Spend Velocity
Spending pattern analysis and anomaly detection
Overview
Spend Velocity (also called Cost Dynamics) applies physics concepts to spending analysis. Just as velocity measures speed of motion, Spend Velocity measures the speed and acceleration of your spending.
This dashboard catches subtle changes that other analyses miss — the slow subscription creep, the viral software adoption, the zombie vendor still getting paid.
The Physics of Spending
Velocity
How fast is spending growing?
Velocity = (Current Spend - Prior Spend) ÷ Prior Spend
Example: If you spent $10,000 last month and $12,000 this month:
Velocity = ($12,000 - $10,000) ÷ $10,000 = 20%
Acceleration
Is growth speeding up or slowing down?
Acceleration = Current Velocity - Prior Velocity
Example: If velocity was 10% last month and 20% this month:
Acceleration = 20% - 10% = 10% (accelerating)
Positive acceleration = spending growth is increasing Negative acceleration = spending growth is slowing
Health Score
The Score
Spend Velocity Health Score (0-100) measures the health of spending patterns:
| Score | Grade | Meaning |
|---|---|---|
| 80-100 | A | Healthy, controlled spending |
| 70-79 | B | Good with minor anomalies |
| 60-69 | C | Some concerning patterns |
| 50-59 | D | Multiple issues detected |
| Below 50 | F | Significant spend control problems |
Higher is better.
Key Metrics
Header Metrics
| Metric | Description |
|---|---|
| Health Score | Overall spend health |
| Total Spend | Spending in the period |
| Average Monthly Velocity | Typical monthly growth rate |
| Acceleration | Change in velocity |
| Accelerating Vendors | Vendors with increasing velocity |
| Alert Count | Number of spend anomalies |
Detection Models
Spend Velocity uses seven specialized detection models:
1. Velocity & Acceleration Tracking
What It Detects: Vendors or categories where spending is growing rapidly, especially when growth is accelerating.
Why It Matters: Catching cost increases early, before they become significant.
Indicators:
- High velocity (rapid growth)
- Positive acceleration (getting faster)
- Velocity above threshold (default 20%)
2. Boiling Frog Detector
What It Is: Named for the metaphor of a frog in slowly heating water. Detects incremental price increases that are individually too small to notice but compound over time.
What It Detects:
- Subscriptions with creeping price increases
- Service contracts with annual escalators
- Vendors slowly raising prices
Example: A vendor charges $1,000/month. They increase prices 3% per year. After 5 years, you're paying $1,159/month. Each increase seemed small, but the total is significant.
Threshold: Default 25% cumulative increase over 180 days.
3. Shadow IT Detection
What It Is: Identifies software and services that spread "virally" through the organization without IT/procurement oversight.
What It Detects:
- Tools adopted by multiple employees independently
- Departmental software subscriptions proliferating
- Unauthorized SaaS spreading
How It Works: Tracks when payments to a vendor appear from multiple cost centers or departments, especially new-to-the-organization vendors.
Threshold: Default 5 employees using a tool within short period.
4. Anomaly Heat Signatures
What It Is: Statistical outlier detection using standard deviations.
What It Detects:
- Spending significantly above or below normal
- Unusual transaction patterns
- One-time large purchases
- Missing expected regular payments
How It Works: For each vendor/category, calculates the mean and standard deviation. Flags amounts outside 2.5 standard deviations.
5. Commitment Cliff Analysis
What It Is: Compares purchase order velocity to actual spending velocity.
What It Detects:
- Overcommitment (POs far exceeding actual purchases)
- Undercommitment (spending exceeding planned)
- Timing mismatches between commitments and spend
Why It Matters: POs are future commitments. If PO velocity diverges from spend velocity, something is changing.
6. Seasonal Pattern Recognition
What It Is: Identifies cyclical spending patterns and flags when current behavior doesn't match expected seasonality.
What It Detects:
- Missing seasonal spend (you usually spend more in Q4, but not this year)
- Unusual timing (spending that normally happens in March happens in February)
- Seasonal vendors with unexpected activity
7. Zombie Spend Detection
What It Is: Identifies vendors you're still paying but haven't actively purchased from.
