The emergence of Internet-enabled computing has made the flaws of the shared-services model difficult to ignore.
For a long time the shared-services approach to managing operations like accounts receivables or human resources was the only option large corporations had to keep down the costs of running support services. Vital support functions were concentrated in centralized departments serving many business units, rather than each business unit managing its own services.
Perceived advantages of this pooling of support activities came from economies of scale and scope, i.e., sharing of expertise. But these benefits blinded many companies to the considerable disadvantages. For one thing, centralized departments are less responsive to the business priorities of individual units and less focused on innovation, because any changes must pass muster with multiple business units with very different needs. Further, with management oversight often lacking, performance standards tend to be extremely low.
Also, very few support activities can be strictly compartmentalized: Processing a supplier invoice, for example, can involve half a dozen or more functions, including purchasing, engineering, IT systems, legal, taxes, banking, and accounting. The intersection of functions is even more relevant if the company is trying to raise efficiency in the total business rather than just tinkering with a single process (e.g., introducing procurement cards in the payables process).
Moreover, what's forgotten in touting the shared-services benefits of scale and scope is the harm to productivity and customer relationships that can occur when things go wrong. Consider a business manager for a German multinational who is trying to address a client's question about an invoice dated several months earlier. He has to contact the centralized accounts support unit in Rotterdam to request copies of the backup data. His budget will be charged 50DM for this information, which he probably won't receive for at least two days (if he's even sent what he actually asked for). The client becomes increasingly frustrated because his query isn't answered, and the business manager, equally frustrated, has also wasted valuable time.
One solution, of course, is to have online access to archived files. But that's expensive. Another option is the call center, a corporate-wide unit set up specifically to respond to questions from customers or suppliers. That, too, is costly to run, and is yet another hidden expense of the shared-services model.
Many companies turned to outsourcing the management of shared services, thinking that by getting even more scale and concentration of like activities they would capture the expected benefits of the shared-services model. But that adds another level of complexity to the already difficult process of overseeing different -- but linked -- activities. Imagine trying to achieve process improvements when your company has outsourced IT applications to a systems company, IT operations to a separate company, legal to a law firm, and transaction processing to an accounting firm.
The solution lies in new, more powerful systems functionality available through Internet-based models at a fraction of the cost of typical IT applications. Interactivity, networking, and robust data search engines are almost free. Imagine creating an interactive environment to give a business unit the appearance of operating as a self-standing department, fully in control of the services it needs. With this approach, it's possible to let each business unit behave and operate as though it were a small independent company, with access to its customer accounts, outstanding receivables, payables, personnel records, and the like. All these would be individually identified and extracted for the business unit from a centrally controlled set of master files.
What does it take to pull this off? Look no further than dot-com startups, entrepreneurs, and independents who see their businesses through the lens of new technology and aren't burdened with the typical large corporation's legacy computer systems and management processes.