One of the hardest things within IT is to properly integrate IT into the business processes. More often than not bureaucracy gets automated, instead of simplifying the work for the people who need it most.
Today we explain how the IT can be embedded so tightly it transparently becomes part of the business process. McCoy & Partners have worked in close co-operation with the customer to improve the review process of process order variances in SAP.
Reviewing the process order variances is a monthly recurring process which takes a lot of effort when done manually. All the sites jointly create about 60.000 process orders each year. With SAP still being rolled out in new countries, this number will continue to grow.
Some order variances are due to one-time issues, whereas others have a more structural nature. At least 80% of the variances have common causes, to manually sift through them is laborious. Each country had their own solutions and processes in place to do this and the diversity made it practically impossible to create uniform reports on a global level.
Summarizing, there were a lot of reasons to harmonize and automate the processes over the sites:
Sheer size of the data (>60.000 orders per year) makes manual labor expensive
Monthly recurring process
Most problems originate from a handful of common issues
Global management does not have proper overview
A few years ago, the global finance department made a first attempt at harmonization by collecting the data from the local sites and storing the data into an MS Access database. This solution was actually quite a big success: the data input was formalized, there was a global way of working and from this central point of data, reports were distributed on a regular basis. The audience varied from production staff members to global finance managers and everybody was happy.
Everybody? Well, at first, yes they were.
But it did not take long before things started to crumble under the heavy weight of great success. Especially the responsible controller was feeling the pressure, because he was The One Guy who chased the business for data, scraped the ECC system, scheduled BW queries, ran EveryAngle extracts and produced the output reports at every month-end.
Instead of spending time on analyzing problems, his precious time was consumed by data collection activities and he had become the single point of failure in the system. Even his holidays had to be carefully scheduled.
A second round of improvements was necessary, including items as
Must be an IT managed application to eliminate the SPOF
Closely tied to SAP, as 90% of the data resides in ECC
Additional master data and customizing parameters can be added to the model
Fast cycle times, especially during month-end closure
User comments are stored in the data
Simplified reporting in close loop with financial actuals processes
Globally accessible from a portal
Enforce data and access authorizations to comply with global policies
Three possible directions were identified to move forward, varying from staying close at home to moving into undiscovered territories.
Each one of these scenarios would have the following implications regarding activities and sources for data.
The scenario closest to the original solution would be to move the database from MS Access to an enterprise platform like MS SQL Server, Oracle, DB2, etc. All existing screen logic and reports could be retained, as well as the well-known way of working. The enterprise database could then be exposed to BusinessObjects for reporting.
However, there would still be a local Access application (non-IT managed), which has to be rewritten to use the new backend database. Developments would be split over front-end and backend, making it more complex. Most importantly: the data collection would still be done manually.
The best in house solution was proposed to be built within the SAP BW system, and if you look at the architecture pictures, it does not really look that much different from proposal 1b (full migration to an enterprise SQL database).
On the SAP application server, a custom-built application would be deployed to maintain the specific master data, and to review the order variances. These data would then be merged back into BW for global reporting purposes.
Possibilities included migrating the BW system to SAP HANA, with custom schemas for the review application, Fiori front-end developments and tight integration with the existing BW finance data. Other data warehouses were also considered, like MicroStrategy, which had out-of-the-box functionality for commenting on data.
Best on market would require substantial investments which were not feasible, both in time and money. Even though there would have also been a lot of benefits for other business areas, there was no business case to venture further possibilities in this direction.
The customer decided to go for option 2b, to build the solution on the Best in house platform, with a custom-built WebDynpro application for the review process.
There were a couple of reasons to build it on the SAP BW platform rather than ECC.
Most of the data needed was already available anyway. Reporting on BW was already done by finance, so now all would move to one platform. Existing reports could also benefit from the additions, so it made perfect sense to extract this data to BW.
Authorizations were already worked out pretty well, just a few extensions had to be made.
The BW OLAP engine could be leveraged for number crunching activities, like auto-reviewing the variances. This auto-review process took off most of the burden for the business, as it is able to categorize around 80% of the variances automatically.
Now that the categorizations are stored and retained, it becomes very easy to refresh the dataset of variances. In principle, the business can now review the orders on a daily basis if they want to prepare most month-end activities many days ahead.
The process has become a fully automated cycle; the figure above explains it. The WebDynpro activities are shown in green. In blue are the BW activities and in yellow are the transitions between BW and the review application.
A custom WebDynpro application was developed which enables users to maintain the custom master data they needed for categorizing the variances (e.g. reason codes, action codes, organizational assignments).
Structural issues can be stored in the application, so that future process orders will be automatically tagged with the appropriate codes by the auto-review process in BW.
The data is then pulled into the BW system to enrich the existing data. InfoObjects like Process Order are enriched with the additional attributes like organization assignments and review codes.
Also the review data itself with the comments is loaded into BW, where reports are built on top of specific InfoCubes.
In short, the review application has brought a rich set of improvements to the business, in bold the major drivers.
Bulk of data is reviewed automatically (>80%)
Review can be prepared any day, not just at month-end
User friendly, one-stop review application
Standardized global reports (plus self-service reports possible)
Authorizations are set up in a compliant way
Closely tied to finance processes, closest to real-time
Application resides on one IT-managed platform with global access
Covered by existing licenses
Full cycle review process, no hiatuses in data flows
User input is retained and is used in following cycles of the process
User comments are present in reports
Globally available BW master data is enriched, making existing report more valuable
Future-proof platform, ready for in-memory solutions like HANA and DB2 BLU for top scalability