Every meaningful model begins with an act of understanding, and this happens before structure, schemas, diagrams, or formal rules.
In the Denodo Platform, this act takes shape within the semantic layer, where data is not merely connected but interpreted. Logical views, business definitions, and relationships are the result of listening to a domain and translating its language into a shared form that both humans and systems can inhabit.
Semantic modeling, in this sense, is never neutral; its role is to interpret, to decide what deserves a name, what distinctions matter, and how reality should be expressed. But because the world being modeled continuously evolves, understanding cannot be fixed. The semantic layer must remain open, revisable, and responsive, or it risks becoming a rigid structure disconnected from the domain it claims to represent.
As Denodo’s semantic models become foundational to analytics, data services, and AI-driven applications, a crucial question inevitably arises: how can we know whether the model still understands the world it represents? Measurement enters here not as reduction, but as observation, a way to verify that meaning remains alive within structure.
The Paradox of Measuring Meaning
At first glance, measuring meaning appears contradictory, because meaning is contextual, relational, and fluid, while measurement suggests precision and control. Yet within Denodo, measurement does not aim to quantify meaning itself, but to observe the conditions that enable meaning to remain coherent and governable once formalized in logical views.
Without such observation, semantics can drift. Definitions embedded in views may lose clarity, relationships may detach from business intent, and AI or analytics layers may continue to operate on technically valid but semantically weakened structures. Measurement, in this context, becomes a practice of attention, making visible what would otherwise remain implicit, allowing misalignment to be detected before it propagates downstream.
Metrics, therefore, do not claim truth, instead they provide perspective, enabling the semantic layer to be continuously assessed as a living interface between data and understanding.
From Reflection to Practice – Classes of Semantic KPIs
If measuring meaning is an act of care rather than control, then the indicators we adopt must reflect this intention.
Rather than a flat list of metrics, it is more useful to think in terms of classes of key performance indicators (KPIs), each illuminating a different dimension of how meaning behaves once formalized within Denodo’s semantic layer.
Taken together, the following classes of KPIs do not attempt to measure meaning itself. Instead, they make visible the conditions that enable meaning to remain coherent, complete, and faithful to reality as it is formalized, governed, and continuously reinterpreted within Denodo’s semantic layer.
These include:
- Completeness
- Quality
- Usage
- Stability
- Cultural Maturity
For each, I’ll provide an explanation and a few examples of KPIs.
Completeness as Semantic Coverage
Completeness is often mistaken for exhaustiveness. Yet in Denodo, a semantic model is not complete because it exposes all available data, but because it captures what is meaningful for the domain and the use cases it serves. Completeness emerges when logical views represent the essential concepts, relationships, and distinctions needed to reason about reality without overwhelming users with unnecessary detail.
Observing completeness means assessing whether key business concepts are consistently modeled across views, whether relationships reflect real dependencies rather than technical joins, and whether the semantic layer avoids both gaps and redundancies.
A complete model is one that provides sufficient expressive power while preserving clarity and navigability, and in this balance between richness and restraint, completeness becomes a form of semantic precision rather than mere coverage.
Completeness KPIs – Representing the Domain
This class of indicators observes whether the semantic model adequately represents the essential structure of the domain. It focuses on conceptual coverage and semantic balance, helping to reveal gaps, overlaps, and blind spots that may weaken the model’s expressive power.
Typical examples include:
- The ratio between modeled business concepts and those identified during domain elicitation
- The average number of meaningful relationships per concept, as a proxy for semantic richness
- The percentage of concepts affected by unresolved ambiguities, duplicates, or inconsistent definitions
- The degree of alignment between semantic views and existing business or regulatory glossaries
Quality as Coherence and Memory
Quality within Denodo’s semantic layer is expressed through coherence. Definitions must remain consistent across views, calculations must be transparent, and meanings must remain stable even as implementations evolve. Logical views act as contracts of meaning, and their quality determines whether trust can be sustained across consumers and applications.
A high-quality semantic model also preserves memory, and Denodo’s metadata, lineage, and documentation capabilities enable each concept to retain its origin, rationale, and evolution, over time.
This memory is not ancillary, it is essential; without it, the model risks becoming opaque, and understanding dissolves into implicit assumptions.
Measuring quality, therefore, means assessing whether the semantic layer remains intelligible to those who use it and maintain it, and whether it can evolve without losing coherence.
Quality KPIs – Preserving Coherence and Intent
Quality-related indicators examine the internal consistency and clarity of the semantic layer. They reflect the model’s ability to remain intelligible, trustworthy, and explainable as it evolves.
Representative examples include:
- The percentage of logical views and concepts with documented definitions, lineage, and ownership
- The consistency of calculations and business rules across different views
- The frequency with which semantic definitions are reviewed or updated
- The proportion of automated integrity and coherence checks successfully passed
Usage as Shared Language
A semantic model reaches its full meaning only when it is used.
