January Weekly
Strongest month
April Weekly
Current month
Monthly Decline Rate
Average monthly drop
R8 Gap in April
Below rolling average
The Descent: Network Production vs R8 Predictions
Each line tracks a day-of-week across 4 months. Dashed lines show the R8 rolling average target.
January benefits renewal inflated the R8 average. Each subsequent month is weaker.
Monthly DOW Summary
| Month | Mon | Tue | Wed | Thu | Fri | Weekly Total | vs R8 |
|---|
Where Are We Heading?
Solid lines = actual Jan-Apr. Dashed = linear regression projection into May-Jun. Shaded band = R8 prediction range.
R8 Rolling Average
Based on 8-week history (includes strong January)
Linear Trend (May)
Based on Jan to Apr trajectory
Weighted MA (May)
Recent months weighted heavier
Key Insight
DOW Prediction Detail
| DOW | Jan | Feb | Mar | Apr | May (pred) | Jun (pred) | Monthly Slope | R8 |
|---|
Declining
Stable
Improving
Net Daily Impact
Decline minus improvement offset
Jan Avg vs Most Recent Month Avg
Dots below the diagonal are declining. Color = trend.
All 83 Practices
| Location ▲ | OD ▲ | Trend ▲ | Jan Avg ▲ | Feb Avg ▲ | Mar Avg ▲ | Apr Avg ▲ | Change % ▲ | R8 Avg ▲ |
|---|
Daily $ Lost by Ops Director
Sum of daily production gap for declining practices under each OD
Ops Director Detail
| Ops Director ▲ | Total Practices ▲ | Declining ▲ | % Declining ▲ | Daily $ Lost ▲ |
|---|
Network Production by Day of Week (4-Month Erosion)
Grouped bars showing each month side-by-side per DOW
Tuesday Crisis
Thursday Crisis
DOW Decline Pattern (Declining Practices Only)
| DOW | Jan | Feb | Mar | Apr | Change % |
|---|
R8 Rolling Average Formula
The R8 prediction is a simple average of the most recent 8 weeks of production data for each day of week. For example, the R8 for Monday = average of the last 8 Mondays of network production. Because the 8-week window includes January (the strongest month due to benefits renewal), the R8 is systematically high relative to current production levels.
$300 Threshold
A daily production value of $300 or less is treated as a non-operating day (practice closed, doctor out, etc.). This threshold matches SGA's internal "one-doctor-day" logic. Days below this threshold are excluded from averages to avoid deflating the metrics.
Data Sources
- 83 VELOX practices: Daily production data pulled directly from VELOX reporting system with day-of-week granularity
- 7 estimated practices: These non-VELOX practices have monthly totals only. Daily values are estimated by dividing monthly totals by observed operating days. This is standard methodology for practices on different reporting systems.
- 90 total practices in the analysis across all months
Prediction Models
- Linear Regression: Fits a line through Jan-Apr data points for each DOW. Slope indicates monthly rate of change. Extended to May and June. Simple and interpretable but assumes constant rate of decline.
- Weighted Moving Average (WMA): Weights recent months more heavily (Apr=4, Mar=3, Feb=2, Jan=1). More responsive to recent trends than linear regression. The
avg_monthly_changeis projected forward for May/Jun estimates. - R8 Rolling Average: 8-week trailing average. Slowest to respond to changes because it includes the strong January data. Useful as a baseline but overstates expected production.
Limitations
- No scheduled production (booked) data is incorporated into forward predictions. Predictions are purely based on historical actuals.
- April data may be partial depending on when the analysis was run.
- Linear regression assumes a constant rate of decline which may not hold if interventions are made.
- Benefits renewal seasonality (strong January) is a known factor but not explicitly modeled.