Easing Data Conversion Woes for an Early Education Client

Project Summary

This early education client wished to upgrade their data architecture from PeopleSoft to Oracle Cloud for enhanced functionality and reporting capabilities. As one of the largest privately-owned child care providers in the US, they wished to bring 10 distinct brands and hundreds of schools to their existing Oracle Cloud platform. Premier was enlisted to perform data migration for this Oracle Cloud Financials and Procurement implementation in 2019. Our project team included the clients’ functional and technical leads, system integrators from a partner consulting firm, and our own project managers and developers. After months of development and test cycles, the project was put on pause in the spring of 2020 due to the COVID-19 pandemic. Premier resumed project activities in September 2021 to iterate on existing data conversions and support the production loads for a February 2022 go-live, completing six conversion cycles in six months to ensure the utmost data quality in the new system.

Project Requirements

The client’s go-live comprised of the following Oracle Cloud data objects: Suppliers, Supplier Banks, Purchase Orders, General Ledger, and Fixed Assets. Source data came from PeopleSoft and Excel reports from business users. To minimize the impact on school operations, the Cutover schedule was incredibly compact. The goal was to load historical General Ledger, Suppliers, Supplier Banks, and Purchase Orders before go-live and load Fixed Assets and the most recent GL period after go-live. So as not to disrupt the pre-go-live processes in Peoplesoft, two separate data snapshots were necessary before go-live and a third was needed for the post-go-live conversions. All data objects were to be loaded into the already live client HCM Oracle Cloud environment.

Key Activities at Cutover

We instilled confidence our process by using our repeatedly enhanced data conversions for our final Cutover loads. At the time of Cutover, we provided 42 distinct files to support the in-scope data objects. We converted over 4 million records, performed pre-load validations using our Critical-to-Quality checks, aided in client validations, and provided load files to the technical leads from the system integrator. To adapt to this client’s high data volume, we generated and provided Excel validation extracts to the client leads as well as Excel extracts and load files to the system integrator instead of the typical Oracle FBDIs (File Based Data Imports). All data provided at Cutover was successfully loaded into the client’s Oracle Cloud system.

Project Risks

Given the project restart, we faced unique and unanticipated challenges. However, we used our agile approach to adapt to the circumstances and support our client in this unprecedented time.

COVID-19-Induced Challenges

The client and System Integrator (SI) both had notable resource changes after the project pause. We worked with client functional leads who did not have insight into the original data mapping conversations nor the functionality of the target system.

We adapted to these challenges by iterating on previous versions of Mapping Specification and discussing logic changes with the new client and SI leads. We modified preexisting conversion code and used our Oracle Cloud knowledge to provide solutions. We ensured that their organizational needs were met through repeated validations and transparent communication.

Client Technology Limitations

Large client data volumes caused server slowdown; conversion processes took substantially longer to run during the first cycle after the project pause.

Premier consultants tuned and optimized our data conversions programs to overcome limitations of the client server. Premier also consulted with our internal software team to troubleshoot issues in real time. We were able to get faster conversion processing times than ever before and delivered data extracts to the client leads with ease.

Data Transformation/Requirements Changes

The Supplier client leads wanted Supplier addresses standardized in a manner that did not align with the best practices address standardization approach.

We worked with the client leads to discuss their desired end state and implemented additional transformation logic to guarantee that their needs would be met. We tested this logic and ran our Critical-to-Quality (CTQ) pre-load validation checks to ensure that all processed records would load successfully into the client’s Oracle environment.

Client leads requested substantial requirements changes on the Purchase Orders data conversion during Cutover.

We typically do not recommend logic changes at Cutover given increased risk with the load to Production. However, based on the client’s needs and understanding of the risk, a consensus was achieved to proceed with the change order. Premier consultants worked with client leads to understand all aspects of their requests and robustly tested the newly coded logic. Client leads thoroughly validated the data extracts with all requirements changes encompassed. The desired changes were reflected and the data loaded fully into the Production environment, saving the client from making substantial manual changes to post-load data.

We found extensive duplication of supplier data within the legacy system. Additionally, the supplier format and structure required for Oracle Cloud was substantially different than of the client’s existing system.

