ERP data cleanup should begin with business processes

Poor data quality can negatively affect various aspects of a company’s operations, especially as AI and advanced analytics continue to gain importance. ERP data cleaning benefits the entire organization because it helps ensure report and dashboard accuracy as well as reliability of data for AI agents and human decision-makers.

Unfortunately, most ERP data cleanup projects simply extract information, run matching algorithms, confirm information against external sources and fill in any blanks. Within weeks or months, duplicate records and missing values start showing up again because the root cause is the business processes through which key records are created or updated.

Achieving long-term ERP data cleaning benefits like faster closes, reduced payment errors, accurate inventory and trustworthy AI requires starting with a review of business processes.

Why low-quality ERP data is often caused by bad processes

Duplicate records for the same customer in an ERP system is usually blamed on an employee. However, examining the upstream business processes can open up other possibilities. For example, the sales department might create a customer without checking the master or each branch office may use a different naming convention.

Duplicates are only part of the picture. For example, fields that are optional in one channel might be required in another, or purchasing may create a second vendor record because the first record lacks a bank account or payment term.

The quality and variability of a company’s processes are the greatest factors in determining ERP data quality.

Begin with the business processes that rely on the data

Achieving lasting benefits from an ERP data cleanup starts with defining the top pain points. Which processes are heavily reliant on clean data?

For example, order-to-cash requires accurate customer information, such as address and contact information, and requires that item records match what was sold and shipped. Duplicates, missing information, inaccuracies or inconsistencies negatively affect the entire order-to-cash workflow.

Procure-to-pay also depends on unique vendor records that include accurate bank details, payment terms and tax IDs. Meanwhile, supply chain planning requires accurate and up-to-date item masters, units of measure, lead times and location-level inventory, and record-to-report depends on a chart of accounts, cost centers and intercompany rules that are the same every period.

Companies should create a team to walk through each process from start to finish. Make note of the points along the way where users may create new master records, an integration might lead to the automatic creation of a new master record or exceptions may prompt a workaround. These are the points that can affect ERP data quality.

Prioritize the right ERP data

Not every field or record is as important as another. The CFO and the COO should set priorities and the CIO should determine what’s technically feasible and cost efficient.

The CFO should consider financial and compliance exposure. For example, if duplicate payments and tax reporting are a pain point, vendors and payables should be addressed first, but if revenue reporting, credit and collections are the primary challenge, then order-to-cash should take precedence. If financial reporting and month-end closing are problems, then focus first on the chart of accounts and cost centers. Compliance, including privacy regulations like GDPR, “know your customer” for financial services, and tax reporting, should be a consideration across all domains.

The COO should rank data quality concerns by their operational risk. If available-to-promise and inventory planning are major issues, focus on items, BOMs and inventory balances. If service levels are an issue, consider prioritizing customer records and ship-to addresses. If the CFO and COO want to prioritize different topics, focus on shared master records first.

Regardless of where the company starts, make a good-faith effort to quantify the cost of doing nothing. How much extra time does the finance team currently spend on month-end closings? What’s the organization’s duplicate-payment exposure? How much are late shipments negatively affecting customer satisfaction and attrition? ERP data quality benefits the bottom line, so take the time to build a business case for the investment.

Be careful to avoid an “all of the above” approach and address the highest-priority processes early. The highest-priority processes are usually high-change domains such as customers, vendors and inventory items.

Define objectives, then measure the resulting improvements.

Who should be part of data cleanup

While the IT department typically owns the platform and tools associated with ERP data cleanup, the overall initiative must be a collaborative effort between IT, data stewards with domain expertise and the front-line employees who work with the ERP system every day.

IT usually possesses the greatest expertise in data extraction, data quality tools and integrations between ERP and external systems. Data stewards are individuals from each department, such as finance, procurement, sales and operations. They own the definitions, standards and business rules for each data element.

Those first two groups should drive ERP data cleanup, but the front-line employees who work with the ERP system every day are critically important as well. They understand the constraints of the current processes and the workarounds that enable them to operate within those constraints.

Develop and document an improved create-change-retire process for each priority domain and address the workarounds by filling any gaps in the process, then clean up existing records so they meet the new standard.

How to maintain ERP data cleanliness in the future

As a company changes, processes must change as well to fit the new reality.

Cleanup must be followed by operating discipline. Establish data quality metrics and monitor performance in each key domain. Know duplicate rates, completeness of required fields, timeliness of information and the frequency and cost of downstream impact.

Data stewardship must be treated as an ongoing responsibility, not a project line item that ceases to matter after go-live. The CFO and COO should partner with the CIO to continually enforce business process controls, keeping in mind that IT only owns the platform and tools, while domain experts own the business rules and process discipline.

Schedule periodic reviews of KPIs and keep the changing needs of the company in mind. Has the organization opened any new sales channels? Has it expanded products or territory in ways that could affect ERP data discipline? Has the company acquired and integrated another entity? Ask line-of-business employees to document constraints and workarounds and share them with the appropriate data quality steward.

ERP data cleaning benefits are maximized when an organization stops regarding data quality as an IT responsibility. Companies must begin by analyzing business processes, evaluate how poor data quality affects the organization and involve the front-line employees who work with the system every day.

James Kofalt spent 16 years at SAP working with SME business applications and was a product manager for integration technology at Microsoft's Business Solutions division. He is currently the president of DX4 Research, a technology advisory practice specializing in ERP and digital transformation.