Payroll Parallel Testing

[Former Member]

Hello everyone,

I am leading the testing stragegy for a HR SAP implementation and would like to know if anyone could share best practices for payroll parallel testing (preparation, cycles..).

I have some experience on that and know it is very time consumming. We did parallel testings with sample populations, but this time business insists in doing it for full populaiton. Does anybody has a recommendation around what would be efficient regarding volume?

Also I would like to know if anybody knows efficient tools to perform the comparisons.

Thanks!

[Former Member · Accepted solution]

Are you implementing Time Evaluation and Payroll Processing? Will you also handle Third Party Processing and Transfer to FI/CO?

Look at all the processes that you will be using, and at the data output from them, and look to what data they could be compared in the Legacy System.

Parallel Test as much as you can (data entry, data processing and data output) to find gaps and problems.

We spent a lot to get a program to extract data from our Legacy Systems into a format that could be downloaded into SAP (Master Data Infotypes, Old Payroll Results, Quota Balances, etc...) and we used it for the integration test, the parallel test and the "Go-Live", but no software was used to compare data after processing.

The Payroll Journal, the Remuneration Statement and the WageType Reporter (in SAP) were used extensively to generate the data to compare with the Legacy System Reports.

[Former Member · Accepted solution]

Hello Remi,

Thanks for your answer. Regarding volume, did you perform parallel testing for the whole population? How many PRNRS do you think would be a reasonable sample (for example a percentage of the population...)?

[sikindar_a · Accepted solution]

Hi

Apart From Percentage Sample say ie 50 % those PERNR Should meet all the legacy conditions other wise we may not get accurate results in Parallel Testing

ie in those 50% you may have Take only Male Employee you may able to test only tax related to Male PERNRS But not Female so please have a look into it

The testing should include all the Scenarios which are applicable to Employee in Legacy and than try to corss the values

[Former Member · Accepted solution]

Hi Sikindar,

Thanks for your advise. I agree with your suggestion, to take a sample but being careful to choose all scenarios

What I need is some help on how to handle parallel testing for high volume of PRNRS.

The problem is that the business insists in testing the whole population (about 7500 PRNRS). I think is very time consuming and not feasible and I am trying to give them some arguments to make them understand the challenge that this would represent for the project schedule. Do you have any experience in parallel testing for huge populations?

Thanks all for your help in advance!

[Former Member · Accepted solution]

We did 11 conversions (from Legacy System to SAP) in 6 different "waves".

In the Integration Testing, we only took a representative sample of the employees to be integrated, in order to validate the configuration and identify the gaps.

In the Parallel Testing, we uploaded the Master Data and Old Payroll Results from all employees (lets say after Pay Period 10. Then the Master Users went through the process of entering the data for PP 11, Time was evaluated and Payroll was processed for all employees, and results in SAP were compared with results in the Legacy System.

It is more intensive, but the users get to experience SAP (after having classroom training), you get to fully test the Data Transfer tools that were built between the Legacy System and SAP, andyou get an acurate picture of the gaps remaining (if any).

[Former Member · Accepted solution]

How many PRNRs and testers did you have for this testing? And how many payroll period did you parallel test? Thanks!!

[Former Member · Accepted solution]

Hello Viviana,

So far in my implementation we have done payroll and time runs for the whole group.

Based on the business:

1. You can prioritize the most common or recurring scenarios. Basically create a test script for every scenario based on the configuration performed.

2. You can also talk to business SME's to find out several scenarios based on their past experience regarding complex payroll calculation, time data processing or any retro scenarios.

3. Once you get the detail outline then setup employees for those test. You can run payroll for the whole group but on parallel validation side, you will still use the same PERNR group you have setup and do the compare.

4. Later on, you can finish postings, TPR and all other subsequent steps for the whole group.

I am insisting on the whole group just to be on safer side.

Perform same study for Off cycle payroll and retro active accounting scenario.

Hope this helps.

Arti

[Former Member · Accepted solution]

Hello Arti,

Thanks for your e mail. Yes, we are planning to do exactly that in our functional testing. Data Conversion will be done at a certain point for the whole population and testers will work with real data (running actions, payroll, time upload, interfaces, etc), but will not be sincronized with legacy. Payroll and time will run for the whole population, but calculations will be checked for a sample.

However, in addition, what I am trying to do is another type of testing which will consist in choosing a payroll period, run it in the legacy system and try to run exactly the same in SAP. Then, compare results in both systems, justifying any discrepance. This means to convert legacy data into SAP just before payroll runs so that both payrolls will use the same master data. This is what we call Parallel Testing. Did you ever do something similar for a hugh pupulation (more than 7.000 PRNRs)?

[Former Member · Accepted solution]

Yes thats exactly what we do.

You can run "Gross to Net" reports on in legacy and SAP and can use those reports as a tool to identify discrepancy for such a big population. If you find discrepancy do the data validation in both systems first and then go for configuration in SAP.

Make sure you test every scenario for doing calculations on samples and document it.

I think you're right on track. GOOD LUCK!!!

Let me know if you need additional details.

Arti

[Former Member · Accepted solution]

I don't remember the exact numbers, but the number of PERNRs varied from 1500 to 15000 per implementation.

For the parallel tests, we had the whole HR/Time/Payroll team on site, but some users stayed only a few days so that they could do their part of the process (Input of HR Master Data, Input and evaluation of Time) while the key players (the future Super-Users) stayed on for the whole 2 weeks.