Best, Worst, and Emerging Practices

in Information Technology Sourcing


Address by Mary Lacity, Assistant Professor of MIS, University of Missouri-St. Louis and Research Affiliate, Templeton College, Oxford University, England

Delivered to the National Commission on Restructuring the Internal Revenue Service, Washington, D.C., January 30, 1997.


Mr. Chairman and Members of the Committee: My name is Mary Lacity, Assistant Professor of Management Information Systems at the University of Missouri-St. Louis and Research Affiliate at Templeton College, Oxford University, England. I am here today to summarize the best, worst, and emerging practices in information technology sourcing so that organizations get the best value for their information technology dollar.

I would first like to credit my co-authors, Leslie Willcocks and David Feeny from Oxford University, and Rudy Hirschheim from the University of Houston. Professor Willcocks will also be presenting to the Commission today.

My co-authors and I have been studying information technology sourcing decisions for the last six years. To date, we have conducted case studies in over 70 US and United Kingdom organizations. We have interviewed over 200 senior executives, information technology managers, business unit managers, and outsourcing vendors. Our research base is unique in three areas, which enabled us to identify best and worst practices.

First, we studied what actually happened in organizations with their sourcing decisions. You may read in the trade press that companies hope to reduce costs, improve service, and access new technologies through outsourcing. Such reports are made when large contracts have just been signed. We studied whether their sourcing expectations had been achieved, and extracted practices that differentiate successes from failures.

Second, we studied multiple sourcing strategies pursued by organizations, including total outsourcing, total insourcing, and selective sourcing. We were able to compare the relative success of each of these three general sourcing strategies. Selective sourcing was the most successful option. With selective sourcing, organizations keep some activities in-house and outsource others, depending on who is best able to deliver value for that activity.

Third, by interviewing multiple stakeholders within organizations, we were able to study the inherent conflicts surrounding information technology decisions. In general, senior executives wanted low costs because they pay for information technology. On the other hand, users generally want service excellence because they use information technology, but do not associate a direct cost for their demands. Information technology managers are subsequently placed in the precarious position of providing Rolls Royce service at a Chevrolet price.

Although the vast majority of the cases we studied were in the private sector, we find many similarities with the public sector. In both the public and private sectors, organizations want to get the best value for the money they spend on information technology. In both sectors, users demand more information technology than can be supplied. In both sectors, multiple stakeholders often create conflicting expectations for information technology. In both sectors, organizations struggle to keep up with the dizzying pace of emerging technologies. Both sectors must adapt systems to changes in the environment, be it federal regulations or business requirements. And both sectors must balance trade-offs between low cost solutions like standardization and high-service solutions like customization.

The most important lesson derived from our research base to get best value for the information technology dollar is the need to properly diagnose information technology problems before making a sourcing decision. There are no white knights or magical elixirs outside of an organization to make information technology problems disappear. You have to know diagnose the right problem and then determine how best to fix it.


Letís look at three examples.

Letís first consider an example from my own institution. Our information technology costs were considered high and our Board of Curators received two proposals from external vendors to reduce our information technology costs by 10%. Our Board of Curators hired a new CIO to oversee the outsourcing decision. This CIO had 35 years experience working for a vendor and instead of immediately outsourcing, he investigated why our information technology costs were high. Our costs were high because we had three data centers for three different academic units on the same campus. He could have outsourced and let the vendor perform the consolidation, but instead he did it in-house and reduced our information technology costs by 20%--double what the Board of Curators expected from their first look at outsourcing.

A second example is a company that has a poor track record with developing new applications. The systems are consistently late and over-budget. Will outsourcing applications development to outside experts help? Not until the company understands why systems are late and over-budget. Systems development may have a poor track record because users participating on the project keep adding functionality that was not in the original project proposal. The systems are late and over-budget because bigger systems are delivered. Such a problem requires a senior level business manager to determine which additional functionality truly adds value, and which functionality is merely bells and whistles. Such a problem requires a method for freezing user specifications during project development. If applications development is outsourced before this problem is fixed, the legacy is merely passed on to a vendor. As users add functionality, the vendor must add excess fees beyond their bid price.

