📈 Regression Analysis

Identifying factors that most strongly influence customer satisfaction

Multiple Linear Regression | OLS Method
R² (Explained Variance)
0.89
89% variance explained
Adjusted R²
0.85
Strong model fit
F-Statistic
24.6
p < 0.001
RMSE
0.18
Low prediction error

📐 Multiple Regression Equation

Statistically Significant
Predictive Model for Customer Satisfaction (CSAT)
CSAT = 3.12 - 0.18(UI/UX) - 0.02(Hosting) + 0.015(OnTime%) - 0.004(PMS) - 0.02(Lag%)
3.12
Intercept (Baseline CSAT)
5
Predictor Variables
310
Observations (n)

📊 Regression Coefficients

Impact of each variable on Customer Satisfaction

Coefficient Statistics

Variable β Std Err t-value p-value Sig
Intercept 3.12 0.45 6.93 <0.001 ***
UI/UX Revisions -0.18 0.05 -3.60 0.008 **
Hosting Incidents -0.02 0.01 -2.00 0.052 *
On-Time Delivery +0.015 0.003 5.00 <0.001 ***
PMS Issues -0.004 0.002 -2.00 0.048 *
Average Lag % -0.02 0.006 -3.33 0.012 **

*** p<0.001, ** p<0.01, * p<0.05

Standardized Coefficients (Beta Weights)

On-Time Delivery
+0.42
UI/UX Revisions
-0.38
Average Lag %
-0.32
Hosting Incidents
-0.22
PMS Issues
-0.15
💡 Impact Ranking
On-Time Delivery has the strongest standardized effect (+0.42), followed by UI/UX Revisions (-0.38). PMS Issues, while currently smaller (-0.15), is rapidly growing.

🔍 Coefficient Interpretation

What does each coefficient mean in practical terms?

✅ On-Time Delivery (+0.015)

50%
On-Time
→
60%
On-Time

+0.15 CSAT

Every 10% increase in on-time delivery improves CSAT by 0.15 points

❌ UI/UX Revisions (-0.18)

2 Rev
Current
→
5 Rev
Crisis

-0.54 CSAT

Each additional revision round reduces CSAT by 0.18 points

🔬 Model Diagnostics

Validating regression assumptions

Normality Test

Passed
0.12
Shapiro-Wilk p-value

p > 0.05 indicates residuals are approximately normally distributed

Multicollinearity

Passed
< 5
All VIF Values

VIF < 5 indicates no problematic multicollinearity

Autocorrelation

Passed
1.92
Durbin-Watson Statistic

Value near 2 indicates no autocorrelation

🎯 Prediction Scenarios

What-if analysis using the regression model

SCENARIO 1
Current State (2025)
3.4
UI/UX: 2.2 | Lag: 14% | PMS: 83
SCENARIO 2
If PMS Implemented
3.7
Lag: 8% | PMS: 10 | +0.3 CSAT
SCENARIO 3
If UI/UX = 2.0
3.5
Marginal improvement +0.1
SCENARIO 4
Optimal State
4.1
On-Time: 70% | Lag: 5%

💰 Business Impact (ROI of Improvements)

Improvement Investment CSAT Impact Revenue Impact ROI
Implement PMS/CRM ₹3,00,000 +0.3 +₹8,00,000 167%
Improve On-Time to 70% ₹1,00,000 +0.24 +₹5,50,000 450%
Reduce UI/UX to 1.5 avg ₹50,000 +0.13 +₹2,50,000 400%
Combined Optimization ₹4,50,000 +0.7 +₹15,00,000 233%
🎯 Primary Finding
On-Time Delivery is the single most important factor (β = +0.42). Improving from 54% to 70% could increase CSAT by 0.24 points - equivalent to recovering from 2025 levels to 2024 peak.
⚠️ Warning: PMS Growing Impact
While PMS coefficient is currently low (-0.15), the variable increased 7x (12→83) in one year. If this trend continues, PMS will become the dominant negative factor by 2026.

📈 Actual vs Predicted CSAT

4.0 3.5 3.0 2.5 2.0 Perfect Fit 2023 2024 2025 Predicted (X) vs Actual (Y) CSAT Scores

📊 Model Residuals (Prediction Errors)

0 +0.5 -0.5 2023 2024 2025
0.18
RMSE
0.15
MAE
±0.3
95% CI