How long should a part number be? It sounds like a simple question with a simple answer. But ask a procurement manager, a database administrator, and an engineer at three different companies, and you will likely get three different numbers. One says 15 characters. Another swears by 18. A third insists 25 is the safe ceiling. None of them is wrong, and none of them is right in any universal sense.
That is the strange reality of part number width specifications. Unlike barcodes or product classification codes, manufacturer part numbers have no single governing authority setting their length. The width is whatever an organization decides it should be, shaped by legacy systems, industry habits, and the limits of the software that stores the data. Understanding why that is true, and how to work with it, saves teams from a surprising amount of pain.
The Core Truth: A Part Number Is a Business Identifier, Not a Regulated Field
A part number is a label a company assigns to an item so that it can be tracked, ordered, and referenced. It is not a public standard in the way a postal code or a currency code is. Because of that, no international body mandates how many characters it must contain.
This distinction matters more than it first appears. Many people assume that somewhere, in some standards document, the correct width is written down. In practice, widths are set by one of two things: an organization’s own numbering scheme, or a specific industry standard that happens to define an identifier. The confusion arises because these two categories get mixed together in everyday conversation.
Where Fixed Widths Actually Do Exist
Some identifiers genuinely have a fixed width. These are worth knowing because people often mistake them for part number rules.
| Identifier | Fixed width | What it identifies |
|---|---|---|
| GTIN-8 | 8 digits | Small retail products |
| GTIN-12 | 12 digits | Trade items (North America) |
| GTIN-13 | 13 digits | Trade items (international) |
| GTIN-14 | 14 digits | Cases and cartons |
| UNSPSC code | 8 digits | Product classification |
The GTIN family is a genuine width specification, but it applies to trade item numbers, not to manufacturer part numbers. UNSPSC codes are likewise fixed at eight digits, structured as four two-digit segments, but they classify products rather than identify a specific maker’s item. Teams that quote these numbers as proof of a “part number standard” are conflating two different things.
What the Standards Bodies Actually Say
Several standards address identification data but set no universal width. ISO 22745 defines the structure and exchange of identification data for items, including how identifiers such as part numbers appear in item records. It governs how the data is structured and shared, not how many characters a part number should contain.
Similarly, IEC 61360 and ISO 13584, often known together as PLIB, provide dictionaries of properties for industrial and electronic component data. These define attribute types and lengths per property rather than one global rule. ECLASS, a widely used classification and property standard for products, follows the same logic: its property definitions include data types and recommended lengths for individual identifiers. Version 10.0 arrived in 2014, and later versions have followed.
The pattern is consistent. Standards define how to describe an identifier and what data type it uses. They do not hand down a magic number for part number width.
Why ERP and PDM Systems Drive the Real Limits
If standards do not set the width, what does? In most organizations, the answer is the software. ERP and PDM systems commonly impose their own field limits, often somewhere around 18, 25, or 40 characters. These limits are vendor- and configuration-specific, so any specific figure should be treated as unverified until someone checks the actual system configuration.
This is where practical trouble begins. A company migrates from an old ERP system to a new one, discovers the part number field is shorter than before, and suddenly has thousands of records that no longer fit. Truncation creeps in. Two different parts collapse into the same identifier. Orders go to the wrong item. The root cause is rarely the standard, because there was never a standard. The root cause is a field limit nobody checked before the migration.
One practical habit is to treat the shortest field across all your systems as the true ceiling. If your warehouse scanner, supplier portal, and legacy reporting tool accept fewer characters than your new ERP, the smallest limit wins by default.
The Myth of the Universal 15-Character Rule
Claims that part numbers must be 15 characters, or 18, or any other specific figure, circulate widely in procurement and ERP circles. They are configuration conventions dressed up as rules. No standards body mandates a universal width, and treating these claims as law causes real problems.
The most common mistake is designing a numbering scheme around a number someone heard in a meeting. A team decides on 15 characters because “that is the standard,” then discovers a supplier whose catalog numbers run to 22 characters. Now the supplier’s items need internal aliases, and the alias table becomes its own maintenance burden. The scheme was built on a myth, and the cleanup takes years.
What actually works is simpler and less glamorous. Document your own width limit explicitly. State it in your supplier onboarding materials. Build validation into the data entry screen so nobody can save an overlong number by accident. And when someone cites a “standard,” ask which document they mean. More often than not, the answer is a memory of a previous employer’s system.
Designing a Width That Survives Contact With Reality
If a team is building a numbering scheme from scratch, the width decision deserves more thought than it usually gets. A few practical guidelines hold up well.
- Audit every system that will touch the number. Include the ERP, the PDM or PLM tool, the warehouse management system, the e-commerce catalog, and any supplier portals. The shortest limit becomes your practical maximum.
- Add headroom for growth. A scheme that exactly fills the available field today will break the first time a new product line needs a longer prefix. Leave a few characters of slack.
- Reserve space for the unpredictable. Supplier part numbers, customer references, and revision suffixes all compete for the same field. Decide early whether these live inside the part number or in separate attributes.
- Avoid embedding meaning that will expire. Smart numbering schemes that encode plant, year, or product family look elegant until the plant closes or the family gets reorganized. Meaningful segments age badly.
