Garbage In … Garbage Out: An Implementation Pitfall!
Hi Guys! I got home late on Tuesday this week. What’s on your mind? Hmmmm, Kenny (as I’m fondly called) was busy configuring an SAP Business One system or doing some kind of remote support. Nope! I was engrossed in a discussion with my Managing Partner, Deola. The crux of our discourse is why implementation fails. Deola attributed failed implementations to “Data” and he said, Kenny, when you put in garbage into the system, what you get is garbage. I nodded my head in agreement. He gave an example of how a CEO lost confidence in the accuracy of his financial reports as a result of incorrect data loaded into the system barely two weeks after go-live. And you know what? Your guess is as good as mine. The multi-dollar project was thrown into the bin! And the implication is bad referral.
At a very high level, I’m discussing the following tasks as it relates to data management during project implementation: Data Definition, Data Collection, Data Clean up, Data Analysis, Data Testing, Data Re-engineering and Data Migration.
Data Definition: The first thing to do is to have a clear-cut idea of the data needed for the system to work. These data forms the bedrock of the master data. This is important so that right from the start, you know what data you are looking at. This is also an eye opener into the size of the data you are expecting to get. Furthermore, a proper definition of data needed can serve as a guide when planning for your hardware, especially the server and backup tools.
Data Collection: Identification of the people to meet for the required data is important. You must be ready to put them on their toes. This can be a daunting task, especially if the company you are implementing for is coming from a non-system environment. It is even possible that the data needed are not readily available or difficult to extract. Hence, the need to do some forms of “pushing”. The data collected could be in hard copy or soft copy or even both.
Data clean up: This is where the bulk of the job is. Cleaning data can be challenging. This involves filtering i.e. separating the shaft from the wheat. In most cases, the data that you get irrespective of your definition contains lots of junks. It’s important that these data be filtered, so as to identify the ones you need and the ones you don’t need.
Data Analysis: It’s important to analyze the data collected, especially when batch loading will be carried out. It is at this stage that coding (material no, business partner no and employee ID) and grouping (item group, customer group) conventions are carried out. This has to align with the supported templates provided (something like the data migration templates in SAP Business One). The customer should be guided in choosing and adopting a naming convention. At this stage, it is important to ensure that all mandatory (primary) fields are filled. As much as possible, ensure that no fields are left blank in your template.
Data Testing: It is also important to test the data before go-live. Test database can be created to simulate what the live system will look like. During training, the test/training database should be used to test all possible business scenarios. Task oriented testing and process (integration) testing should be carried out. Ensure you stretch the limit of the testing. If possible, exceed the elastic limit!
Data Re-engineering: After data testing, as it is expected, updates need to be made to the data. This may require “throwing out” and “bringing in” data. This can be done in parallel as the need arises or after the full testing.
Data Migration: Final migration of data to the live system should only be done after the data has been thoroughly tested. It is pertinent to note that, before final data migration, both the consultant and the client must sign off.
Thus, we can say that, data is an important determinant of whether a project will succeed or not. Right from data definition to collection and then to final data migration, lots of work has to be done to ensure that the right data are loaded into the system. Each state of data management is important as the others. This is because the output of a stage is the input to another stage.
At a very high level, I’m discussing the following tasks as it relates to data management during project implementation: Data Definition, Data Collection, Data Clean up, Data Analysis, Data Testing, Data Re-engineering and Data Migration.
Data Definition: The first thing to do is to have a clear-cut idea of the data needed for the system to work. These data forms the bedrock of the master data. This is important so that right from the start, you know what data you are looking at. This is also an eye opener into the size of the data you are expecting to get. Furthermore, a proper definition of data needed can serve as a guide when planning for your hardware, especially the server and backup tools.
Data Collection: Identification of the people to meet for the required data is important. You must be ready to put them on their toes. This can be a daunting task, especially if the company you are implementing for is coming from a non-system environment. It is even possible that the data needed are not readily available or difficult to extract. Hence, the need to do some forms of “pushing”. The data collected could be in hard copy or soft copy or even both.
Data clean up: This is where the bulk of the job is. Cleaning data can be challenging. This involves filtering i.e. separating the shaft from the wheat. In most cases, the data that you get irrespective of your definition contains lots of junks. It’s important that these data be filtered, so as to identify the ones you need and the ones you don’t need.
Data Analysis: It’s important to analyze the data collected, especially when batch loading will be carried out. It is at this stage that coding (material no, business partner no and employee ID) and grouping (item group, customer group) conventions are carried out. This has to align with the supported templates provided (something like the data migration templates in SAP Business One). The customer should be guided in choosing and adopting a naming convention. At this stage, it is important to ensure that all mandatory (primary) fields are filled. As much as possible, ensure that no fields are left blank in your template.
Data Testing: It is also important to test the data before go-live. Test database can be created to simulate what the live system will look like. During training, the test/training database should be used to test all possible business scenarios. Task oriented testing and process (integration) testing should be carried out. Ensure you stretch the limit of the testing. If possible, exceed the elastic limit!
Data Re-engineering: After data testing, as it is expected, updates need to be made to the data. This may require “throwing out” and “bringing in” data. This can be done in parallel as the need arises or after the full testing.
Data Migration: Final migration of data to the live system should only be done after the data has been thoroughly tested. It is pertinent to note that, before final data migration, both the consultant and the client must sign off.
Thus, we can say that, data is an important determinant of whether a project will succeed or not. Right from data definition to collection and then to final data migration, lots of work has to be done to ensure that the right data are loaded into the system. Each state of data management is important as the others. This is because the output of a stage is the input to another stage.
posted by Kehinde Disu @ 5:04 PM
0 Comments:
Post a Comment
<< Home