Harmonization
The objective of this post is to make clear two concepts of SAP Demand Signal Management (DSIM)
Harmonization engine:
Mapping and Matching and
Attribute Harmonization.
Harmonization refers to how DSIM deals with the problem of storing master data from different origins that refers to the same “physical object”.
Example: suppose we want to load two product masterdata entries into DSIM which refer to the same “physical product”. One entry is originating from our ECC system, the other from a Nielsen data file:
Given that DSIM aims to create one “single version of the truth” out of the supplied internal and external data, the system tries to create unique
Harmonized Records that represent the truth about each of these “physical objects”. It does so trough
Mapping and Matching and
Attribute Harmonization.
Mapping and Matching
Mapping and Matching refers to how DSIM matches the source data (
Source Records, in DSIM terminology) against the
Harmonized Records:
DSIM contains multiple out of the box Harmonization Rules to automate this process (as well as an extensible rules framework to add additional functionality).
For instance, DSIM is able to create one
Harmonize Record and map against it all
Source Records which:
- have the same GTIN/EANUPC (as in the example above)
- have the same Material number OR Description
- have the same Category AND “similar” Description (using fuzzy scoring)
- etc.
Management of DDAs UI – setting up Harmonization by GTIN:
Harmonized Products UI – viewing assigned Source Records:
Attribute Harmonization
Attribute Harmonization refers to how DSIM fills in the attribute values from the
Harmonized Records:
DSIM allows different possibilities to automate the process of filling in these attribute values.
Data Origin Prioritization
In general, DSIM fills in the
Harmonized Records values using the data from the assigned
Source Records. Given that multiple
Source Records may be mapped to the same
Harmonized Record, we need to decide how to prioritize the use of this source data. Configuration possibilities include the following:
- The Source Record values from the Data Origin configured to have highest priority are used to fill in the Harmonized Record
- The non-empty Source Record values from the Data Origin configured to have highest priority are used to fill in the Harmonized Record. For the empty attribute values, the record with the next highest priority Data Origin is used (example above)
- Data Origin prioritization is defined not generally, but per attribute. E.g.: in general, all values are retrieved from the ECC Source Record, but the Category value should be retrieved from the Nielsen data Source Record
Attribute Value Derivation
With Attribute Value Derivation, the DSIM system is configured not merely to copy data from the
Source Records, but to “derive” the data for the Harmonized Records using a set of rules. Possible rules are, for example:
- Dictionary Rule: we want to “re-write” one of the attributes. E.g. : we can set up the system so that every time we receive the Category value Cook from the Source, we rename it to Cookies in the Harmonized Record
- IF THEN Rule: we can define the value of an attribute from the Harmonized Record using multiple attributes from the Source. E.g. : we can set up the system so that every time we receive the Brand value Dream and the Description contains the string Choc, the Subcategory value of the Harmonized Record is Chocolate Bar
Harmonized Products UI – viewing the data in the Harmonized Record: