LSMW Estimation

[Former Member]

How can we do an effort estimation for LSMW projects ?

Any insights on this topic can be helpful generally to all the members. Kindly post your ideas.

BR

Tanmoy

[Former Member]

Hi,

Effort estimation is decided based on the priority and complexity of the object. So know the priority level and some times effort estimations differ from project to project.

Thanks & Regards,

Goutam Kolluru.

[Former Member]

Thanks for your reply.

I am thinking about the parameters that can be taken into consideration for the effort estimation of an LSMW project.

BR

Tanmoy

[JL23]

what do you understand under LSMW project?

what is the scope of the project?

do you mean establishing LSMW as a tool for data migration and mass changes in an IT organization? Preparing and conducting training for all the IT folks? replace existing migration programs by new developed LSMW objects?

or do you mean a migration project, where the scope is to migrate material, vendors, customers etc from a legacy non-SAP system to SAP?

or the same for an existing SAP system to a new SAP system?

including data harmonisation, or just the technical LSMW execution part?

should it be one-time objects - or objects that can be reused in the next project without any bigger change?

these are already parameters that infuence the effort estimation tremendously.

I give you an example: I loaded inventory for about 1800 materials last week on short notice.

I got the first extract in the morning at 9 from a functional collegue who knows what information has to be available to load inventory. so I started with a new LSMW object from scratch using BAPI import method. at 11 I was done with a complete test load

including consolidation lists. this was my techncial part, just 2 hours.

then the business had to reconcile and validate the test load. They were satisfied, I did not have to change anything in my LSMW. However, they amended some vaues and gave me a new extract.

I analyzed this new extract, compared it with old extract, and found that they had deleted a cell in Excel, which caused that all cells from beneath got shift one up. so inventory of material 1 had the value of material 2, while inventory of material 2 got value of material 3...

so I gave the extract back, said that I do not trust this anymore. Finally I got a third extract. had to redo my validations and loaded it.

So instead of 2 hours it finaly had taken about 10.

And this confirms my experience with many projects, the technical developement of LSMW is probably only 10 % of the time needed to complete the work. it varies based on the data that has to be loaded, a material master takes much more time than inventory.

loading SAP to SAP takes much more time than other legacy system to SAP, because the people want keep everything what they had in SAP, so you have to care about 10 times more fields than from a legacy system to SAP.

loading in other languages like Chinese, Korean, Russian is more difficult than for a western country.

Loading from templates that were prepared by people who are not familiar with Excel is probably the worst part. carriage returns within one cell, content does not start from character 1, data has much more characters than SAP can take,.....

do you load into new system or into an already running system?

do you load into an exsting organisation or into a new organsiations?

into new organisation means that you may have to extend esiting master data to new organisation levels.

but if you load into an existing organisation, then you have to take care that you do not overwrite existing data, that you are merging data instead of extending, or even both.

do you do the project with a well-established team, or is the team to be built new?

will you continue working with this team, or is it setup each time individually.

does the LSMW guy have technical and functional knowledge, or is a functional without muh technical knowledge, or is it a technican without functional knowledge?and want to mention the technicans without technical knowledge - I had one who was supposed to support me, finally it took me more time together with him than I had needed if had had all done myself.

I think that was a lot information and parameters which infuence the time needed for a project.But you still cannot say how much more time is needed for case A and in case B. Finally it is the slowest individual in a team who defines how much effort is needed.

[madhu_vadlamani]

Hi Tanmoy,

Jurgen answered almost all.That is depending on project what you are panning to upload.If there is a standard method available then you can map it and make a template it will be ready.if there is no standard method then you gone for recording it may take time.If there is ant table controls then it may take time.This all depends on practice and experience to handle that.

Regards,

Madhu.

[Former Member]

Hi J,

Thanks for your detailed answer.

What I mean is;

Its a Migration project, where the scope is to migrate material, vendors, customers etc from an existing SAP system (4.6C) to a new SAP system (ECC 6.0). I will mention the Business scenario for you. The SAP 4.6C is used as a part of the Extraction process by the Informatica team. After the Extraction, Informatica team goes on with the Transformation that has specific Business Rules. After this step, we are getting a flat file that contains the data records. These data records are used by a LSMW (Project, subproject & object) for the Loading Execution.

It doesn't include data harmonisation, but just the technical LSMW execution part.

The objects must be used in the production & Quality servers.

BR,

Tanmoy

[JL23]

I have currently such a SAP to SAP project, and I had already such in the past.

still it is quite different from project to project. I had done some in about 1 or 2 weeks for the big five (vendor, customer, material, batches, inventory) and my current is scheduled for more than a year effort

It depends often how differently the system and processes are setup. and of course how big the companies are, whether it is just one company and plant, or several companies with hundreds of plants.And never forget how long legacy system existed and how good the data quality is.

Does the to be migrated company have to adopt all existing processes?

or will some processes of them be implemented in the existing SAP?

are the rules already fix or is this still in ongoing discussions?

what is extracted? just a couple fields or the entire tables?

Do you only have the good data in the extract, or everything ?

e.g. you could have plant data for material master for plants that are out of business already for many years.

And if something is closed or sold, then nobody is there to clean this data up, so it is without deletion indicator, and probably among the good data.

I often had the case where rules were built based on entries found in customizing tables.

E.g. sales group to sales group.

But in the data I then found sales groups that were not mapped, because they did not exist in the customizing tables of legacy system. Means they deleted their unwanted sales groups from customizing, but did not care about existing data that had used these groups.

back to the extract. a download from tables is pretty work intensive, I often send the legacy data with an ALE scenario per IDOC to a file and use this file as source for LSWM. Especially here you have all data and you need to clean this up in LSMW.(create rules to ignore certain organisations) Means the effort switches from extract to migration.

[Former Member]

I would like to share a parameter too, what if the data in the flat file is garbage data; that kills a hell lot of time.

BR

Tanmoy