Your ERP Data Is an Asset. Start Treating It Like One.

The Asset Nobody Maintains

Walk through any well-run plant and you will see the evidence of maintenance discipline everywhere. Equipment has owners. Preventive maintenance runs on a schedule. Spare parts are stocked. Downtime is tracked and its causes are analyzed. Everyone understands that a machine left alone degrades, and that a degraded machine costs money.

Now open the item master in the same company's ERP. You will often find thousands of records nobody owns, descriptions entered three different ways, units of measure that disagree with how the part is actually bought, lead times set during implementation, duplicate suppliers with slightly different names, and bills of material that no longer match what the floor builds. The data degraded just like an unmaintained machine would. The difference is that nobody was assigned to notice.

Gartner has estimated that poor data quality costs organizations an average of $12.9 million a year.

Gartner has estimated that poor data quality costs organizations an average of $12.9 million a year. A mid-market manufacturer will not have a number that large, but the mechanism is the same: wrong data produces wrong plans, wrong purchases, wrong quotes and wrong shipments, and each one costs something.

Where Bad Data Costs a Manufacturer

The costs are rarely labeled as data problems. They show up as operating problems.

A wrong lead time produces a late order and an expedite fee. A wrong unit of measure produces a purchase of ten boxes instead of ten each, or the reverse. A bill of material that is missing a component produces a shortage discovered at the work center. A duplicate supplier splits your spend in two and weakens your negotiating position. A product with the wrong weight or dimensions produces freight charges you cannot recover. An incorrect standard cost produces a quote that loses money.

Each of these is a small event. Added up across a year, they explain a meaningful share of the expediting, overtime, write-offs and margin leakage that every plant assumes is simply the cost of doing business.

Why It Is About to Matter More

There is a second reason to care about this now. The industry is moving toward analytics, machine learning and automation that consume ERP data directly, without a person in between to notice that something looks wrong.

A planner who sees a lead time of zero days knows it is a mistake. A forecasting model does not. A person reading a bill of material can spot a missing part. An automated scheduling tool will happily plan around it. Every technology on the horizon that promises to make decisions faster, from demand sensing to predictive maintenance to automated purchasing, assumes that the data it is fed is right. When it is not, those tools make wrong decisions at machine speed.

Companies that clean up and maintain their data now will be able to adopt those tools quickly and trust what they produce. Companies that do not will spend the first year of every new technology project discovering that their data is not ready.

Give the Data Owners

The fix begins the same way maintenance discipline does: with ownership. Every major data domain needs a named owner who is accountable for its accuracy. Engineering owns bills of material and routings. Purchasing owns supplier records and purchasing lead times. Planning owns safety stocks and planning parameters. Finance owns costs. Sales owns customer records.

Ownership means more than a name on a chart. It means the owner decides the rules for that data, approves changes to it, and receives a regular report of exceptions to correct.

Measure It Like Uptime

What gets measured gets maintained. A small set of data quality measures, reviewed monthly, changes behavior quickly. The percentage of active items with a lead time that matches recent actual receipts. The number of duplicate supplier and customer records. The share of bills of material changed in the last year without a corresponding engineering change. The count of items with no movement in two years that are still active.

None of these requires new software. Most can be produced from the ERP with simple queries. What they need is someone to review them and the authority to fix what they reveal.

Clean Up Before You Build Anything New

When a manufacturer is planning an ERP upgrade, a new analytics platform or its first automation project, the data cleanup should come first, not as a migration task squeezed into the final weeks. Clean data makes every subsequent project faster, cheaper and more trustworthy. Dirty data makes every one of them slower and less credible, and it is the most common reason those projects disappoint.

The Bottom Line

ERP data is a production asset, and like every production asset, it degrades without maintenance. Bad data already costs manufacturers in expediting, write-offs and lost margin, and it will cost more as analytics and automation begin acting on it without a human check. The discipline is familiar: named owners, clear rules, regular measurement and cleanup before any new project begins.

At Cherry Street, this is the work we do. We help manufacturers assess the condition of their ERP data, assign ownership, build the measures that keep it accurate, and clean it up before it becomes the foundation for something new.