In Denodo, usage manifests through the consumption of logical views by analysts, data scientists, applications, and AI systems. Each query is an act of interpretation, where each reuse is a sign of shared understanding.
Usage is not a measure of popularity, but of resonance. When users rely on Denodo’s semantic views to explore data, explain outcomes, and support decisions, they express epistemic trust, and when they bypass the semantic layer or recreate logic elsewhere, it signals a fracture in shared meaning.
Observing usage enables organizations to understand whether the semantic layer truly functions as a common language rather than merely an intermediate technical artifact.
Usage KPIs – Measuring Shared Meaning
Usage KPIs observe how the semantic model is actually consumed and whether it functions as a shared language rather than a technical abstraction.
Examples in this class include:
- The number and diversity of users actively querying semantic views
- The reuse rate of logical views across reports, applications, and AI workloads
- The average time required for non-technical users to correctly understand and use a view
- Qualitative feedback on clarity, trust, and perceived representativeness of the model
Stability as Governed Evolution
Stability does not mean immobility, because domains evolve, regulations change, and business strategies shift. The semantic layer must adapt accordingly, and stability lies in the ability to evolve in a governed manner, preserving meaning while integrating change.
Denodo’s versioning, impact analysis, and dependency management capabilities support this continuity. Measuring stability means observing the rhythm of change, the propagation of updates across dependent views, and the absence of semantic regressions that could undermine trust.
A stable semantic model is one that changes without disorienting its users, maintaining continuity even as structure evolves.
Stability KPIs – Governing Change Over Time
Stability KPIs address the temporal dimension of meaning, observing how the semantic model evolves, and whether the changes preserve continuity rather than introducing confusion.
Typical indicators include:
- The frequency and magnitude of changes to concepts and relationships over a given period
- The average time required for semantic updates to propagate across dependent views
- The number of regressions or inconsistencies introduced by model revisions
- The stability of key business definitions across versions
Cultural Maturity and Semantic Governance
Beyond technical correctness lies cultural maturity.
A Denodo semantic model becomes truly effective when governance is not merely enforced but internalized, when definitions are discussed rather than imposed, when semantic decisions are shared rather than centralized, and when language becomes a collective concern.
Cultural maturity emerges through active stewardship of the semantic layer, through roles dedicated to semantic governance, shared glossaries, and recurring moments of alignment. These elements resist strict quantification, yet their presence can be observed through organizational behavior and adoption patterns.
At this stage, the semantic layer ceases to be just an architecture and becomes a space in which the organization reflects on itself and negotiates meaning.
Cultural Maturity KPIs – When Semantics Becomes Practice
The final class concerns cultural maturity, the extent to which semantics has become embedded in organizational behavior rather than remaining a purely architectural concern.
Examples include:
- The presence of formally assigned roles for semantic governance and stewardship
- The existence and active maintenance of a shared enterprise glossary
- The frequency of cross-functional alignment sessions focused on meaning and definitions
- Narrative indicators drawn from real cases where the semantic layer enabled shared understanding or resolved ambiguity
Toward an Ethics of Measurement
Measuring a semantic model is an ethical act, and every metric expresses a judgment about what deserves attention. In Denodo, where the semantic layer mediates access to data, analytics, and AI, this responsibility is amplified.
Metrics should not be instruments of control, but tools of care. They should preserve openness rather than enforce rigidity. Used responsibly, they keep the semantic layer aligned with reality, resilient to change, and faithful to meaning.
Measurement does not close interpretation. It keeps it alive.
The Measure of Meaning
To measure a semantic model within Denodo is to acknowledge that meaning, once formalized, remains fragile. Logical views, governance rules, and metadata structures do not replace understanding. They sustain it.
Metrics become gestures of attention, enabling the semantic layer to continue functioning as a coherent, complete, and accurate representation of the world it describes.
The true measure of meaning lies not in precision alone, but in the model’s ability to remain faithful to reality while the world – and the data that describes it – inevitably changes.
He is passionate about philosophy, which he considers an indispensable discipline for those wishing to deal with Artificial Intelligence. He also deals with non-verbal communication, offering specialized seminars and advice (https://www.decorporisvoce.com/) and, in his spare time, is a Food Blogger (https://www.trapignatteesgommarelli.com/).
- Why Trustworthy AI Requires Semantic Transparency - June 4, 2026
- The Semantic Nervous System – How Cognitive Enterprise Architecture Transforms Fragmented Organizations into Living Intelligent Systems - May 28, 2026
- The Measure of Meaning – How Semantic Models Stay Faithful While the World Changes - May 8, 2026
Very interesting blog and easy to read. The explanation about semantic models and how they stay useful even when data changes was really good. I learned something new from this article. Thanks for sharing this simple and informative content.