H34DATA

Automated insight surfacing

Continuously analyzes the firm's metrics for changes worth noticing — unusual movements, emerging trends, anomalies, segments performing differently than expected — and pushes a digest of what matters to the people who should care. Different from a dashboard because the pattern decides what's worth surfacing rather than letting humans scan everything; different from generic anomaly detection because the pattern speaks in business language ('your premium segment churn is rising while standard is stable') rather than statistical jargon. The pattern's value is replacing the 'I look at the dashboard every Monday and hope something jumps out' habit with proactive flagging of what's actually moving.

WHERE THIS FITS

BUSINESS SHAPES

B2B servicesProduct companyDirect-to-consumerMarketplace / two-sided

VOLUME THRESHOLD

Below 30 tracked metrics a month, the payback rarely earns the build. Patterns this shape reliably pay back at 200+.

FITS BEST

Operations-heavy SMBs, e-commerce, marketing teams running multiple channels.

PAYBACK · 6-12 moBUILD · MediumVALUE · $30k-$120kWHEN · multiple business metrics tracked

FAILURE MODE TO DESIGN AROUND

Noise treated as signal → tune sensitivity per metric, not flat.

REQUIREMENTS · 4 REQUIRED, 2 OPTIONAL

Requirements describe capabilities the pattern needs in your environment, not the vendors you must buy. Any system that fills a requirement satisfies it — that’s what makes the catalog portable across the long tail of SMB tooling.

metric_time_series_source
REQUIREDREADbatch
The metrics being monitored over time, at appropriate granularity.
DATA SHAPE
Per-metric per-time-period values with relevant dimensions (segment, region, product line).
COMMONLY FILLED BY
- data warehouse with metric tables
- BI platform with metric layer
- structured time-series store
metric_definitions_and_owners
REQUIREDREADcorpus
What each metric means, who cares about it, what level of change is meaningful, what's normal seasonality.
DATA SHAPE
Per-metric: definition, owner (who gets the alert), expected range or trend, known seasonality, threshold sensitivity.
COMMONLY FILLED BY
- semantic layer documentation
- structured metric library maintained by the data team
- metric ownership matrix
context_data_for_explanation
RECOMMENDEDREADcorpus
Supplementary data the pattern uses to explain why a metric moved: campaign launches, releases, external events, related metric movements.
DATA SHAPE
Time-indexed contextual events with descriptions and affected areas.
IF MISSING
Pattern still surfaces movements but can't explain them. Alerts become 'X moved' rather than 'X moved because Y happened'. Recommend even basic context capture.
COMMONLY FILLED BY
- release notes log
- campaign calendar
- operational event log
- decisions captured by C9 if live
insight_delivery_destination
REQUIREDWRITEevent
Where insights reach the owner, in their normal working surface.
DATA SHAPE
Per-insight: what moved, by how much, when, possible causes, recommended attention, deep-link to investigate.
COMMONLY FILLED BY
- chat digest to the metric owner
- email summary at expected cadence
- card in the BI tool with the insight surfaced
- morning briefing to executives
alert_calibration_loop
REQUIREDREAD + WRITEevent
How owners flag whether insights were genuinely useful or noisy, used to tune sensitivity.
DATA SHAPE
Per-insight feedback: useful, already-knew, noise, wrong-cause-attribution.
COMMONLY FILLED BY
- thumbs widget on each insight
- weekly retro with the data team
- structured feedback capture per insight
narrative_archive
RECOMMENDEDWRITEcorpus
Long-term archive of insights and what happened next, valuable for understanding business history.
DATA SHAPE
Per-insight: what was flagged, what action (if any) was taken, what happened after.
IF MISSING
Pattern surfaces in the moment but the institutional memory is weaker. Recommend for any data-driven business.
COMMONLY FILLED BY
- structured insight log in the data platform
- annotated time-series in the BI tool
- narrative archive accessible to leadership

RUNTIME FLOW · 8 STEPS

01
On regular cadence (often daily, sometimes hourly for fast-moving metrics), pull recent metric values
metric_time_series_source
02
Apply seasonality and trend models to identify genuine movements vs. expected variation
metric_definitions_and_owners
03
Detect anomalies, trend shifts, and segment-level divergences
04
For each meaningful movement, search context data for possible causes
context_data_for_explanation
DECISION Skip if context_data_for_explanation not filled; just describe the movement.
05
Compose business-language insight: what moved, by how much, in what direction, possible causes
06
Route to the metric owner through their delivery destination
insight_delivery_destination
07
Archive the insight in the narrative archive
narrative_archive
08
Capture feedback for tuning sensitivity and improving causal attribution
alert_calibration_loop

EMISSIONS · 3

Structured outputs this pattern produces. Other patterns and client systems can subscribe to them, which is how the catalog composes over time.

insight_quality_signal
Useful-vs-noise ratio per metric and metric type, used to tune.
CONSUMED BY
- data team workflows
- alerting calibration
- metric ownership reviews
metric_health_history
Long-term annotated time series with insights and actions taken, useful for retrospectives and pattern recognition.
CONSUMED BY
- business retrospectives
- annual strategy reviews
- executive learning
explanation_quality_signal
Whether the pattern's cause attributions were correct, used to improve the context-to-cause mapping.
CONSUMED BY
- model tuning
- data team retrospectives