Use this page to choose the right web app for your question, then click through to run the analysis with guided inputs, visuals, and narrative reporting.
New to statistical analysis? Start with these tools to understand your data before running tests:
Explore a single variable's distribution, central tendency, spread, and outliers. Perfect for understanding your data before running hypothesis tests. See histograms, boxplots, normality tests, and descriptive statistics.
Data: one numeric or categorical column from your dataset.
💡 Pro tip: Always run exploratory analysis (like the Univariate Analyzer) before conducting hypothesis tests. Understanding distributions, outliers, and data quality helps you choose the right test and interpret results correctly.
Match your research question to the right tool based on your data structure:
Unsure where to start? Use the Test Picker helper below or read the "TEST OVERVIEW & EQUATIONS" section inside each app for assumptions, notations, and links to references.
Not sure which test to use? Follow this guided path to find the right statistical tool for your marketing question. Each question narrows down your options based on your data structure and research goal.
Download a complete flowchart of all test selection paths as a study guide
The outcome (also called dependent variable or Y) is what you're trying to predict, explain, or measure. In marketing:
A predictor (also called independent variable or X) is what you think influences or explains the outcome. Examples:
Tip: Start with one predictor to understand the basic relationship, then add more for modeling.
Your predictor's type determines which test you'll use:
This is about whether the same people/accounts appear in both groups:
Key question: Can you draw a line connecting each observation in Group 1 to a specific paired observation in Group 2?
Tools are organized by analysis type. Each card shows the tool's purpose, complexity level, and data requirements.
Explore a single variable's distribution, central tendency, spread, and outliers. See histograms, boxplots, normality tests, and comprehensive descriptive statistics. Perfect for understanding your data before running hypothesis tests.
Data: one numeric or categorical column from your dataset.
Explores whether two continuous marketing metrics move together (for example, ad spend vs. revenue or time on site vs. conversion rate). Includes correlation coefficient, significance tests, and scatterplots.
Data: pairs of numeric values for each case.
Compares average performance between two independent groups when variances and sample sizes may differ (for example, Control vs. New Subject Line). Includes effect sizes, confidence intervals, and power analysis.
Data: group means/SDs and sample sizes, or raw outcomes for each group.
Evaluates before/after or A/B measurements on the same people or accounts (for example, pre-course vs. post-course scores). Tests whether the mean difference differs from zero.
Data: paired numeric observations or paired summary statistics.
Tests whether three or more groups differ on a numeric outcome (for example, four landing page designs vs. revenue per visitor). Includes post-hoc pairwise comparisons.
Data: group means/SDs and sample sizes, or long-format raw data.
Compares conversion rates or other binary outcomes between two independent variants (for example, Newsletter CTA A vs. CTA B). Shows relative lift, confidence intervals, and sample size recommendations.
Data: conversions and sample sizes, or raw 0/1 outcomes by group.
Focuses on "switchers" in paired yes/no data, such as matched pre/post responses (for example, same customers under two creative treatments). Tests marginal homogeneity in 2×2 tables.
Data: 2 × 2 matched table or raw paired outcomes.
Checks whether two categorical variables are related (for example, device type by signup status or segment by offer selection). Shows expected vs. observed frequencies and standardized residuals.
Data: contingency table counts or raw categorical outcomes.
Fits a straight-line relationship between one predictor and one outcome, either for a continuous predictor (for example, weekly ad spend vs. revenue) or a binary group indicator. Shows slope, intercept, R², and residual diagnostics.
Data: summary stats (n, means, SDs, r) or raw paired observations.
Models a numeric outcome using multiple predictors (continuous and categorical) with diagnostics and effect plots (for example, customer value vs. recency, frequency, and segment). Includes multicollinearity checks, residual plots, and coefficient interpretation.
Data: raw rows with one numeric outcome and several predictors.
Models a binary outcome using multiple predictors to estimate odds ratios and predicted probabilities (for example, conversion vs. not convert based on traffic source, offer, and device). Shows log-odds, odds ratios, ROC curves, and classification metrics.
Data: raw rows with one binary outcome and several predictors.
Models a categorical outcome with 3+ levels using multiple predictors, framed as a choice model (for example, product choice based on price, brand, and customer segment). Estimates relative risk ratios for each outcome level.
Data: raw rows with one categorical outcome and multiple predictors.
Segments records into a specified number of clusters based on numeric predictors, assigning each case to the nearest cluster center. Includes elbow plots, silhouette scores, and cluster profiles. Perfect for customer segmentation or market analysis.
Data: raw rows with two or more numeric predictors for clustering.
Calculate probabilities for repeated independent events using the binomial distribution. Answer questions like "What's the probability of 7+ conversions in 23 trials with 10% success rate?" Includes exact binomial, normal approximation, Poisson approximation, Monte Carlo simulation, and extensive educational content.
Data: event probability, number of trials, and target success count.
Analyze text sentiment using lexicon-based scoring methods. Process customer reviews, social media posts, or survey responses to extract sentiment scores and classify text as positive, negative, or neutral. Includes word-level analysis and aggregated statistics.
Data: text strings (reviews, comments, survey responses).
Plans how many observations you need to estimate or detect a change in a single mean or single proportion (for example, overall conversion rate or average order value). Supports both estimation (precision-based) and hypothesis testing (power-based) scenarios.
Input: effect size, significance level, desired power or margin of error.
Plans total sample size for A/B tests comparing two means or two proportions, given the lift that would be worth acting on, your confidence level, and desired power. Essential for experiment planning before launching tests.
Input: baseline rate/mean, minimum detectable effect, alpha, power.
Plans how many paired observations you need to detect a non-zero correlation or simple regression slope between two metrics (for example, media spend vs. revenue or engagement vs. satisfaction). Accounts for expected effect size and desired statistical power.
Input: expected correlation, significance level, desired power.
Plans the per‑arm and total sample size for tests with a control and multiple variants (A/B/C/…), for either proportions or means, under goals focused on minimum lift vs. control and/or any meaningful difference across arms. Handles multiple comparison adjustments.
Input: number of arms, baseline rate/mean, minimum detectable effect, alpha, power.
Shows how simple random, stratified, cluster, systematic, and convenience sampling select individuals from the same population, and how each design affects the composition and variability of sample estimates. Interactive visual tool for understanding sampling methodology.
Interactive simulation: compare sampling methods on the same population.
Explores the probability that specific items are selected from a finite population under sampling with or without replacement, showing exact hypergeometric/binomial math, live equations, and simulated draws. Deepens understanding of sampling probability mechanics.
Interactive simulation: explore selection probabilities with/without replacement.
These apps are designed as teaching tools: they show equations, diagnostics, and APA-style reporting so you can see how the math connects to marketing decisions.