Let us consider a third example to illustrate the importance of proper diagnosis. I recently met with a team investigating outsourcing at one of our government-funded research labs. For every 5 dollars that comes into the lab, 1 dollar is spent on information technology. The lab directors have charged the information technology management to reduce costs, possibly through outsourcing. Information technology costs are high because each division in the lab buys what they want and then expects the information technology staff to connect it to the network and integrate it with other systems. The lab needs to determine if the decentralized procurement is necessary to achieve the labís mission. It may be that a physics division needs different technology than the biology division. Or it may be that significant savings can come from standardization and consolidation without hindering the labís objectives. Such decisions must be made before looking at the outsourcing market.


Once an organization properly diagnoses their information technology problems, they can then intelligently evaluate sourcing options. In cases we have studied, the most successful decisions involved two separate teams: a decision-making team and an internal bid team.

The successful decisions appointed a decision-making team comprised of senior level managers and information technology managers. Senior level managers provided the larger business perspective--such as the mandate to reduce costs--and the power to enforce decisions. Information technology managers provided the necessary technical expertise to develop requests for proposals, to evaluate vendor bids, and to negotiate, and manage sound contracts.

The internal bid team is a separate group of information technology specialists and information technology managers that submit a competitive bid along with external vendor bids. Some senior executives in organizations we studied were reluctant to allow internal bid submissions, asking, "Hey, if the information technology staff could do this better, why havenít they done so already?" In many cases we studied, information technology managers did not have the power to enforce change or implement best practices due to organizational politics. Users may not want practices that reduce costs because they think it will lower service. But outsourcing evaluations empower internal information technology managers to behave like vendors. Senior managers communicate the mandate for change to users, whether it is implemented by external providers or their own information technology staff. In 8 of the 9 cases we studied where the internal bid prevailed, the information technology managers successfully implemented their proposals. In instances where the vendor bid prevailed, the organization was confident that the vendor had an advantage that could not be copied in-house.

Turning to the decisions which lead to outsourcing, what differentiates outsourcing successes from failures? In successful outsourcing decisions, the organization knew exactly what they wanted from their information technology provider.

· The organization selected activities which were well understood and they could therefore negotiate a sound contract.

· The organization signed contracts only for the duration for which requirements were stable.

· The organization practiced hands-on management of the contract--they did not throw the contract in the drawer and think-- "now the vendor can do this for me."

As I always tell my students, great classes are made by great students, and not great professors alone. The same is true for outsourcing--great outsourcing relationships are made by great customers, and not great vendors alone.

One common example that illustrates the successful use of the information technology services market is transitional outsourcing while migrating from mainframe systems to smaller platforms, such as client/server. This new world of client/server creates a resource shortage within organizations. But what should be outsourced? The support of the old world or the development of the new world? Many organizations successfully outsource their old world and insource the development of the new world. Because organizations understand their requirements and the costs to run their mainframe legacy systems, they can negotiate a sound contract with an external provider. The contracts are typically only signed for 2 or 3 years while the requirements for the legacy systems are stable. The in-house team may also seek external help for the development of client/server technology, but this outside help is managed by the internal project team. We call this sourcing option "buying-in" vendor expertise. The vendor transfers learning about the new technology to in-house personnel so they are in a position to support the system after development. Examples of companies pursuing this strategy of outsourcing legacy systems and insourcing the development of client server include Sun Microsystems, Elf Alochem, Owens-Corning Fiberglass, NASDAQ, Sears, and Esprit.

Finally, Iíd like to mention some emerging practices in the outsourcing market. Many of these practices are a result of the organizational learning which leads to more sophisticated use of the information technology services market. One example is performance-based contracts. Rather than negotiate a fixed-fee contract, customers pay the vendor based on the vendorís performance. For example, the City of Chicago negotiated a contract with EDS to develop a ticket processing system. EDSís payments are based on the revenue generated from the new system.

Other emerging sourcing options have names such as value-added outsourcing, co-sourcing, multi-sourcing, and flexible sourcing. It does not matter what labels we give these emerging options, because the best practices for ensuring value for your information technology are the same: proper diagnosis of problems, rigorous evaluation of sourcing options, and sound contract negotiation and management. Obviously there are other factors to consider. It is difficult to summarize 6 years of research in 15 minutes. But the procedures for proper evaluation makes a big difference to an organization like the IRS. Using them effectively can mean success which benefits the whole organization, and in the case of the IRS, all Americans.