Applaud’s powerful and flexible data matching engine was used to de-duplicate the client’s supplier data. Premier provided ample reporting tools to the client for them to determine which records should persist in their new Oracle solution. Across multiple test cycles, we gathered, implemented, and tested complex transformation logic that required cross-functional input and validations. Our custom error log reflected the coded changes and any cases that required further client investigation. We received some substantial structural Supplier changes during SIT2 and UAT, but adapted accordingly to meet the client’s data expectations.

Challenges Faced During Data Loads

Oracle Cross-Validation Checks (CVRs) were turned on for Purchase Orders (PO) during UAT without our knowledge. We loaded POs in the Production instance with CVRs on as well.

We dealt with significant PO fallout during UAT due to the CVRs being left on for the loads. We took a multi-step approach to mitigate the situation. First, we created a delta process to capture the POs that fell out in a prior load so we could try to re-load them again during the same mock cycle. The Oracle CVRs sometimes over-exclude records, so this process helped us to iteratively attempt to load previous fallout to eventually arrive at just the ‘problem’ records. Next, we conducted a second UAT test cycle to re-load POs with additional account transformation logic captured in the PO conversion. The objective of this approach was to mimic the back-end transformations that would be happening in Oracle so that validators could easily identify improper accounting combinations. These combined strategies, coupled with ample pre-load validation, helped us achieve 100% PO load success at Cutover.

The client’s Oracle HCM Cloud system was already live; all Procurement and Financial data was to be loaded/merged into this Production environment.

We needed to ensure that all newly converted data would be compatible with the already live HCM Production data. Our previous test cycles loaded data into a mock environment, so the Cutover loads were unique in this manner. We found inconsistencies with certain Location and Department configurations in Production that did not align with our repeatedly tested logic. Given our suite of Critical-to-Quality checks, we were able to quickly identify issues and subsequently notify the client team. We worked alongside the client team to guarantee that all affected Locations and Departments were updated accordingly. All issues were resolved prior to the formal data loads, paving the way for Cutover success.

Given the large data volumes of this client, the System Integrator requested that we provide data for loads using flat text files  instead of the typical FBDIs (File-Based Data Imports).

While this may seem like a slight and nuanced change, the data delivery process was substantially impacted by this request. Usually, the clients validate the data in the FBDI, provide sign-off, and that same file is used for the SI load resource to perform the associated data load. In this case, we took a multi-step approach to ensure that all functional and technical parties would have their needs met. First, we provided an enhanced Excel extract to the client with additional legacy information included for easier validation. Once the client approved that data, we would exclude any records that were flagged by our Critical-to-Quality Checks and output that data in two formats for the System Integrator: 1) an Excel file that mimicked the format of the FBDI and 2) a CSV with the same data for load purposes. Several data conversions required multiple sets of files be generated; we managed the data delivery with ease. Specifically regarding the Fixed Assets conversion, we provided data in chunks of 100,000 records or less for an effective and efficient load into Oracle.

Project Timeline

After the project pause, the client schedule was extremely compact (6 test cycles, including Cutover, occurred in 6 months).

We understood the needs of the client and the pressure to meet their ambitious go-live date. We leveraged our many years of experience performing data migration to reduce risk along the way. Premier worked closely with the client and the System Integrator to create test cycle schedules and set expectations for data snapshots, data conversion turnaround time, and data validation. Premier also went above and beyond to work with individual client leads to instill confidence in the converted data.

We were faced with incredibly tight timing around Cutover, particularly for Purchase Orders.

We worked relentlessly during Cutover to determine that all necessary steps were taken for our client’s success. We completed the entire Cutover PO data conversion cycle (from raw data to pre-load validation) in 24 hours. This timeline was substantially faster than other test cycles. To meet this timeline, we planned and conducted numerous test runs off of an earlier data snapshot as practice for the final Cutover PO conversion. This robust validation using the prior data snapshot allowed us to expedite the final Cutover PO sign-off. Premier worked around the clock with client vendors and leads to ensure that all of our bases were covered during this critical time.

The Results

Our client went live on-time and with 100% load success for all data objects. Shortly after the go-live, the client team rolled out the Oracle Procurement and Financials system across their entire organization. The client business users had confidence in the converted Oracle data and few manual updated were needed. In addition, our expeditious turnaround at Cutover allowed operating schools to function as per usual—school facilitators were able to order food and supplies with minimal interruption. Child care has become more important than ever given the COVID-19 pandemic; we are glad to know that our client is able to operate effectively to serve working parents and teach the leaders of tomorrow.

‍