It's time we acknowledged that ERP systems are here for the long haul and start adapting to permanent life with them. At this year's Sapphire conference for SAP users, I couldn't help but reflect on how much money is involved in the business of ERP.. One tip-off that there are profits to be made was the items being given away: not the typical coffee mugs, golf balls and cheap pens, but Porsche Boxsters, VW Beetles and Harley-Davidson motorcycles. I didn't win anything, but the excess of it all inspired a few thoughts.

One involved how long all this will go on. The conventional wisdom is that ERP sales will decline substantially after 2000, when organisations are no longer trying to solve their Y2K problems with a fast (right!) ERP installation.

Perhaps in anticipation of this problem, several ERP vendors' stocks are disproportionately down.

I, for one, don't predict any big decline in ERP fortunes. It's already too late for a rational person in a big company to purchase an ERP system and expect to jam it in by the mother of all New Year's Eves. In fact, just yesterday someone from a large chemical company told me that his firm was putting off buying an enterprise system until after 2000, so I wouldn't be surprised to see a surge in sales then.

No, ERPs are not purely a result of millennial craziness. For vendor and user companies, they are not a flash in the '90s pan, but rather a long-term phenomenon.

Unfortunately, many companies that have bought the packages are treating the systems not as a till-forever-do-us-part marriage but as a brief cohabitation.

However, no matter how depressing it may seem, "going live" with an installation is just the beginning (but presumably still better than "going dead").

We have treated ERPs as projects, with the assumption that some day the projects would end. But an enterprise system is not a project, it's a way of life. No organisation that I've talked to-and I've talked to over a hundred on this subject-has yet to say, "We're finished." And few ever will. There will always be new modules and versions to install, acquired or divested business units to deal with and better fits to be achieved between business and system.

Even if an organisation could declare final victory on implementation of ERPs, many additional years could be spent in getting real business value from them.

The realisation of a lifelong commitment has begun to dawn upon several companies. Some of the early adopters of SAP, for example, who implemented the R/2 mainframe version of the package, are now faced with an expensive, time-consuming conversion to R/3. One presumes that there will eventually be an R/4; then many more companies will be faced with the same prospect. Non-IT employees and managers who served as functional experts, or "super users," during implementation still can't return to their previous jobs because their help is needed in support roles. And companies that justified an ERP project on the basis of reduced IT head count are finding that they are unable to lay off anybody; people no longer needed for implementation are now needed for support and maintenance.

The permanent nature of ERPs has numerous implications. Here are just a few:Organisational StructureMost firms have created implementation project offices and appointed project managers with the assumption that the project will end and life will go back to normal. But it won't. So what they need is not a project office, but a new organisational structure that reflects the ongoing need for ERP-related activity. "Sponsorship" is one example of the need. Many companies appointed senior executive sponsors-often a CFO or COO-for implementation projects. Their expectations were probably that the CIO could own the systems after installation was "complete." But many companies are postponing business change until after they ram a basic system into place. Who's going to oversee those changes and ensure that they fit with the rest of the business? If the executive sponsor is temporary, who will ensure that the system and the business evolve hand in hand?One response is to create a hybrid role that oversees change in processes, systems and information. There aren't many of these around, but they're not unprecedented. An organisation I once worked with had a senior vice president of strategic change (we're not talking tactical change here, so you need an SVP to do the job). Another option is that CIOs can oversee long-term ERP change, but you'd better let everyone know before accepting the job that you're going to venture far beyond the realm of technology.

Of course, there are other organisational implications of these systems. A substantial support capability will have to be created, but that can probably just be glommed onto the existing IT support organisation-though it will become a lot bigger. More important, the long-term use of ERP could change thebroader organisation in terms of its structure and function. It may be, for example, that there is less need for middle managers and that the organisation can be a bit flatter. Maybe that's a responsibility for the "change czar."Roles and SkillsThe post-ERP organisation will need a different set of roles and skills than those with less integrated kinds of systems. At a minimum, everyone who uses these systems needs to be trained on how they work, how they relate to the business process and how a transaction ripples through the entire company whenever they press a key. And don't think that this training will ever end; new people will always be coming in, and new functionality will always be entering the organisation.

Most companies use consultants to help with the implementation process. I'm hardly going to say this is a bad idea, as I get a monthly check from a consulting firm. But there is a big problem with how most companies use consultants in ERP implementations: They don't transfer knowledge from consultants to internal employees. If these systems will be around forever, it's important to have internal talent that knows as much as any consultant about how they work and how they can be configured to fit the organisation.

Make sure your consultants allow your people to work side by side with them on the implementation project, and before they leave tap the most knowledgeable consultants on long-term system evolution issues.

In every business function and department that's affected by ERP (which means pretty much everybody except janitorial services, although I hear the JS module is coming soon), you'll need one or more people who know the system and its relationship to departmental processes. In the early days after you turn the system on, these people will save your bacon. They'll answer questions, find needed work-arounds and let you know what's working and what isn't. But, as I suggested earlier, the super user role is a job for life. You might as well put super users' responsibilities on paper and perhaps get them some sort of dual reporting relationship to the IT function and their department manager. It will also be useful to convene meetings of these hybrids every once in awhile so that they can share knowledge and group therapy.

Knowledge Management

If you don't want to hire everyone in your standard metropolitan statistical area to help resolve ERP-related problems, you had better adopt a dose of knowledge management to leverage human support. Again, big companies like Johnson & Johnson have had knowledge networks and repositories to allow knowledge sharing during implementation, and this is a good idea. And if consultants have helped you implement, make sure that what you've learned about the system during that project is captured in a repository. But how about the post-implementation knowledge of how to use the system and related new business processes?Most companies will benefit from a knowledge management initiative in which both ERP experts and garden-variety users can put what they learn about the system and processes into a knowledge base and then search it for what knowledge others have added. Since system and process knowledge will often be inextricably linked, it's important to capture both.

I am excited about a software application that performs this function from a startup called Ventix Systems Inc. (Warning: As an adviser to this company, I have a minor stake in its success.) It allows hot-key access to a knowledge base from within SAP R/3 and eventually other systems. If they're smart, all ERP vendors will have some sort of knowledge management capability embedded in their systems, but today it's only available from Ventix Systems.

Application Management

Further evidence that living with ERPs is an important issue is provided by several vendors that are bringing out ERP application management tools. Envive Corp. and BMG Entertainment have separately announced capabilities to monitor SAP systems and pass on any troubles they discover to humans or service management systems. And for a while now ServiceWare Inc. has offered prepackaged SAP support knowledge for help desks. It's all more grist for the mill that attention is shifting from putting ERPs in place to living with them comfortably for a long time.

Just as courtships and honeymoons are different from marriages, living with ERP systems will be different from installing them. Projects for implementing the systems get a lot of resources and attention, but I believe that how we live with ERPs will have more to do with the value we receive from them than the quick decisions made during installation.

(Thomas H. Davenport is a professor of management information systems at Boston University School of Management and director of the Andersen Consulting Institute for Strategic Change. He welcomes reader comments at thomas.h.davenport@ac.com.)