We request our prospecting firm to send us few lines of sample data for us to do the evaluation and process it. During the evaluation process, we would analyse the input data in terms of precision, completeness and standardization. A customized cleansing process is defined based on the initial data analysis.
Materials are sourced from the respective Manufacturer’s / Vendor’s / Supplier’s websites or
catalogues to identify the material and its specifications concisely. The materials are sourced as per the hierarchy given below:
- MANUFACTURER: The first step is to try to source the material information from the actual Manufacturer’s website to obtain authenticated, reliable and precise results.
- VENDOR: Suppose the product listing or required information is not present on the Manufacturer’s website then the next logical step is to source the material information from the Vendor’s websites.
- SUPPLIER: If the product information is not found either in the Manufacturer’s or Vendor’s website, then the next logical step is to source the material information from any Supplier’s
websites.
- NOT FOUND: When material details or any relevant information is absolutely not found even after an exhaustive search, we conclude that the material is Not Found. For not found items,
attribute values are captured from the client’s input data (Legacy data)
Many a times, Vendor details are tagged as Manufacturer details. As part of the process, we will validate and identify the exact manufacturer details. Example:
LEGACY MATERIAL DATA
3VX71, Grainger, 24x1/2 Belt
CLEANSE/ENRICHED & CLASSIFIED MATERIAL DATA
MANUFACTURERE: NORTON
MANUFACTURERE PART NUMBER: 78072728564
VENDOR: GRAINGER
VENDOR PART NUMBER: 3VX71
Manufacturer and Vendor names are standardized as per the legal name of the entity. Part numbers
are validated and standardized as per the exact availability on the website/catalogues. Example: The
following names are standardized to Allen-Bradley
- ALLEN-BRADLEY
- ALLEN BRADLEY
- A&B
- A-B
- A/B
The following part numbers are standardized as to one format as given in Allen-Bradley
| Legacy Part numbers |
Standardized Part numbers |
| 440LR4K1600YD |
440L-R4K1600YD |
| 440LT-4K1600YD |
440L-T4K1600YD |
| 440LT-4K1600YD |
440L-R4K0320YD |
| 440.LT4K0320YD |
440L-T4K0320YD |
| 440.LT4K.0640.YD |
440L-T4K0640YD |
Materials are identified by defining the Noun Modifier. Noun modifiers are defined based on legacy
descriptions, and/or based on external source information. Example:
| Legacy description |
Noun, Modifier |
| DODGE BALL INSERT 4 "# 037645
SCREEN FOR CH5 DEINKING |
BEARING, ROLLER |
| SHOCK # MAS3325 SA 1/2 x 1 ACE
CONTROLS FOR WINDER LINE CHECK |
ABSORBER, SHOCK |
| 435Y89 48x16 SANDING BELT |
ABRASIVE, BELT |
Currently SoftNis has more than 10,000 unique Noun modifier combinations in the dictionary. The
Noun modifier dictionary is a two-tiered classification schema with consistent naming convention.
Each Schema consists of Noun/Class and Modifier/Sub-Class with associated characteristics that
describes the general part attributes.
GAUGE
MEASURING RANGE
PROCESS CONNECTION SIZE
CASE MATERIAL
STEM MATERIAL
FILL FLUID
TEMPERATURE RATING
CERTIFICATION / STANDARD
ADDITIONAL FEATURES
As per the Attributes defined for a Noun-Modifier, values are captured for each material from respective input
description or from source (if).
Referring to the Standard Units (SI units) and Military Standards (MIL-STD-12D), we selected the
most used units of measurement abbreviations in the industry. Example:
| TYPE |
UoM |
SoftNis STANDARD LIST OF
ABBREVIATIONS |
| ANGLE |
DEGREE |
DEG |
| ANGLE |
RADIAN |
RAD |
| LOAD |
KILOGRAM |
KG |
| AREA SQUARE |
CENTIMETER |
CM2 |
For each material Long and short descriptions are generated from the Standardized and Enriched data. Short descriptions were abbreviated, and length was truncated to 40 chars to fit in to SAP’s short text. Format of the descriptions are customized as per the unique customer requirement.
Throughout the data cleansing/enrichment process, the following two types of duplicates items are identified.
Exact Duplicates: Exact duplicates are the ones for which manufacturer’s name and part Number are exactly the same.
Function equivalent Duplicates: Function equivalent duplicates are the ones, for which manufacturer’s name and/or part number are different but having the same specifications, like size, material, and other specifications.
As per the client requirement, materials are classified to any standard taxonomies like UNSPSC, ECLASS, CPV or any custom build taxonomies. UNSPSC / ECLASS consists of four levels of hierarchy. A Complete 4 level hierarchy codes with descriptions are provided. Example:
During the process of enrichment if we come across any material which is marked as obsolete, such parts are tagged as “Obsolete”. If there is any alternate part number suggested in the manufacturer’s website then the same will be captured.
A list of materials which could not be identified are marked and submitted to the client for physical verification. These materials can be processed after receiving the required additional information.
| Material Number |
Description |
| F2_17964 |
#20301 UNIVERSAL REF |
| F2_17955 |
#213968,two 5309M,MR |
| F2_17948 |
#2154 MULTI-MATERIAL |
Data Evaluation and PoC
