MRO Material Master Description Analysis Study

An SPC Results study of more than 700K material master descriptions from multiple Energy related MRO material Catalogues

Target Audience: Energy Industry (O&G, Pipelines and Utilities) Business Process Owners & Leads (SCM & Maintenance), Material Master Administrators, ERP Support Managers and influencers

Short descriptions from Oil & Gas and related industry MRO material catalogues were analysed for their usability in searches in support of ERP business processes. The analysis included 714,588 descriptions from six unique catalogues. All the catalogues had similar development histories. The catalogues were compiled over the course of multiple events including ERP implementations where several local site catalogues were centralized and amalgamated, corporate merger projects where the acquired companies MRO catalogue items were merged, and ongoing business operations where new items were added through administrative processes. Rigorous description standards are not apparent in any of the catalogues studied. In short, the descriptions studied are likely a representative sample of much of the industry as most have experienced a common evolution of their material master data records.

Information on the assets and business system landscapes of the companies studied are in note 1.

Key Findings

  1. The most fitting characterization of the item descriptions reviewed is they are unstructured and non standardized. This makes it difficult to search, compare and identify what each item represents and causes negative outcomes.
  2. The frequency distribution of the description key word is very skewed.  The description key word is also known as the description noun or class identifier. A tiny percentage of all nouns are used very frequently, followed by a very long tail of very low usage nouns. Here are samples to paint the picture of the frequency skewedness:
    • Gasket was the most frequently used noun, representing 5.21% of all items
    • The 10 most frequently used nouns were used in 25.02% of all items
    • The 40 most frequently used nouns accounted for 50.22% of all items
    • The top 500 most frequently used nouns were used in 89.9% of all items
    • The top 1000 most frequently used nouns were used in 94.19% of all itemsThe 13,595 least frequently (all after the top 1000) used nouns were used in 5.81% of all items.
    • The most frequently used noun is used about as frequently as the least used 93% of nouns (approximately 5% in both cases)
  3. This paints a frequency distribution profile that is highly skewed to a small number of frequently used nouns with an extremely long tail of low occurrence nouns.  This causes problems for both the frequently used nouns and the sparsely used nouns. The frequently used nouns rely on their associated attribute values to distinguish one item from another. Since there is neither a standard set of attributes nor order of them per noun, finding the right item for commonly used nouns is highly challenging. For the infrequently used nouns, the searcher’s challenge is to guess the name of the noun that was used to identify it.
  4. Reviewing the 10 highest frequently used nouns, there was limited commonality of the attributes used in the descriptions, and no common order in organizing the attributes. This shows the randomness in the attributes and the order they are presented. Partially offsetting this observation is the fact that the items came from six independent catalogues, with no expectation of common standards between companies.
  5. Short description field lengths were analysed in three sample SAP material catalogues containing 350,000 items.  A common assertion is that the 40-character short description field size is a significant barrier to identifying items. It was found that 7% of items used the full 40-character description.  Perhaps more meaningful is 26% of items had 36 or more characters. This suggests that about a quarter of items have descriptions that may be impacted by the field length. It points to the value of consistent use of abbreviations and punctuation in short descriptions.

