of materials and the supply of expertise. This is because IT enables deep linkages to be established between organizations. These linkages are much richer than when you are supplying car seats or windshields. It is reasonable to expect that DirectTV s customer service (or telemarketing) representatives have to be able to see the same version of customer information in Delhi as those at DirectTV s headquarters. This establishes very deep information-based linkages between the two outfits. So at one level, these are two companies engaged in a market-based transaction, but at another level you could view this as an extended organization.
Knowledge@Wharton: How have such extended organizations and business process outsourcing in general evolved?
Aron: Business process outsourcing began with the setting up of captive service centers by large transnational corporations. These centers, such as the ones set up by Citicorp and American Express in India, began by executing enterprise-wide operations that involved the conversion of data from one medium (such as documents) to another (digitized data in corporate databases). Companies such as American Express and Citicorp started moving more and more of their information extraction and reporting tasks overseas. Two factors that made this possible was the convergence in corporate computing platforms and the rapid advances made in communications technology. As corporations standardized a few enterprise wide platforms (such as Relational Databases, networking standards etc.) and with the availability of software tools that made it easy to port large data sets between dispersed information systems, the flow of data and information between geographically dispersed branches of the same corporation became a viable and nearly costless option.
Knowledge@Wharton: What difference did this change make?
Aron: As the flow of data between computers that talked to each other increased, so did the extent of human intervention and the degree of expertise required of the information worker to transform data into information. As a result, the costs of providing accurate and timely information to support middle management decision making increased by orders of magnitude. Corporations were faced with a two-pronged cost escalation they had to hire more information workers and at the same time ramp up the expertise levels of existing information workers who provided operational support to the managers. The move to a centralized operations factory in a lower-cost labor regime was an obvious response to the cost frontier faced by these corporations. Hong Kong Shanghai Bank Corp., for instance, has an office in India handling back-office work which employs about 1,000 information workers, and HSBC plans to triple this office size in the near future. America Online s customer service operations are supported from India. Initial reports suggest that some firms benefit to the tune of up to 60% cost savings in these lower wage markets.
Knowledge@Wharton: On what specific aspects of this phenomenon does your research focus?
Aron: I m studying at least three issues. Let s place this in its theoretical context there are two theories that broadly explain the existence of the firm: the transaction cost and the principal-agent theories. If you look at the extended organizational form, you can argue that people are sending work out to places like India for cost gains. But that also increases the transaction cost. What if the firm behaves opportunistically? So one of the things I m studying is, what kind of information is shared between organizations?
Secondly, I m looking at questions such as, what tradeoffs take place between monitoring and control; what are the metrics of performance and who agrees on them; what is efficiency and effectiveness and how are these measured; and how do you align the incentives of firms that ship out the work with those to whom it is sent? In addition, I am studying the so-called KIF problem or the knowledge-intensive firm. These include firms in industries such as insurance, banking, financial services, brokerage, health care these are now huge and sprawling organizations. Firms have no reason to be that big, and they are starting to unravel to some extent. Ford Motor realized back in 1915 that there was no reason for it to own rubber plantations in the Amazon so that it could make its own tires; someone else could make the tires. In the same way, health care firms, insurance firms and brokerages are beginning to recognize that they don t need to have all these back-end processes going on within their organizations. That expertise could be acquired much better and cheaper from someone else.
The third issue I m studying is pricing: contracts, fee systems, etc. Also, I m looking into whether pricing defines the employees behavior in some way. If you take your back office and make it someone else s front office, what sort of pricing structures do you use? Out of 50 or 60 things that you can study, these are the three I m focusing on.
Knowledge@Wharton: Have you reached any preliminary conclusions?
Aron: I m looking closely at these relationships in the financial services industry. Early results show that initially, as the process of organizational unraveling begins, firms look at the most easily definable tasks such as document reconciliation or call centers. After that, they look at a factor I call revenue distance.
The concept of revenue distance is simple: The point at which a financial services firm captures revenue where a customer works through its doors and agrees to buy a credit card or another service is the point at which its revenue distance is zero. It is a final and visible process that leads to the sale, but supporting that is a series of processes that run behind it. If you take one step back, before you call a customer to sell him a credit card or another service, you may have looked at his profile, your existing relationship with him, and so on. You may have sifted through a list of 10,000 people and found 1,500 potential prospects. That s a back-end process. If you take yet another step back, you may have acquired that list of 10,000 names from a credit rating company or some other vendor. As you take more steps away from the point of revenue capture, your revenue distance increases.
Initially companies ship out back-end processes that have really long revenue distance. At that point, their main motivation is cost. But by the time the relationship starts deepening, they realize that costs are relatively unimportant. Many back-end jobs in the U.S. are manned by people who are under-qualified and bored. Attrition rates in call centers in the U.S. range from 70% to 120%. It takes a month and a half to train a person so that he or she can hit the ground running, and then three months later that person is gone.
In contrast, in India, at its worst, the attrition rate is between 12% and 35%. A back-end job in the U.S. is a serious job for someone in India because it pays serious money. This applies all the more as you move away from call centers into other back office processes in financial services and other industries. Pentamedia Graphics is a great example. It is a large animation shop, which also has a U.S. office that does business development. Just as you can find talented engineers and software writers in India you can find talented animators. So you can argue that initially U.S. companies outsource operations to India for cost reasons, but by the time the relationship deepens, it s driven by anything but costs.
Knowledge@Wharton: What are the main drivers in addition to costs?
