Three steps. Design a tree or two, connect your data, and go live — within a week — on the warehouse and data stack you already have.
Templates for every business model — and every domain, from finance to growth, product, and ops.
Start with the KPIs that matter — revenue, retention, margin, engagement — and map the key input metric drivers and segments.
AI dramatically accelerates setup while keeping metric calculations governed and deterministic.
Trace's agent writes the spec — SQL, metrics, entities, attributes — from your schema or existing definitions.
The tedious work of analysis runs automatically — so teams can focus on decisions, not dashboard hunts.
Metric trees provide the model. Encoded analytical skills provide the expertise.
Short answers to frequently asked questions.
A metric tree is a computable model of how business metrics relate. It connects KPIs to their underlying drivers, giving both people and AI a structured model of how the business works.
No. Most customers start with 1–2 trees around the KPIs that matter right now, then expand as new workflows and questions emerge.
Dashboards monitor metrics. Trace connects them into a model of the business and applies analytical methods — automatically surfacing the key drivers and what matters most.
In a sense, Trace is an enhanced semantic layer plus an analytical layer.
Metrics and semantic layers define how numbers are calculated consistently. Trace goes further: metric trees encode the metric relationships, while analytical skills encode the workflows used to investigate them — turning governed metrics into automated analysis.
The industry is working hard to make AI-on-data reliable, but most systems still center on metric retrieval or SQL generation. Trace gives AI richer, more reliable business context — plus analytical methods — so it can run multi-step workflows across metrics and drivers.
For sensitive metrics like revenue, retrieval has to be exact. And deep, useful analysis requires more than a prompt — it relies on repeatable methods and workflows.
Trace keeps metric logic and analytical calculations deterministic, while AI orchestrates the analysis across the business — compressing days of work into minutes.
They package the expertise behind an analysis: analytical methods, workflow logic, business-specific nuance, interpretation, and narrative.
Methods include decomposition, attribution, anomaly detection, segment scans, mix analysis, variance-to-plan, and more. AI agents combine them into complete analytical workflows.
Most teams are live within two weeks. In week one, you design 1–2 metric trees and connect them to your existing data. In week two, the trees are live and automated analysis begins.
No. Trace reads directly from your existing data platform and works with the data you already have. When needed, our enhanced semantic layer can transform and shape your data into metric-tree models — so you don’t need to provide pre-modeled or structured datasets.
Trace can consume your existing metric definitions and models, enhance them into metric-tree models, and add the analytical logic needed to power automated analysis.
You do. Configs are clean, read-only, and version-controlled in git. Trace doesn’t become another source of truth — it’s a source of intelligence on top of the datasets you already govern.
Add a segment, edit the metric, update the tree — and the analyses automatically follow the new structure, so the methods and workflows stay aligned with how the business now works.
Trees evolve with your business, and every change remains versioned and auditable.