Spike Detection

Highlight demand points that are unusually high relative to each item's own pattern.

← Back to Reports

Executive framing

Executive Summary — Answer First

Decision rule: review this page first, then move to deeper drill-down only when the result is economically meaningful.

Use this page to focus on forecast rows most likely to break the plan.
Spikes drive capacity risk, supplier stress, and potential service failures.
Not every spike is wrong, but every major spike needs an explanation.

SCQA / Context

Situation: some items experience sharp increases relative to their normal pattern. Complication: spikes may be valid, promotional, seasonal, or erroneous. Question: which forecast rows are abnormal enough to require active management? Answer: flag values exceeding the item's expected range and investigate them.

Issue Tree / Diagnosis

Magnitude and timing of unusual spikes
Operational feasibility of the revised demand
Root cause: promotion, seasonality, forecast error, or data issue

Analysis & Insights

Spike detection is an exception-management report, not a judgment that the value is wrong.
Near-term spikes are most critical because they compress response time.
Large spikes on constrained or high-volume items deserve immediate escalation.

Recommendations

Review near-term spikes in the weekly planning meeting.
Validate commercial drivers and supply readiness before accepting the plan.
Tune the alert threshold as the team gains confidence.

KPIs

Spike count by batch
Spike volume as a share of total
Near-term spike count
Resolved spike explanations

Implementation Roadmap

Phase	Key action	Owner	Success metric
1. Review	Run the report and isolate the highest-value exceptions.	Planning / Operations	Material issues identified within one review cycle.
2. Validate	Confirm whether the result reflects real business change, timing shift, or data-quality issue.	Forecast owner	Exceptions explained and documented.
3. Act	Translate validated findings into supply, capacity, purchasing, or governance actions.	Cross-functional team	Action owners assigned with due dates.

Risks, Gaps, and Next Steps

Do not treat a statistical exception as a business conclusion without validating the commercial or operational driver.
Where economics are missing, pair this page with item margin, lead time, or inventory exposure before escalating.
Use the summary line above the grid to confirm row volume; unexpected row counts often indicate filter or data issues.

Unable to load the report: @StdDevMultiplier is not a parameter for procedure CP_forecast_spike_detection.

Std Dev Multiplier

Analysis output

Data Table

Use the table below to validate the hypothesis, size the issue, and identify the items or periods that need action.

No rows returned for the selected filters.