Aron: Cost benefits usually just open the doors. You may start by saying that your call center or back office operations are very expensive, and you want to look for other options. Thereafter you discover some interesting things. Consider quality. You might believe that if you are willing to throw money at quality in the U.S., you will get it. That s not true. You ll have to pay anything from a 40% to 80% cost premium (in direct wages alone) to get the kind of quality you could get in India. Inaltus, a processes outsourcing company based in Britain which specializes in F&A services (Finance and Accounting), estimates that premium in Britain to be upwards of 80%. In many cases, they point out that the quality is simply not available in the labor pool that competes for these positions in the UK (and several other Western countries). Many of the people in the U.S. who have talent and education don t want to do back office jobs. In the U.S. there is a serious mismatch between the kind of skills and temperament that delivers quality and the kind of people that are needed to deliver them. That mismatch doesn t exist in India. So even though a company may enter into a relationship with an Indian firm because of costs, it discovers that it gets much better quality than it had expected.
Moreover, companies that outsource back-office operations to India discover other intangible benefits. At one financial services firm, the back office in India started making suggestions about how processes could be streamlined in Europe and the U.S. That not only led to productivity gains, but also made for much faster customer processing and proximity. Earlier, if you wanted to cross-sell an insurance product to someone who had already bought a credit card or mortgage, it took six to 12 days to turn that around. You can now do that in 24 hours. You can find really smart people who are willing to make a career of this in India.
So companies start out for cost, stay on for quality, and then realize that they get a lot of managerial initiative. It takes a lot of headaches off their hands and allows them to stay focused on their core competence and to remain close to their ideal customers and serve them.
Knowledge@Wharton: What are the implications of your research for developing countries to which the supply chain of expertise extends?
Aron: Initially many companies looked at setting up back-office operations in Ireland and Canada; now the cost advantages have gone away in those markets. Ireland has a small population, and back-office operations are a sunset industry there. Malaysia and Singapore are trying to get in on the action. They have the infrastructure, unlike India. Getting your telecom infrastructure set up in India is like a doing a root canal without Novocain. You don t have this problem in Singapore; it s amazingly agile. In 48 hours you can get full connectivity.
You might also see hybrid situations where more skill- and labor-intensive jobs might migrate to India and parts of Malaysia, and and as a second tier, jobs which carry higher managerial content which require more Western management techniques and human GUI (graphics user interface) and more of a professional touch such as offshore CRM services – those might migrate to Singapore.
Knowledge@Wharton: Have you seen any examples so far?
Aron: Architecture offers a good example. There s a large firm which does a lot of work in Indonesia, but this is how it works: the Australian firm ships out part of its work to Singapore, where the cost differential is very high (in fact, they have many Indian architects working there), and this relationship is managed out of the Sydney office. This allows the firm to be in the region and yet not part of the same cost regime. So you might find middle offices move to countries like Singapore, Thailand or the Philippines. The single biggest advantage India has over these countries is the existence of a large, math-friendly, English-speaking population.
Knowledge@Wharton: What about risks? How should companies deal with the risk of intellectual property being poached as a result of outsourcing arrangements?
Aron: The principal risks are what my Wharton colleague Eric Clemons calls PSOR poaching, shirking and opportunistic renegotiation. These are very early days, and so most of the time you find companies under-committing and over-delivering. These risks can come up when the market matures, but right now it s still in the expansion phase.
Still, you have to be careful in a couple of areas. First, if the firm to which you have outsourced work takes ownership of your customer information, that greatly increases its bargaining leverage. Outsourcing firms that have this information can start renegotiating prices or contracts.
Secondly, there is the issue of lock in. When people think about lock in, they usually think about information technology. This is a minor part of lock-in; in fact, there is a great deal of standardization with web services. The greater risk is process-level lock-in. If employees get used to seeing the same screens and doing the same things while someone supports them from a back office, it s a tremendous bargaining advantage for outsourcing companies (the providers) vis-à-vis the principals (the users). Even if the outsourcing firm increases its fees, it becomes difficult for the principal to discontinue operations or change the way things are done. The human level lock-in of the person, the process, and the information manipulating system is much more dangerous and costly. Disruption of those processes is usually unaffordable; and that gives the outsourcing firms tremendous advantages.
There are of course ways to guard against these risks. One is to distribute risk. The other is to have a dummy standby mechanism for every process. And going back to the notion of revenue distance, as you come closer and closer to the point where revenue is captured, for those processes you should keep a skeleton back up, which might let you operate for two weeks without interruption if your supply is disrupted. That skeleton backup is enough because it changes your bargaining power.
Knowledge@Wharton: What opportunities do you see in health care? Medical transcription is being outsourced, but do you see other opportunities as well?
Aron: In the next four to eight years, the U.S. is going to see three important trends in health care. One is boutique medical care: For those who are able to pay, there will be a level of service that goes beyond what is now provided by medical care firms. Every time you hear the term, high quality service, it means there is a great deal of information manipulation. So boutique medical care is likely to take off and this will aim at providing health services priced at $60,000 a year to $400,000 a year. There are more than 1 million millionaires in the U.S., so this is not a tiny market.
Since not everyone is going to get the same suite of health care services, other parts of the industry will segment themselves into groups that make more intelligent use of customer information. In the U.S., some of this will be aimed at better capacity utilization. If, for example, a hospital such as the Columbia Presbyterian Hospital is accepted by 20 health care providers, you might want to make the case that it should be accepted by 200 of them. You can do several levels of fine segmentation, and decide how those costs are going to be covered among the providers.
The finer you price your services, the better it is for people at the lower end. If someone who costs $500 a month to service pays a premium of $250, and someone who costs $50 a month to service also pays a premium of $250, if you can make the premium reflect the risk status a little better, you might be able to reduce the number of the 40 million uninsured in the U.S. Finer segmentation means faster information throughput and movement in the back office. Two very large players in the health care business are already looking for outsourcing opportunities in this area.