7 KPIs Every Geotechnical Dashboard Must Track

On complex works, the difference between “monitoring” and “management” is whether the geotechnical view drives clear hold/accelerate/mitigate decisions at the right time. A decision-grade dashboard does four things well: it consolidates historical and live data (boreholes, CPT, lab, piezos, inclinos, settlements, maintains a continuously calibrated ground model, projects performance against serviceability limits, and records auditable triggers with the action taken. That is the essence of a modern, defensible Observational Method—data in, model checked, forecast made, decision logged.

The seven KPIs

1) Pore-pressure exceedance & trend. Pore pressure is the earliest indicator that ground behaviour is diverging from assumptions. A useful signal is not the latest value—it’s whether exceedances are becoming frequent and whether a rising trend persists once rainfall, pumping and staging are accounted for. When this rises, teams can adjust drainage, alter surcharges or slow excavation before displacement shows up elsewhere. This is the leading indicator that protects both stability and consolidation-driven programs.

2) Settlement forecast with time-to-limit. Owners and site teams need a date-certain view, not just a curve. A credible dashboard shows settlement with uncertainty and a live countdown to the serviceability threshold for each critical point. After significant events (storms, loads, sequence changes) and after model recalibration, the countdown should update. This turns settlement from documentation into planning: preload brought forward, drainage capacity boosted, heavy moves resequenced—avoiding last-minute stoppages.

3) Lateral displacement utilization & velocity. Serviceability often governs walls and deep excavations long before ultimate capacity. Two numbers matter weekly: how much of the allowable movement has been consumed, and how fast that utilization is changing. Mapped by panel/section, this isolates where support stiffness, strutting or excavation pace must be tuned. Velocity is the truth-teller: a stable utilization can still be risky if the rate is quietly accelerating.

4) Model-to-measurement fit & parameter drift. The Observational Method only works if the model continues to explain reality. Routine review of residuals (by sensor family) and of how key parameters have shifted since baseline (stiffness, permeability, effective strength, OCR) shows whether changes are noise, boundary-condition artefacts, or true ground response. Sustained drift demands a documented recalibration and a design note; a stable fit supports leaner, confident decisions without hidden optimism.

5) Probability-weighted geo-risk index. Executives need one comparable, defensible number per site to allocate attention and budget. Combining likelihood and consequence across the key hazards (excess settlement, slope/retaining instability, uplift, etc.), and tracking the trajectory, prevents teams from arguing plot-by-plot. It surfaces the real risers so inspections, mitigations and contingency releases go where they matter most.

6) Data health & completeness. Bad data equals bad decisions. A dashboard must make data SLAs visible: latency (how fresh), completeness (how much arrived), calibration and device health (battery, housing, comms). When this slips, engineers should see it before they act on a spurious alarm. Fast maintenance tickets, redundancy fallbacks and clear flags on any analysis that used stale inputs protect credibility.

7) Observational Method Action Index. Senior stakeholders benefit from a single “act / watch / good” signal—provided the why is visible. Rolling up the indicators above into a weighted, explainable index creates a common language across regions and phases. Paired with short “because” chips (e.g., “Δu trend rising”, “residuals widening”, “utilization stable”), it drives accountability: an owner, a deadline, a scheduled reassessment.

Applying the framework in DAARWIN

DAARWIN’s workflow aligns with this structure: Digitization → Data Management → Predictions vs Data → Back-Analysis → Governance. Historical investigations are digitised and standardised (units, metadata, QA/QC). Live streams from piezometers, inclinometers, settlement points are ingested on a UTC time base with device-level SLAs. Predictions vs Data presents settlement forecasts with time-to-limit and pore-pressure trend cards with automatic rainfall/pumping annotations.

Back-Analysis maintains a calibrated model, logging parameter changes with version history and approver notes, and ensuring modelling choices remain aligned with staging and boundary conditions. Governance completes the loop: each KPI card displays data lineage, model version, trigger rule and the recorded decision (what, who, when, why), supporting internal audits and owner reporting without additional administrative work.

Deployment and governance

A pragmatic rollout begins with a single pilot zone surrounding one primary limit state and three to five instruments. Historical data are ingested and cleaned; live feeds are connected; residual bands and threshold logic are agreed with reference to serviceability limits, geological context, instrument accuracy, and dual lookbacks. The seven KPIs are then enabled and operated on a fixed cadence: review time-to-limit and the composite index; execute mitigations; log decisions; adjust thresholds once based on observed behaviour. Portfolio scaling should follow only after the full loop—data → model → forecast → action → audit—runs reliably. This approach provides earlier warnings (pore-pressure trends, utilisation velocity), fewer stop-work events (credible time-to-limit), evidence-backed model updates (fit and drift), defensible prioritisation (risk index), and transparent traceability across regions and asset classes.

Want this live? With the out-of-the-box widgets above, teams commonly stand up a decision-ready pilot in a single working cycle in DAARWIN—giving stakeholders a clear, shared language for risk, cost and schedule across USA, Europe, Gulf and Oceania. Book a live demo now!