Data Academy · Layer 9 · Contextualize

Metrics & Knowledge Fabric

Parse BI reports, dashboards and semantic models into a governed metrics layer.

Lesson 11 of 17 · Find it on the platform map

01 What it is

What this layer does

The metrics and knowledge fabric turns the business logic buried in BI reports, dashboards and semantic models into a governed, searchable layer of metrics and knowledge. SCIKIQ discovers metadata in tools such as Power BI, Tableau, Looker, Qlik, Excel and dbt, extracts each metric’s formula and dimensions, attaches business meaning, and links it to the domain, process and decisions it supports. It also holds the documents, policies and rules that explain how the business operates.

02 Concepts

Four ideas to hold on to

Metric definition

A metric is more than a number: it has a formula, the dimensions it can be sliced by, a target, thresholds and a direction that says whether up is good. SCIKIQ records KPI definitions with target, thresholds and direction, alongside dimensions and metric history.

Metric reconciliation

Comparing how the same metric is calculated in different tools and systems, and resolving or documenting the differences. SCIKIQ reconciles metric implementations across systems so that variants are visible rather than hidden.

BI lineage and usage

Lineage traces a number on a visual back through its calculations and queries to the source data; usage shows which reports are actually opened. SCIKIQ traces BI lineage through Power Query, DAX and visuals to metrics, and flags reports as used, unused or file-fed.

Driver decomposition

Breaking a change in a metric into the contributions of its underlying drivers, so you can answer why it moved rather than only that it moved. SCIKIQ provides driver decomposition to answer the question: why did it change?

03 In SCIKIQ

What the layer contains

BI discovery and metadata (Power BI, Tableau, Looker, Qlik, Excel, dbt)Metric extraction: formula and dimensionsBusiness meaning and definitionsMapping to domain and processKPI to decision to actionDocuments, events, policies and rulesOperational knowledge: hybrid keyword and vector searchKPI definitions: target, thresholds, directionDimensions and metric historyWhy did it change? Driver decompositionMetric implementations reconciled across systemsBI lineage to metrics (Power Query, DAX, visuals)BI report usage: used, unused and file-fed

04 What good looks like

Signs it is working

Each key metric has one agreed definition with its formula, dimensions, target, thresholds and direction recorded.
Where two systems calculate the same metric differently, the difference is documented and has an owner.
Any number on a dashboard can be traced back to its source fields and calculations.
Reports that nobody uses, or that depend on manual files, are identified and reviewed.

05 Diagnostics

Questions to ask your team

1
How many different definitions of our top five KPIs exist across our BI tools today?
2
Which of our dashboards are actually used, and which depend on someone’s spreadsheet?
3
When a KPI moves sharply, how long does it take us to explain which drivers caused it?
4
Is each KPI linked to the decision and action it is meant to inform, or is it reported for its own sake?

06 Keep going

Frequently asked

Do we have to rebuild our dashboards to use a metrics layer?

No. The fabric reads existing BI reports and semantic models to extract metrics and their lineage; the dashboards can stay where they are while definitions are governed centrally.

What is the knowledge part of the fabric?

Alongside metrics it holds documents, events, policies and rules, searchable through hybrid keyword and vector search. This gives people and agents the operational context behind the numbers.

Why link a KPI to a decision and an action?

A metric earns its place when someone acts on it. Mapping KPI to decision to action makes clear who responds when a threshold is crossed and what they are expected to do.