Metric Definition Glossary Maintenance
Most BI teams launch a metrics glossary with real energy — a wiki page or a dedicated tool listing every core business metric with its definition and calculation — and the glossary is accurate for about as long as it takes for the first dashboard to be rebuilt with a slightly different calculation that nobody updates the glossary entry to match. Within a year, half the glossary describes how a metric used to be calculated, new metrics that have become important never got added at all, and new analysts trust the glossary anyway because it's the official source, which means they build their next dashboard against a definition that's already wrong.
STARTING PRICE
From €299
Standard tier · Multi-step workflow with AI extraction/decisioning and 2-3 integrations.
Get a quote →Saves roughly 3-5 hrs/month of manual glossary upkeep plus a documentation source new analysts can actually trust.
How the automation works
We compare your published metrics glossary against the actual calculation logic currently live in your BI tools and semantic layer, flagging entries where the documented definition has drifted from what dashboards are actually calculating, and separately flagging metrics that are in active, widespread use across dashboards but have no glossary entry at all. Each flagged drift shows the documented definition next to the current live calculation so the glossary owner can update the text in minutes rather than reverse-engineering what changed, and new-metric candidates are ranked by how many dashboards actually use them, so the glossary gets built out in order of what matters most rather than an arbitrary list.
Process flow
- 01
Pull the current published glossary integration
Existing glossary entries and their documented definitions are pulled from wherever they're maintained — Confluence, Notion, or a dedicated metrics catalog tool.
- 02
Extract live calculation logic integration
Actual current calculation logic for each metric is pulled from BI tool definitions and the semantic layer.
- 03
Compare documented vs. live definitions ai
Glossary entries are compared against live calculation logic to flag drift where the documented text no longer matches what's actually running.
- 04
Find widely-used metrics missing from the glossary ai
Metrics in active use across many dashboards but absent from the glossary are identified and ranked by usage breadth as candidates for new entries.
- 05
Route to the glossary owner for update output
Drift flags and new-entry candidates route to the glossary owner with the current live definition attached, ready for a quick documentation update.
Inputs
- Published metrics glossary content
- Live BI tool and semantic layer calculation definitions
- Dashboard-to-metric usage frequency data
- Glossary ownership assignment
Outputs
- Documentation drift flags with side-by-side comparison
- Undocumented high-usage metric candidates ranked by breadth
- Glossary freshness score over time
- Owner-routed update queue
Works with
Prefer a fully custom build instead of an off-the-shelf integration? We scope both options during your free consultation — most jobs like this one work fine on standard connectors, but higher-volume or non-standard systems sometimes need bespoke API work, reflected in the complex tier.
Where this goes wrong if you get it wrong
- A glossary entry documenting a deliberately simplified version of a metric for a general audience, distinct from the more precise technical calculation used in the actual dashboard, will look like drift to a naive text-versus-logic comparison even though the simplification is intentional — flagged drift needs a quick human check for intentional simplification versus genuine inconsistency before treating it as an error.
- Ranking undocumented metrics purely by dashboard count can miss a metric that's used in only one dashboard but is extremely high-stakes, like a compliance-driven calculation reviewed by auditors — usage breadth is a useful signal but shouldn't be the only factor deciding what gets prioritized for documentation.
- A glossary that's updated reactively, only when drift is detected, never captures new business context for why a definition changed — the update flow needs a field for the reasoning behind a definition change, not just the corrected calculation text, or the glossary becomes accurate again but no more useful for someone trying to understand why.
- Metrics calculated inconsistently across different teams (see KPI definition drift) will generate confusing, contradictory drift flags against a single glossary entry that was never meant to describe multiple variants — glossary maintenance and cross-team definition drift detection need to be coordinated, not run as two unrelated processes pointing at the same underlying inconsistency.
Frequently asked questions
Does this rewrite glossary entries automatically?
No, it flags drift and shows the current live definition alongside the documented one, but the glossary owner writes the actual update, since some drift reflects an intentional simplification rather than an error to correct.
How does it decide which undocumented metrics to prioritize?
Candidates are ranked primarily by how many dashboards actively use the metric, though high-stakes metrics used in fewer places, like compliance calculations, are worth reviewing even if they rank lower on pure usage volume.
Does this fix the underlying problem of teams calculating the same metric differently?
No, that's a separate, related problem — this keeps the glossary in sync with whatever the live calculation actually is, while cross-team definition drift detection identifies when the live calculations themselves disagree with each other.
How often does the glossary get checked against live definitions?
Typically on a monthly cadence, though it can run more frequently if your BI tools and semantic layer change often enough to warrant it.