Key Analytics Formulas & Benchmarks

After the Analytics Day 0 Cheat Sheet, the next priority is fast recall of the formulas and benchmarks that repeatedly appear in analytics interviews. Key Analytics Formulas & Benchmarks helps candidates revise the bias-variance trade-off, core machine learning evaluation formulas, and India-relevant business analytics benchmarks in one place.

• Underfitting = Model too simple; misses patterns in training data (high bias).
• Overfitting = Model memorises training data; fails on new data (high variance).
• Accuracy = (TP + TN) / (TP + TN + FP + FN) - Classification accuracy can be misleading with imbalanced classes.
• F1 Score = 2 × (Precision × Recall) / (Precision + Recall) - Harmonic mean of Precision and Recall - use with imbalanced data.
• RMSE = √[ Σ(y_actual - y_predicted)² / n ] - Root Mean Squared Error - penalises large errors more.
• CLV = AOV × Purchase Frequency × Customer Lifespan - Customer Lifetime Value - a North Star for retention strategy.

Big Picture Overview

This cheat sheet fits together in three parts: the bias-variance trade-off visual, key formulas for model evaluation and business analytics, and common analytics formulas with Indian context benchmarks. Use it to quickly connect model behaviour, metric selection, and business performance indicators.

Bias-Variance Trade-off Visual

Underfitting = Model too simple; misses patterns in training data (high bias).

Overfitting = Model memorises training data; fails on new data (high variance).

Regularisation (Ridge/Lasso), cross-validation, pruning, more data.

Key Formulas

These formulas cover classification, regression, variance explained, and retention strategy. TP means True Positive, TN means True Negative, FP means False Positive, and FN means False Negative.

CLV = AOV × Purchase Frequency × Customer Lifespan. Customer Lifetime Value - a North Star for retention strategy.

Common Analytics Formulas - Indian Context

These business analytics formulas connect acquisition, retention, advertising, customer satisfaction, margin, and engagement with India-relevant benchmarks.

All salary and market figures are tentative based on open-source data. Actual figures may vary.

Conclusion

Key Analytics Formulas & Benchmarks is a Day 0 revision sheet for moving from model behaviour to metric formulas to business benchmarks. The core takeaway is to match each formula to the problem it is meant to evaluate, especially when accuracy can be misleading with imbalanced classes.

The most frequent error is relying on classification accuracy without checking whether classes are imbalanced. Accuracy can be misleading with imbalanced classes, so candidates should connect Precision, Recall, and F1 Score to the actual evaluation problem.