Scoop Login
User Guide
Start typing to search…

The Power of Why - Autonomous Investigations

The Power of Why: How Scoop Investigates Like a Senior Analyst

Stop Asking "What" - Start Asking "Why"

Traditional analytics tools answer "what" questions. Scoop is built to investigate "why" - just like a senior data analyst would, but in seconds instead of days.

How "Why" Questions Trigger Deep Investigations

Example 1: "Why did sales drop in March?"

What happens behind the scenes:

Your Question: "Why did sales drop in March?"

Scoop's Investigation Plan:
├── Probe 1: Measure sales change magnitude
│   └── Finding: -23% vs February, -18% vs March last year
├── Probe 2: Analyze by dimension breakdowns
│   └── Finding: Enterprise segment down 45%, SMB up 5%
├── Probe 3: Investigate enterprise segment specifically
│   └── Finding: 3 major accounts churned
├── Probe 4: Analyze churn patterns
│   └── Finding: All 3 had support tickets > 7 days
├── Probe 5: Check historical support correlation
│   └── Finding: 72% correlation between 7+ day tickets and churn
└── Synthesis: Poor support response driving enterprise churn

Root Cause Identified + Action Plan Provided

What you get: Not just "sales are down" but WHY and WHAT TO DO.

The Investigation Process

1. Strategy Formation

When you ask "why", Scoop:

  • Identifies the metric or change to investigate
  • Creates a hypothesis-driven exploration plan
  • Determines which probes will reveal root causes

2. Multi-Probe Execution

Each probe can:

  • Analyze different dimensions
  • Compare time periods
  • Find statistical anomalies
  • Identify correlations
  • Detect pattern changes

3. Intelligent Chaining

Unlike simple queries, each finding influences next steps:

  • If regional difference found → Investigate that region
  • If segment pattern found → Deep dive on that segment
  • If correlation found → Test causation hypothesis

4. Synthesis & Recommendations

Scoop doesn't just list findings. It:

  • Identifies primary root causes
  • Ranks contributing factors
  • Suggests specific actions
  • Explains in business terms

Types of "Why" Investigations

Performance Changes

  • "Why did conversion rate drop?"
  • "Why are costs increasing?"
  • "Why did productivity improve?"

Scoop investigates: Period comparisons, dimension analysis, correlation discovery, anomaly detection

Business Outcomes

  • "Why do some customers churn?"
  • "Why do certain deals close faster?"
  • "Why are some regions outperforming?"

Scoop investigates: Behavioral patterns, success factors, environmental variables, predictive indicators

Operational Issues

  • "Why are deliveries delayed?"
  • "Why did quality scores decrease?"
  • "Why is inventory building up?"

Scoop investigates: Process bottlenecks, cascade effects, systemic issues, root dependencies

Real Investigation Examples

"Why did customer satisfaction drop?"

Scoop discovering patterns in customer data over time

Investigation reveals:

  1. Response time increased 40% starting March 1st
  2. Correlated with new support system deployment
  3. Specific issue: Ticket routing to wrong teams
  4. Impact: 15% of tickets delayed > 24 hours
  5. Customer segment most affected: Enterprise
  6. Predicted churn risk: $2.3M in next quarter
  7. Fix: Reconfigure routing rules (specific instructions provided)

"Why are we losing deals to competitors?"

Decision tree analysis revealing what drives win/loss outcomes

Investigation reveals:

  1. Loss rate increased in deals > $50K
  2. Common factor: 3+ stakeholders involved
  3. Competitor advantage: Better multi-stakeholder tools
  4. Our sales cycle: 45 days vs their 32 days
  5. Key moment: Week 3 when additional stakeholders join
  6. Strategy: New collateral for technical buyers, accelerate POCs

"Why did profit margins shrink?"

Investigation reveals:

  1. Revenue stable but costs up 12%
  2. Primary driver: Shipping costs (+34%)
  3. Root cause: Order size decreased 23%
  4. Reason: New free shipping threshold too low
  5. Customer behavior: Splitting orders to hit minimum
  6. Solution: Adjust threshold or incentivize larger orders

Advanced Investigation Patterns

The Cascade Investigation

"Why did revenue drop?" leads to:

  • Why did sales volume decrease?
    • Why did conversion drop?
      • Why did trial-to-paid decrease?
        • Why did feature adoption fall?
          • Root cause: Onboarding bug in v2.3

The Comparison Investigation

"Why is Region A outperforming Region B?" reveals:

  • Different customer mix (found through ML segmentation)
  • Different competitive landscape
  • Different sales team structure
  • Key insight: Region A's model is replicable

The Prediction Investigation

"Why might we miss our target?" uncovers:

  • Current trajectory analysis
  • Leading indicator changes
  • Risk factor assessment
  • Scenario modeling
  • Preventive actions needed now

When to Use "Why" Questions

✅ Perfect for:

  • Understanding changes in metrics
  • Diagnosing problems
  • Finding improvement opportunities
  • Discovering hidden relationships
  • Getting actionable insights

💡 Try These Powerful Starters:

  • "Why did [metric] [change]?"
  • "Why are we seeing [pattern]?"
  • "Why do some [entities] [behavior]?"
  • "Why is there a difference between [A] and [B]?"
  • "Why isn't [expected thing] happening?"

The Magic: You Don't Need to Be Specific

Traditional BI Challenge:

You need to know exactly what to ask and where to look

Scoop's Approach:

Vague questions trigger comprehensive investigations

Example:

  • Vague: "Why are things weird this month?"
  • Scoop: Investigates all major metrics, finds anomalies, identifies that customer acquisition cost spiked due to a marketing channel issue

What Makes This Possible?

1. Reasoning Engine

  • 859-line prompt system for analytical thinking
  • Multi-step reasoning with validation
  • Context awareness across queries

2. ML Integration

  • Automatic pattern detection
  • Clustering for segmentation
  • Correlation analysis
  • Predictive modeling

3. Business Intelligence

  • Understands metric relationships
  • Knows investigation patterns
  • Applies domain expertise
  • Maintains context

Start Investigating Now

Export your investigations to PowerPoint with one click

Your First "Why" Question:

  1. Think of a metric that changed
  2. Ask Scoop: "Why did [metric] [change]?"
  3. Watch the investigation unfold
  4. Get root causes, not just data

Power User Tips:

  • Don't pre-filter your question
  • Let Scoop determine what's relevant
  • Ask follow-up "whys" on findings
  • Trust the investigation process

Beyond "Why": Other Investigation Triggers

  • "What's driving..." → Factor analysis
  • "What influences..." → Correlation investigation
  • "What explains..." → Statistical modeling
  • "What's causing..." → Causation analysis

The Bottom Line

Every business has "why" questions that go unanswered because investigating them manually takes too long. Scoop is your always-available senior analyst who can investigate any "why" in seconds, not days.

Stop wondering. Start discovering.

[Try Your First Why Question] | [See Investigation Demo]

Updated 21 days ago