Are Incidents Really Declining? Trend Analysis with Rolling Averages

NexGenIT deployed a new monitoring system in July. The NOC manager claims incidents dropped 30% — from 20 in June to 14 in September. But the analyst must manually compute rolling averages, month-over-month changes, and year-over-year comparisons to determine whether the improvement is a real trend or just random fluctuation.

# Aplly.xyz Case Study Submission

## Title
Are Incidents Really Declining? Trend Analysis with Rolling Averages

## Type
Technology/IT

## Difficulty
Intermediate

## Estimated Time
75 minutes

## Overview
NexGenIT deployed a new monitoring system in July. The NOC manager claims incidents dropped 30% — from 20 in June to 14 in September. But the analyst must manually compute rolling averages, month-over-month changes, and year-over-year comparisons to determine whether the improvement is a real trend or just random fluctuation.

## Case Details

Function Focus: Multi-method trend analysis, rolling average computation, month-over-month vs year-over-year comparison, separating signal from noise

Scenario:
NexGenIT's new AI-based monitoring system was deployed on July 1. The NOC manager reports that total monthly incidents dropped from 20 in June to 14 in September — a 30% reduction. The IT director wants verification. The analyst has 12 months of current-year incident data by severity, plus prior-year totals. She must manually compute three trend indicators — month-over-month change, 3-month rolling average, and year-over-year comparison — to determine whether the decline is a real trend, a seasonal dip, or just month-to-month noise.

Dataset Structure:
- 12 months of 2026 incidents by severity (Critical, High, Medium)
- 12 months of 2025 total incidents for YoY comparison
- Deployment date: July 1 (month 7)

Tasks:
1. Compute month-over-month change (%) for total incidents Jan–Dec — identify the largest single-month drop and the largest single-month rise
2. Compute the 3-month rolling average of total incidents for each month (Mar onward) — compare the pre-July average (Jan–Jun) vs post-July average (Jul–Dec) and state whether the post-July level is a sustained decline or a temporary dip
3. Compute year-over-year change for each month (2026 vs 2025) — identify whether post-July months show improvement beyond the pre-existing YoY trend
4. Using all 3 indicators, answer the NOC manager's claim with specific numbers — is the 30% reduction claim supported, or is the evidence inconclusive?
5. Only after submitting your manual analysis, re-run all calculations in a spreadsheet or AI tool and report any discrepancies — recommend what additional data would strengthen the conclusion

Expected Output:
A trend analysis memo containing: (a) month-over-month change table, (b) rolling average comparison (pre vs post July), (c) YoY comparison table, (d) a conclusion on the 30% claim with specific evidence, and (e) a round-trip discrepancy note.

Evaluation Criteria:
Correct month-over-month percentage arithmetic, correct rolling average computation, correct YoY comparison, a well-supported conclusion that references specific numbers from all 3 indicators, and honest discrepancy reporting.

## Data Sources

2026 Monthly Incidents by Severity:
| Month | Critical | High | Medium | Total |
|---|---|---|---|---|
| Jan | 1 | 5 | 10 | 16 |
| Feb | 2 | 4 | 12 | 18 |
| Mar | 1 | 6 | 11 | 18 |
| Apr | 0 | 5 | 13 | 18 |
| May | 2 | 7 | 10 | 19 |
| Jun | 1 | 6 | 13 | 20 |
| Jul | 2 | 4 | 9 | 15 |
| Aug | 1 | 3 | 10 | 14 |
| Sep | 0 | 4 | 10 | 14 |
| Oct | 1 | 5 | 11 | 17 |
| Nov | 0 | 3 | 9 | 12 |
| Dec | 1 | 4 | 8 | 13 |

Prior Year (2025) Monthly Total Incidents:
| Month | 2025 Total |
|---|---|
| Jan | 14 |
| Feb | 16 |
| Mar | 17 |
| Apr | 15 |
| May | 17 |
| Jun | 19 |
| Jul | 18 |
| Aug | 17 |
| Sep | 16 |
| Oct | 18 |
| Nov | 15 |
| Dec | 14 |

Context:
- New monitoring system deployed: July 1, 2026 (cost: ₹35 lakh)
- NOC manager's claim: "30% reduction in total incidents since deployment"
- IT director's question: "Did the system actually work, or is this coincidence?"

## Solution Frameworks
Trend analysis, rolling average smoothing, month-over-month vs year-over-year comparison, separating signal from noise

## Solver Guidance & Tutorials
Link to: "Trend Analysis Without Dashboards — Manual Calculation Methods" tutorial

## What You'll Learn
- Computing and interpreting rolling averages by hand
- Distinguishing a real trend from seasonal or random fluctuation
- Using year-over-year comparisons to control for seasonality
- Building a defensible conclusion from multiple trend indicators

## Tags
incident trends, rolling average, year-over-year, IT operations, trend analysis

## Registration Links
- Register as Solver
- Register as Evaluator

2026 Monthly Incidents by Severity:
| Month | Critical | High | Medium | Total |
|---|---|---|---|---|
| Jan | 1 | 5 | 10 | 16 |
| Feb | 2 | 4 | 12 | 18 |
| Mar | 1 | 6 | 11 | 18 |
| Apr | 0 | 5 | 13 | 18 |
| May | 2 | 7 | 10 | 19 |
| Jun | 1 | 6 | 13 | 20 |
| Jul | 2 | 4 | 9 | 15 |
| Aug | 1 | 3 | 10 | 14 |
| Sep | 0 | 4 | 10 | 14 |
| Oct | 1 | 5 | 11 | 17 |
| Nov | 0 | 3 | 9 | 12 |
| Dec | 1 | 4 | 8 | 13 |

Prior Year (2025) Monthly Total Incidents:
| Month | 2025 Total |
|---|---|
| Jan | 14 |
| Feb | 16 |
| Mar | 17 |
| Apr | 15 |
| May | 17 |
| Jun | 19 |
| Jul | 18 |
| Aug | 17 |
| Sep | 16 |
| Oct | 18 |
| Nov | 15 |
| Dec | 14 |

Context:
- New monitoring system deployed: July 1, 2026 (cost: ₹35 lakh)
- NOC manager's claim: "30% reduction in total incidents since deployment"
- IT director's question: "Did the system actually work, or is this coincidence?"

Trend analysis, rolling average smoothing, month-over-month vs year-over-year comparison, separating signal from noise