Marketing & Growth

Empirical analysis of viral cascades on a retailer's recommendation network; foundational for understanding k-factor and referral mechanics.

VAR model showing WOM elasticity is 20-30x higher than traditional marketing channels.

Decomposes viral influence across awareness, interest, and decision stages; practical framework for campaign design.

Referred customers are 16% more valuable with higher margins and retention. Foundational empirical study on referral program economics. ~1,600 citations.

Field experiments show hub/bridge seeding outperforms random targeting for viral campaigns. ~1,400 citations.

Randomized trial on 1.4M Facebook users; passive-broadcast generates 246% more contagion than active-personalized messaging. ~1,800 citations.

Seminal formalization of direct/indirect network effects and compatibility choice; 6,800+ citations.

Original network effects paper introducing critical mass concept for telecommunications.

Accessible synthesis on network effects, installed base dynamics, and platform competition.

Introduces 'excess inertia' and coordination failure; explains why inferior standards persist.

THE foundational two-sided markets paper; pricing structure, chicken-and-egg dynamics. ~8,000 citations.

Introduces competitive bottlenecks—when one side multi-homes, platforms compete fiercely on single-homers. ~4,500 citations.

Formalizes platform launch problem; divide-and-conquer strategies for entrants. ~1,800 citations.

14 empirical generalizations about promotion effects (brand switching, stockpiling, category expansion); most-cited promotion paper.

Econometric model for coupon effects showing self/cross-coupon elasticities; practical for optimization.

Comprehensive update to 1995 generalizations; bridges academic research with practitioner insights.

8-year panel showing promotions increase price sensitivity while eroding brand equity. ~2,500 citations.

Foundational theory: coupons as self-selection price discrimination devices. ~1,800 citations.

Foundational Google paper introducing Bayesian MMM with adstock/saturation curves; basis for LightweightMMM.

Extends Bayesian MMM to geo-level data; provides tighter credible intervals and reduces targeting bias.

Technical documentation for Meta's open-source Robyn; covers ridge regression, evolutionary optimization, calibration.

Methods for calibrating MMM with experiment results; bridges MMM and incrementality testing.

Introduces persistence modeling via unit-roots and impulse response; distinguishes short-term from permanent effects. ~2,200 citations.

INFORMS Practice Prize winner; 1-4% profit improvement from optimized allocation vs. heuristics. ~650 citations.

Promotions have virtually no permanent effects—short-term ROI metrics mislead. ~1,200 citations.

Original uplift paper introducing the 'Persuadables' segmentation framework (Persuadables, Sure Things, Lost Causes, Sleeping Dogs).

Introduces Qini curves for uplift model evaluation; practitioner-focused methodology comparison.

Causal forests for HTE estimation with valid confidence intervals; foundation for grf package; 2000+ citations.

Adapts information theory for uplift decision trees; extends to multiple treatments; widely implemented.

Introduces S-learner, T-learner, and X-learner; optimal for imbalanced treatment groups. ~1,100 citations.

Foundation for causal ML—Neyman-orthogonal scores + cross-fitting enable root-n consistent inference. ~4,000 citations.

DR-learner with minimax optimality bounds for heterogeneous effects. ~100+ citations.

Bayesian multi-touch attribution with carryover/spillover across funnel stages; AMA/MSI Paul Root Award. ~600 citations.

Seminal conceptual framework for touchpoints across pre-purchase, purchase, post-purchase. ~3,500 citations.

Foundational 5-challenge framework for omnichannel research. ~2,000 citations.

Markov chain attribution showing substantial divergence from last-click heuristics. ~350 citations.

Foundational best practices from thousands of experiments at Microsoft/Amazon; OEC design, validity threats. ~1,200 citations.

Handles network interference via cluster randomization; exponentially lower variance under spillovers. ~600 citations.

Establishes Thompson Sampling as practical standard for explore-exploit in ad optimization. ~1,500 citations.

Defines always-valid p-values maintaining validity regardless of peeking; solves continuous monitoring problem. ~400 citations.

CausalImpact methodology for geo-experiments when individual randomization impossible. ~1,400 citations.

Artificial progress (pre-stamped cards) significantly increases completion; foundational loyalty psychology. ~1,800 citations.

Framework identifying controllable CLV drivers; outperforms RFM models. ~2,400 citations.

IV correction shows 86% of apparent LP effect disappears when controlling for self-selection. ~900 citations.

Heterogeneous effects: heavy buyers don't change; light buyers benefit most from programs. ~1,400 citations.

Causal evidence reviews affect sales; 1-star reviews hurt more than 5-star help. ~4,500 citations.

High-arousal emotions (awe, anger) increase sharing; positive content more viral. ~3,500 citations.

WOM dispersion (not volume) predicts growth; conversations across communities matter most. ~2,200 citations.

5,800+ campaigns: originality, follower size, sponsor salience enhance effectiveness; inverted-U for activity. ~300 citations.