What It Detects:
- Subscriptions for departed employees
- Unused software licenses
- Services no longer needed
- Auto-renewing contracts forgotten
Threshold: Default 90 days of payments without new purchases.
Dashboard Sections
Velocity Overview
Summary gauges showing:
- Overall velocity trend
- Acceleration direction
- Number of anomalies
Vendor Velocity Table
| Column | Description |
|---|---|
| Vendor | Vendor name |
| Current Spend | This period's spending |
| Prior Spend | Last period's spending |
| Velocity | Growth rate |
| Acceleration | Change in velocity |
| Flags | Active anomaly alerts |
Detection Alerts
Grouped by detection type:
- Boiling Frog alerts
- Shadow IT discoveries
- Statistical anomalies
- Zombie vendors
Each alert shows:
- Vendor/category affected
- Magnitude of issue
- Recommended action
Trend Charts
- Velocity over time
- Top accelerating vendors
- Anomaly frequency trends
Using the Dashboard
Weekly
- Review any new anomaly alerts
- Check acceleration trends
- Investigate flagged vendors
Monthly
- Review Boiling Frog detections (slow creep)
- Check for Shadow IT spread
- Audit Zombie vendors
- Analyze seasonal patterns
Quarterly
- Comprehensive spend velocity review
- Vendor negotiation planning (based on velocity data)
- Policy adjustments for spending control
- Shadow IT cleanup initiatives
Configuration Options
Access via Settings → Spend Velocity:
Velocity Thresholds
| Setting | Purpose | Default |
|---|---|---|
| High Velocity | Flag when monthly growth exceeds | 20% |
| Acceleration Threshold | Flag when acceleration exceeds | 15% |
Boiling Frog Settings
| Setting | Purpose | Default |
|---|---|---|
| Detection Window | Days to look back | 180 |
| Cumulative Increase | Total increase to flag | 25% |
Shadow IT Settings
| Setting | Purpose | Default |
|---|---|---|
| Employee Threshold | Users before flagging | 5 |
| Growth Period | Time window for spread | 90 days |
Anomaly Detection
| Setting | Purpose | Default |
|---|---|---|
| Standard Deviations | Outlier threshold | 2.5 |
| Minimum Data Points | Required history | 6 periods |
Zombie Detection
| Setting | Purpose | Default |
|---|---|---|
| Inactivity Period | Days without new purchase | 90 |
| Minimum Amount | Threshold to flag | $500 |
Best Practices
Investigate Acceleration, Not Just Velocity
High velocity might be planned (new project, seasonal). Acceleration indicates things are changing — that's when to dig in.
Review Boiling Frog Quarterly
Slow creep is designed to avoid notice. Set a calendar reminder.
Shadow IT Isn't Always Bad
Sometimes employees find great tools. Review Shadow IT discoveries as potential adoption candidates, not just violations.
Kill Zombies Promptly
Every zombie vendor is pure waste. Audit quarterly and cancel aggressively.
Connect to Contracts
Pair velocity analysis with contract renewal dates. Negotiate before renewal when you have leverage.
Common Questions
Why is high velocity sometimes okay?
Growth requires spending. A vendor with 50% velocity growth may be supporting a 50% revenue growth. Context matters.
How do I distinguish signal from noise?
Focus on:
- Recurring patterns (not one-time spikes)
- Multiple alert types on same vendor
- Material amounts
What's a good velocity target?
Spending velocity should be at or below revenue velocity. If revenue grows 15%, spending growing 15% is neutral. Growing faster means shrinking margins.
Can I track by category instead of vendor?
Yes, the dashboard shows both vendor-level and category-level analysis.
How far back should I look?
Default lookback is 180 days for trend analysis. Longer periods provide more stable baselines but may miss recent changes.
Related Dashboards
- Procurement — Vendor performance and relationships
- Profitability — Margin impact of spending
- Sentinel — Unusual spending patterns and fraud