- Test with real data, not invented examples. Pull a few hundred actual part numbers from your catalog and run them through the proposed rules. The edge cases surface fast.
For teams comparing how different property standards handle identifier lengths, the variation across classification systems is instructive. The same discipline applies to property data in residential property listings, where fields like apartment sizes follow their own local conventions rather than one global rule. The lesson transfers: local convention beats imaginary universal standards every time. (Modern apartment interiors)
Where Part Number Width Meets Real-World Data Exchange
The width problem gets sharper when data crosses organizational boundaries. A manufacturer sends a product data sheet to a distributor. The distributor’s system accepts 20 characters. The manufacturer’s part numbers run to 24. Something has to give, and it is usually the part number, truncated in a way that breaks the match.
Identification-data standards were designed to address this scenario, even though they stop short of fixing a width. The goal is structured, exchangeable item data so that both parties describe the same item the same way. When organizations skip that discipline and rely on informal conventions, the width mismatch becomes a silent data quality problem that surfaces only when an order fails.
A useful practice is to maintain a canonical part number alongside any shortened display version. The canonical number is the source of truth. The display version exists only for systems that cannot handle the full length. Keeping these separate prevents the slow erosion that happens when truncated numbers quietly become the real ones.
Common Mistakes That Cost Teams Time
Beyond the myth of a universal rule, a handful of errors show up again and again.
- Assuming the longest field is the constraint. It is the shortest one that matters, and it is often buried in an old reporting tool nobody remembers.
- Ignoring leading zeros. Many systems strip them, which silently changes an identifier’s width and breaks matches.
- Mixing classification codes with part numbers. A UNSPSC code and a manufacturer part number serve different purposes and follow different rules.
- Forgetting about revisions and suffixes. A part number that fits perfectly today may not fit once a revision letter is appended.
- Skipping documentation. Undocumented width rules live only in the heads of long-tenured staff, and they leave when those staff do.
Each of these is avoidable with a short written policy and a validation rule in the data entry system. The cost of prevention is a few hours. The cost of cleanup is measured in months. (Forena apartment location)
How to Audit Your Current Part Number Width
An audit does not need to be elaborate. Start by exporting every part number from every system that stores one. Measure the actual lengths and look at the distribution. A histogram usually reveals a cluster of normal lengths and a long tail of outliers, and the outliers are where the trouble hides.
Next, identify the field limit in each system and compare it to the longest real number. Any system whose limit sits below the longest real number is a truncation risk. Finally, check for duplicates that appear only after truncation. If two distinct parts become identical when shortened to the smallest field limit, you have found a live data integrity problem.
Running this audit once a year catches drift before it becomes an incident. It also gives teams concrete evidence when arguing for a field expansion in a system upgrade.
Conclusion
Part number width specifications are less a standard than a negotiated reality. No authority sets a universal length, and the numbers people quote in meetings are usually inherited conventions rather than documented rules. Fixed widths do exist for GTINs and UNSPSC codes, but those identify trade items and product categories, not manufacturer part numbers. For everything else, the width is whatever your organization and its systems can agree on.
The practical path forward is unglamorous but reliable. Find the shortest field across your systems. Document it. Validate against it at data entry. Keep a canonical number separate from any shortened display version. When someone cites a standard, ask to see the document. Without a universal rule, clear local practice is the only thing that keeps the data clean. Teams that treat width as a design decision rather than an inherited accident avoid years of quiet, expensive cleanup.
Frequently Asked Questions
What is the standard character width for a part number?
There is no universal standard width. Part numbers are business identifiers, and their length is set by each organization’s numbering scheme or by the limits of its software systems. Any specific number quoted as “the standard” is almost always a configuration convention rather than a documented rule from a standards body.
Do GTIN and UNSPSC codes have fixed widths?
Yes, but they are not part numbers. GTIN-8, GTIN-12, GTIN-13, and GTIN-14 are fixed at 8, 12, 13, and 14 digits respectively, and they identify trade items. UNSPSC codes are fixed at eight digits and classify products. Neither governs the width of a manufacturer part number.
Why do ERP systems impose their own part number limits?
ERP and PDM systems are built with database field limits that vary by vendor and configuration. These limits are practical constraints of the software, not reflections of an industry standard. Because they differ between systems, the shortest limit across all your systems effectively becomes your real maximum.
What happens if a part number is longer than the system allows?
The number typically gets truncated, and truncation creates two problems. First, the shortened number may no longer match the original, breaking references across systems. Second, two distinct part numbers can collapse into the same shortened string, causing orders to go to the wrong item. Both problems are silent until something fails.
How much headroom should a new numbering scheme include?
Enough to absorb new product lines, supplier catalog numbers, and revision suffixes without hitting the field limit. A common approach is to reserve a few characters beyond the longest number you expect today. A scheme that exactly fills the available space will break the first time requirements grow.
How can a team check whether its part numbers are at risk?
Export part numbers from every system that stores them and compare the longest real values against each system’s field limit. Any system whose limit sits below the longest real number is a truncation risk. Then check whether any two distinct parts become identical when shortened to the smallest limit, which reveals live data integrity problems.
