The Mint's Public Ledger

Opening Scene

A mint publishes a clear, public record of exactly which denominations it currently issues — what coins exist, what they’re worth, what they look like. Without that public ledger, people would have no reliable way to know what currency actually exists and is valid to use, forcing them to rely on rumor, guesswork, or their own limited personal experience with whatever coins happened to pass through their hands.

A metrics catalog needs this exact same public, discoverable role, and its absence leaves people re-deriving metrics that already exist, simply because they didn’t know to look for them.

In Plain English

A metrics catalog is a searchable, browsable inventory of every metric definition available in a semantic layer, along with its documentation, owner, and current status. It solves the discoverability problem: someone needing “customer lifetime value” should be able to find the existing, authoritative definition rather than building their own from scratch, simply because they didn’t know it already existed.

The Old Way

Even organizations with genuinely well-governed semantic layers sometimes struggled with a more basic problem: people didn’t know what metrics already existed to look for in the first place. A well-defined, well-tested, well-documented metric definition provided little practical value if the person who needed it had no way to discover it existed before building their own redundant version.

This created a frustrating, avoidable pattern: someone would spend real effort building a metric that already existed elsewhere in the organization, discovering the duplication only by accident, much later, when comparing notes with a colleague who happened to already be using the original.

What’s Changing (and Why AI Is the Reason)

AI-assisted natural-language search can make metric discovery dramatically easier than manually browsing a catalog. Rather than requiring someone to know the exact name of an existing metric to find it, AI-assisted search can match a natural-language description of what someone needs — “how many customers bought again within 90 days” — to the existing, relevantly-named metric definition that already answers it.
AI-assisted duplicate detection can flag when a proposed new metric significantly overlaps with an existing one. Rather than relying entirely on the metric author to search the catalog thoroughly before building something new, AI-assisted analysis of a proposed metric definition can proactively flag likely overlap with existing definitions, catching redundant effort before it happens rather than discovering it after the fact.
AI-assisted catalog summarization can make a large, complex catalog genuinely navigable. As a metrics catalog grows to include hundreds or thousands of definitions, AI-assisted tooling can help organize and summarize it meaningfully — grouping related metrics, surfacing the most commonly used ones, highlighting recently changed definitions — rather than leaving users to navigate an undifferentiated, overwhelming list.

The Metaphor, Fully Extended

Mint Element	Metrics Catalog Concept
A mint’s public ledger of currently issued denominations	A metrics catalog of currently available metric definitions
Someone unable to find out what currency actually exists	Someone unable to discover an existing metric definition that already answers their need
Rebuilding a coin design that already exists elsewhere in the kingdom	Rebuilding a metric definition that already exists elsewhere in the organization
A clerk helping a visitor find the specific denomination they actually need	AI-assisted natural-language search matching a described need to an existing metric
A royal auditor noticing two mints have accidentally struck nearly identical currency	AI-assisted duplicate detection flagging significant overlap with an existing metric

For Beginners: What to Actually Do

Before building a new metric from scratch, practice checking your organization’s metrics catalog first, even if you’re not entirely sure what it might be called.
Get comfortable using AI-assisted natural-language search, where available, to describe what you need rather than requiring you to know an existing metric’s exact name.
If you discover you’ve built something that already existed, treat that as useful information about a catalog discoverability gap, not just a personal inefficiency to feel bad about.
Notice how much redundant effort a genuinely good, well-used metrics catalog prevents across an organization, even when that prevented effort is invisible by nature.

For Practitioners and Leaders: The Deeper Layer

Invest in a genuine, actively maintained metrics catalog as a companion to your semantic layer, recognizing that well-governed definitions provide little value if nobody can discover they exist.
Use AI-assisted natural-language search to lower the barrier to metric discovery, letting people describe their need rather than requiring them to already know a metric’s exact name.
Use AI-assisted duplicate detection during the metric proposal process (Article 13) to catch redundant effort proactively, before a near-duplicate metric is built and published.
Use AI-assisted catalog summarization to keep your catalog genuinely navigable as it grows, rather than letting it become an overwhelming, undifferentiated list that people give up searching.

Quick Recap

A metrics catalog is a searchable, browsable inventory of available metric definitions, solving the discoverability problem that leaves people rebuilding metrics that already exist.
Without genuine discoverability, well-governed metric definitions provided little practical value, since people had no reliable way to know they existed before building redundant versions.
AI-assisted natural-language search and duplicate detection make existing metrics easier to find and reduce redundant effort proactively, before it happens.
AI-assisted catalog summarization keeps a growing catalog genuinely navigable, rather than letting it become an overwhelming list nobody actually uses.

Where This Fits in the Series

Article 14 covered making a definition legible long after its author is gone. This article covered making metrics genuinely discoverable in the first place. Article 16 looks at what happens when machines start spending the currency directly.