Finding the Data Problem Before It Reached the Merger

The Situation

The client was negotiating a potential acquisition and merger, and their financial data needed mapping to the acquiring company's metadata structure beforehand. Each company's structure had been designed independently, making one-to-one mapping extremely complex.

Resolving data quality first was not optional. Mapping bad data accurately only produces accurately mapped bad data, and every integrated report built afterward would inherit the problem.

The Challenge

Our prior knowledge of the client's data, plus additional review during due diligence, surfaced issues originating in the SAP source system.

Billing documents were being coded to the incorrect product line. Product lines were not separately defined on invoices. Taken together, there was no reliable way to determine which products had actually been sold, which is a difficult thing to discover in the middle of a merger.

Our Approach

The client was advised that measures had to be taken to clean up the data in SAP itself. If the source system is wrong, nothing downstream can be right.

In parallel, we began extracting data from an Oracle database and creating a new Essbase database. A total of 185 fields were migrated. Working through the dataset revealed multiple columns holding blank data or duplicate members, along with fields that had been imported from other sources and were irrelevant to the mapping project. Those were eliminated.

Through ongoing communication with client stakeholders, the data was scrubbed and the fields genuinely needed for mapping were identified.

The Results

The project was completed on time and within budget.

In under four weeks the data was scrubbed, narrowing the dataset from 185 fields to 95. With clean data in place, the full integration of financial data could proceed, along with replication of the forecasting and profitability cost management allocation processes in alignment with the acquiring company.

"An organization of our size depends heavily on the talent and deep technical knowledge of subject matter experts and consultants."

Senior Director, Energy Company

Why It Mattered

Nearly half the dataset turned out to be noise. Finding that before the mapping work rather than after is the difference between a four-week cleanup and a merger-long argument about whose numbers are right.