Analytical Conclusions

  1. The extremely large variations in primary identifiers, attributes and random ordering of description elements provides the evidence to confirm repeated anecdotes with client’s personnel that searches in MRO material catalogues are challenging in the absence of structured consistent descriptions and a highly significant productivity impediment in ERP materials related business processes. 
  2. The business processes effected by the material master search challenge are central to the safe and effective operations of the operator.
    • The involved processes for the SCM function are:
      • purchase requisitioning to payment processes
      • sourcing and contracts processes
      • all inventory and physical inventory processes
    • The involved processes for the Maintenance function are:
      • work planning
      • job scheduling
      • maintenance execution.
    • The involved processes for the Projects function (for projects executed with ERP business processes) are:
      • materials planning
      • project execution
  3. Applying a suitable Material Description Taxonomy that covers a high majority of operations and maintenance required material masters will improve findability and user confidence in your business processes. 
  4. The nouns, modifiers, attributes in a set order effectively creates the meta data for structured standardized material descriptions.
  5. A material description taxonomy of between 500 and 1000 nouns and modifiers (class identifiers) will be suitable to identify 90 to 95% of the materials items needed.
  6. As a supplemental point, a question I have had on several occasions from involved SMEs is why not just make use of maintenance BoMs to bypass the low-quality description problem? BoMs give a direct connection to the equipment or functional location the material supports bypassing the material master search challenge. My best answer was found in one company’s attempt at a BoM creation project that failed after being well over schedule and over budget. The low quality of the source data made constructing the BoMs problematic, several variations of a BoM could be built from the indeterminate items, exactly the same problem other business users had with the material masters.  

What the Data Showed?

The MRO material master catalogues of six companies with 714,488 material short descriptions were analysed.

The benchmark elements of a structured material description are:

  • A description noun also known a class identifier that is the primary identifier of the nature of the subject item
  • A modifier term that where needed narrows the scope of the primary noun allowing standard unique templated lists of attributes per noun modifier combination that are useful and meaningful to identify an item
  • A set of identifying attributes in a fixed order moving from most to least significant to complete the standard template to describe this class of materials
  • Standardized use of abbreviations and units of measure to provide consistency

When separated out, the descriptions have 14,595 unique class identifiers.

The usage frequency of description nouns

The usage frequency of description nouns:

Description Length

Short description field lengths were analysed in three sample SAP material catalogues containing 350,000 items.  A common assertion is that the 40-character short description field size is a significant barrier to identifying items. Only 7% of items reviewed had a full 40-character description.  Perhaps more meaningful is 26% of items had 36 or more characters.

Note: The larger catalogue study included one non-SAP catalogue and it allowed longer descriptions than 40 characters. Of interest there was on one item that exceeded 40 characters (perhaps that catalogue was previously in an SAP system?)

Notes:

  • Overview of the Companies studied:
    • The companies whose catalogues were reviewed had the following asset and MRO materials characteristics:
      • They are asset intensive
      • Their O&M costs and the success of their operations are highly influenced by the effective integration of Supply Chain and Maintenance personnel and business processes
      • Some included Drilling and Completions project materials, all catalogues included operating assets like oilsands, refineries, pipelines or offshore platforms
      • The plants involved within the companies had both inventory supported operations, nonstock material processes and mixed use of the two methods
  • The companies whose catalogues were reviewed had the following ERP characteristics:
    • They are multi plant, multi company code enterprises
    • All used a single production client ERP for their business processes. (The MDM benefit of distributing common material masters to multiple productive systems is not significant)
    • Operations may or may not include multiple countries and purchasing organizations
  • Their MRO material catalogues consist of spare parts, assemblies (actual assemblies not IBAU type materials), equipment, bulk O&M materials, operating supplies. Materials may have single or multiple valuation, be batch or non batch managed, be non serialized or serialized, be repairable items or not and be capital or noncapital spares. (Typical SAP material types are HIBE operating supplies, ERSA spare parts, with smaller numbers of NLAG non-stock, FHMI production resources – tools, UNBW non valuated materials)
  • The terms Noun and Modifier, and Class Identifier are used interchangeably in this analysis. Both refer to the primary object descriptor used to identify a material in a material master description. Class identifier is used where Classification functionality is used in support of material master identities. An appropriate class is assigned to a new record, and the class has the appropriate characteristics to assign values to, uniquely identifying the item (and allowing meaningful duplication checking). Often an automated program generates short and long descriptions from the item’s classification data.
  • There were 14,595 unique nouns (or noun phrases) in the catalogues analysed. 5 of the 6 catalogues were formatted allowing the noun to be broken out (comma delimited for example). The last file was supplied with manual work done to normalize the nouns and move them to the beginning of the description. The long tail of nouns used just one time reflects the lack of structure in identifying the items