Fundamental Concepts Related to Matching

A thourough understanding of the fundamental concepts enables you to define an effective matching strategy.

To get us started, here are some fundamental concepts that are helpful to understand or review as we learn about matching.

Operational Behavior of Reltio Matching

As stated, by default Reltio matching operates continuously and in real-time. That is, anytime a record is created in the tenant or updated it is first cleansed in real-time and then processed by the match engine to find matches from the existing set of records in the tenant. (Note: if the real-time match and merge behavior is not desired, it can be disabled and instead explicitly executed as a batch job on-demand. See Tenant Level Match Strategy)

Survivorship is Unrelated to Merging

Unlike other systems, Reltio Survivorship has nothing to do with merging. for more information, see Attribute-Level Survivorship.

Entity Types and Rules

Each entity type (Contact, Organization, Product) defined within your tenant’s metadata configuration has its own set of match groups. Each match group holds a single rule along with other properties that help dictate the behavior of the rule within the match group. A rule is composed of Comparison Operators, such as Exact, ExactOrNull, and Fuzzy and attributes you select from the entity type definition.

Within each of Reltio’s OOB solution accelerators, a best-practice set of match groups are already defined and included. They can be removed, modified, and supplemented with additional ones of your design.

Match Tokens and Comparators Within a Match Group

  • Match Tokens are used to help the match engine quickly find candidate match values from potentially millions of other profiles.
  • The comparison formula within the match rule is used to adjudicate a candidate match pair, and will evaluate to true or false (or a score if using relevance based matching).

Outcome of a Match group

  • For each candidate-pair of profiles presented to a match group, the match group will do one of three things with the pair as a result of how the match group evaluated the level of match for the pair.
    • Nothing (if the comparison formula evaluates to false)
    • Issue a directive to merge the pair
    • Issue a directive to queue the pair for review by a data steward

Cleansing While Matching

Separate from the cleansers that you may have configured within your tenant (typically address, email, and phone), Reltio offers an out-of-the-box (OOTB) set of additional match-level cleansers that can be used exclusively by the match engine, on-the-fly during matching.

Internal vs External matching

As mentioned earlier, the match rules you develop can be used to match and merge records within a tenant. You can also match records held within an external file against records within your tenant. This occurs in a non-invasive manner and the results of the match job are written to an output file for you to review. External matching can be achieved programmatically via the External Match API. The functionality is also available through the External Match Application found within the Reltio Console. The SCOPE parameter controls Internal vs External matching.

Matching of Data Tenant Records to your Tenant Records

The match rules you develop for your tenant are not used by DTSS to match Data Tenant records to your tenant records. The rules used by DTSS are held in the Data Tenant’s configuration. If your tenant subscribes to a Reltio Data Tenant, your DTSS subscription may modify the behavior of the Data Tenant’s match rules but not significantly.