Sometimes the ideal solution may turn out to be a single data center with a couple different databases but one application code where you can handle different divisions or collections of divisions that have radically different data models and are difficult to consolidate on the database level. But at the very least, the goal ought to be consolidating down to a very small number of databases and data models and a single application instance -- or a couple application instances at most -- that support all of your different operations.
The biggest challenges turn out to be not difficulties with hardware or software, but rather with the people involved in the process. Five to 10 years ago, when we talked a lot about global single instance, the technology was there to support it, but only in major cities in developed countries around the world. While it requires careful planning, at this point it is fairly certain that a company can get the bandwidth with maximum latencies required to connect users from wherever they are in the world. Certain parts of the developing world still have some challenges with connectivity, but most locations where a company might set up a distribution hub or a manufacturing facility, or where a sales manager may need to access customer relationship management (CRM) software are adequately served.
Regional barriers including multiple languages, time zones, currencies and units of measure are typically handled effectively by a modern enterprise software product. There may be limits to the number of languages and currencies supported by some products, and that ought to be a factor in selecting ERP for a global instance. But the primary barriers to this high level of global consolidation have to do with people rather than technology.
One concern is whether or not all of your different divisions can agree on a common set of standards like product nomenclature, customer naming conventions, customer views and credit management policies across different divisions around the world. As you move into a single data model, each entity has some ability to make adjustments to deal with their regional issues, but fundamentally, you are dealing with similar processes and a common set of master data. From a business standpoint, this can be a very positive thing, but each division may find it unappealing to, for instance, renumber hundreds or thousands of parts or restructure some of their supply chain or remanufacturing data to comply with a new corporate data standard.
Other impediments are inherent in how the different divisions or entities want to use the software. Most software today like IFS Applications is highly configurable and flexible, and can be modeled to conform to various business processes within an operating unit. But there is often a limit to how much variability you can achieve with a single instance of an application. That means that achieving global ERP requires overcoming some resistance to change, overcoming resistance to commonality. All of these are management problems rather than software problems, so they can be overcome by a strong ERP project team that has solid management backing.
Sign up for MIS Asia eNewsletters.