Among three variations for SAP software re-implementations there are two key approaches. You either make the changes to your existing production system (or a cloned copy of it) or you make the changes in a pristine, newly designed environment.
SAP Cloned Production System “Re”-Implementations
Making changes in your existing production system (more likely in a cloned instance of it) helps ensure data consistency and ease of adjustment, however there are several difficulties involved. If you have a significant number of custom-coded solutions you will have to fight them every step of the way. You will have to work around them, deal with them throughout the process, adjust any out of date coding, and most likely will end up keeping many of them. As you can tell, I’m not a big fan of an SAP reimplementation in a production system with lots of custom coding.
You either make the changes in your existing production system or in a newly installed instance with no data.
For example if you decided to consolidate organization structures from a multi-system environment you might quickly discover lots of hard coded values in custom programs. These hard-coded values in the programs themselves, rather than using table driven values and parameters, can cause system consolidation nightmares. This is just one type of problem from many of the custom-coded solutions so often provided.
Another problem occurs with any of the existing system configuration. If you make changes to existing objects that are already in production you have the challenge of timing and coordinating your cutover to prevent disruption to existing processes. Depending on your circumstances if you decide to do a transition in your own production system with the eventual goal of moving away from it and into cleaner environment then it may be best to create all new custom configuration objects. You have to make that determination.
Clean SAP Re-Implementation with Old Legacy SAP for History
The other approach to SAP re-implementation is to do a re-implementation in a clean, non-modified system. That approach assumes that after conversion to the newly designed environment you will leave your old “legacy” SAP system in place for reference and historical data only. Using a new system for a re-implementation means that you do not have to work around any of the bad setup or design decisions that were made previously. You avoid all of the headaches with the custom programs and only bring in those custom programs that are really business critical.
If necessary, and if you already have a BI / BW / or other reporting system it will require some additional work to integrate old data structures with new. However even that will be easier with standard functionality. The SAP BI / BW / BObj reporting options already contain a number of standard extractors that can be used more easily and with less expense.
The Optimum Solution is a Phased SAP Global Instance Harmonization
The most cost effective way I have found over the years to do a reimplementation is to bring in an operation that is moving to SAP in a “clean” environment. It is not particularly complicated to integrate two SAP systems using ALE (Application Link Enabling). In this way you create a new environment, with more up to date and more standard functionality that you can eventually migrate other business units into.
As upgrade projects occur it is only incrementally more expensive to migrate the upgraded companies into the less customized environment. With an upgrade you still have to do the custom ABAP program reviews, code validations, etc. With a cleaner environment that does not have all of the custom coded artifacts it is much easier to pick and choose what is really of value and what can be replaced by new, or better understood functionality.
For additional rollout locations there is virtually no additional cost over the rollout project for bringing those companies or organizations onto the more standard SAP environment. In fact, the reduced custom coding would tend to be less expensive because the amount of time spent regression testing custom functionality, or fixing any organization specific settings, as well as training people how to deal with some of the custom functionality would be lower. Consulting time, and therefore consulting cost, would be lower as well because the closer you stay to standard the larger pool of resources there is available to make or adjust system settings rather than work with custom programs.
SAP Organization Structure and Master Data Harmonization
One other possible project approach is to do the SAP Org Structure harmonization in all of the separate SAP global instances and then agree on the common master data types. At “go-live” you extend all of the existing data in each production instance to begin executing with the new structure and master data types in each production instance. By doing this, the “legacy” data and “legacy” org structures stay in place so that little or no business disruption occurs. A transition period of approximately a year is needed to complete at least one full annual financial close under the new structures and data in the existing production system.
By using this approach you are actually making the transition in two-steps. First you build out the new state in your existing system, then after flushing out and adjusting most of the issues you do a conversion or cutover to a clean system after a financial close. This approach allows for an orderly transition from the old to the new with relatively little business disruption. While the old SAP org structure elements and master data types are being made “obsolete” they are still available for all processing and reference purposes.
Some of the key considerations for this approach involve what to do with custom coding and how to transition the master data. It is impossible to know what custom coding is in place that might be replaced with standard functionality. However some new data types may help resolve the issue of moving off the custom coding in the same system. Eventually the goal would be to upgrade away from the custom coding and into more standard functionality unless there is some clear business justification preventing this.
- https://usngerip datamanagement it