For the complete documentation index, see llms.txt. This page is also available as Markdown.

Quality & confidence scoring

How Delpha scores Ultimate Parent records: four of its six data quality dimensions — Completeness, Validity, Accuracy, and Consistency — plus where confidence scoring is calculated.

Delpha scores data across six data quality dimensions — Completeness, Validity, Uniqueness, Consistency, Accuracy, and Timeliness. For Ultimate Parent Accounts, four of them apply, automated across CRM, API, and MCP:

  • Completeness: Verifies successful execution of the hierarchy analysis without system processing errors.

  • Validity: Validates input accuracy, ensuring valid website domains match company names.

  • Accuracy: Confirms the assigned parent matches either the primary or a discarded recommendation.

  • Consistency: Enforces strict alignment between the assigned parent and Delpha's primary recommendation.

Key Takeaway: For Ultimate Parent, each dimension is operationalized at record level: while Completeness tracks system-level processing and Validity checks input account details, Accuracy and Consistency measure how closely your current Ultimate Parent assignment matches Delpha's primary recommendation.

Confidence scoring on individual recommendations — the score attached to each recommended parent and how it is calculated — is covered in Understanding the Ultimate Parent Analysis.

1. Completeness

Completeness measures whether Delpha successfully executed its hierarchy analysis on the record without system interruptions.

  • true: The record was processed successfully by Delpha's processing engine.

  • false: A system processing error occurred during execution (unrelated to user input data).

2. Validity

Validity evaluates whether the baseline input data on the Account record is sufficient and correctly formatted for hierarchy matching.

  • Valid: The input data is valid (e.g., a website domain is present and correctly aligns with the company name).

  • Invalid: An input data error is detected (e.g., missing website, or the domain does not match the company name).

  • Unknown: Unknown is set when completeness is false, to avoid penalizing the score

3. Accuracy

Accuracy checks whether the currently assigned Ultimate Parent is a legitimate parent entity identified within Delpha's analysis payload.

  • Ok: The record is Valid AND the current Ultimate Parent matches either the top recommended Ultimate Parent OR any discarded potential parent option.

  • No: The record is invalid, or the assigned Ultimate Parent does not match any recognized recommendation options.

  • Unknown: Unknown is set when completeness is false, to avoid penalizing the score

4. Consistency

Consistency measures strict alignment between the current record assignment and Delpha's primary, top-ranked recommendation.

  • true: The record is Valid, Accurate, AND the assigned Ultimate Parent exactly matches Delpha's primary recommended Ultimate Parent.

  • false: The assigned Ultimate Parent matches a discarded recommendation option instead of the primary choice (or the record is Invalid/Inaccurate).

Quick Reference Matrix

Data Quality Dimension

Evaluates

Status: true

Status: false

Completeness

Engine Execution

Record processed without system errors

Internal Delpha processing error occurs

Validity

Input Data Quality

Website domain exists and matches company name

Missing website or domain mismatch

Accuracy

Entity Legitimacy

Valid record + assigned parent matches a recommended or discarded parent

Assigned parent does not exist in analysis results

Consistency

Primary Recommendation Alignment

Valid & Accurate + assigned parent exactly matches top recommendation

Assigned parent matches a discarded/secondary recommendation

Last updated

Was this helpful?