Key Driver Analysis

Definition

Key driver analysis is a set of analytical techniques that identify which factors most strongly influence a business outcome. Methods include correlation analysis, regression-based importance (standardised coefficients, Shapley values), and machine-learning-based feature importance.

Why It Matters

Organisations collect vast amounts of data on customer experiences, but without key driver analysis they cannot determine which touchpoints actually move the needle on outcomes. A variable that scores poorly in satisfaction may not be a key driver if it has little influence on overall loyalty; conversely, a moderately rated feature could be a primary driver of retention. By quantifying the relative importance of each factor, key driver analysis enables targeted interventions that maximise return on investment, whether the goal is reducing churn, increasing Net Promoter Score, or boosting conversion rates.

Example

A telecom company surveys 10,000 customers on 15 service attributes and overall satisfaction. A Shapley-value regression reveals that network reliability and billing clarity are the top two drivers of satisfaction, accounting for 45% of the explained variance. Although customer service response time scores lowest, it ranks only sixth in driver importance. The company reallocates investment from call-centre expansion to network infrastructure upgrades.

Related Terms

• Correlation
• Ordinary Least Squares (OLS)
• Multicollinearity

Software Notes

• SPSS: Analyze > Regression > Linear with standardised coefficients; use the Regression module for stepwise approaches
• R: lm() for standardised coefficients; shapley.value() from the relaimpo package for Shapley values; caret::varImp() for ML-based importance
• Stata: regress y x1 x2 ..., beta for standardised coefficients; shapley user-written package for Shapley decomposition

Contact Us for Support → /contact/