A lot of companies routinely refer to all their partnerships as "strategic," but they're kidding themselves. A hard look at the partnerships a company has assembled will generally show that they do too little to support its overarching strategy and may even work against it.

The mismanagement of relationships will become an even bigger problem as partnerships become key to the identity of just about every organization. In decades past, companies were generally portfolios of products. In the 1960s and '70s, some companies began to see themselves as portfolios of companies and built conglomerates. In the 1990s, companies have decided they are portfolios of core competencies and have begun outsourcing non-core operations—such as computer operations—distribution centers, and cafeterias, or even product design and manufacturing. But as companies cut back to focus on what they do best, they still need all their previous capabilities, so they are going to deal much more with outsiders. Companies will, in effect, become portfolios of partnerships.

To manage your relationships as a group—as opposed to individually, as most companies do now—you have to first understand what relationships you currently have. That requires mapping them into four categories, based on how much expertise your partners are providing and how much they are customizing that expertise for your company. The categories are:

MARKET EXCHANGE: The most basic and common kind of relationship, a market exchange doesn't involve much expertise or customization. A good example is telephone service. You could buy that from any number of companies, and your needs are pretty standard.

These relationships aren't necessarily simple. But how to manage market exchanges is well understood. You, the buyer of the services, do the managing. You base decisions mostly on price by lining up competitive bids. You keep your transaction costs as low as possible by automating as much as you can about your relationship with the supplier.

PERFORMANCE CONTRACTS: This type of relationship involves using a standardized process, but customized for a specific environment. The primary example is outsourcing. Your partner takes a standard approach to managing your mailroom but adapts to the way you do business.

You and your partner need to share the planning and control, but you can do so at relatively low levels, with account managers, because the process of outsourcing is straightforward. The best basis for this kind of partnership is a contract with clear goals—such as minimal downtime at a computer center—and incentives for meeting those goals.

SPECIALIZED RELATIONSHIPS: This kind of partnership involves coming up with a customized way of doing a rather common task. Let's say you want to become better at logistics. You may hire a consultant with fairly standard knowledge about inventory management, but that partner will help you create a different system, designed specifically for the way you operate.

This kind of relationship involves more risk than the first two categories because you and your partner are more tightly linked as you try to develop something together. So, you need someone more senior, a relationship manager, to act as your liaison. You can't set the sort of clear goals that go into a performance contract, but you can do some benchmarking to make sure the process that you and your partner come up with is world-class. You need to do that benchmarking periodically, to ensure that you continually improve the process and remain world-class.

UNIQUE RELATIONSHIPS: The last kind of relationship involves considerable expertise that is applied in a unique way for both you and your partner. Microsoft and NBC, for instance, set up a unique relationship when they established the MSNBC network on cable TV and the Internet. The two sides took specialized capabilities—Microsoft's technology knowledge and experience with its MSN Internet operation, plus NBC's news-gathering capabilities and experience in broadcast media—to do something truly innovative.

These relationships need to be managed at the most senior levels of your business and your partner's. These partnerships are the most important because they give you ways of experimenting with new ideas. But you have to be careful that you don't share so much of your expertise that you transfer your core competencies to your partner. You also have to be ready to start managing your partnership differently, because unique relationships generally change into one of the other types of partnerships over time as competitors catch up with your innovation.

Once you've done your mapping, you'll manage the portfolio differently because you'll start to notice some things:

You can, of course, make sure you're not miscategorizing a partnership and, thus, managing it badly. IBM made this mistake with Microsoft on the personal computer. IBM thought in 1980 that it was establishing a market exchange, that it was simply buying rights to an operating system from a vendor. But Microsoft immediately saw that the two were setting up a unique relationship and managed accordingly. Microsoft, which had just 31 employees when IBM approached it, drew so well on IBM's credibility that it became a juggernaut that now has more than twice the market value of IBM.

You can make sure that you have partnerships in all four categories. You should, because all these types of relationships can help you. In particular, you need to be sure that you really do have what companies generally think of as strategic partnerships—the unique relationships that can help you innovate faster than you could on your own.

You can go to your partners and make sure that they view the relationship the same way you do. If you place a high value on your partnership, while your partner places a low value, you're generally going to have trouble.

You can analyze the mapping to see what it suggests your corporate strategy really is. When I did this for Sears Canada, based on 200 relationships, my interpretation of its strategy differed radically from what some members of the corporate management board said the strategy was.

That sort of discrepancy is dangerous and will become more so. Success is becoming less a matter of who you are than of whether you can create a formidable "competitive footprint" that is far bigger and more powerful than the company itself. Such a footprint can only be accomplished through a well-managed portfolio of partnerships.

Henderson, a professor of management at Boston University, has done 15 years of first-hand research on partnerships at prominent corporations. He can be reached at jchender@bu.edu.