Pricing Power in the Context of Competition Between Platform Sales and Live Streaming Sales

The study addresses who effectively holds product pricing power in the competitive context of live-streaming sales versus traditional e-commerce platforms. As live-streaming reshapes online retail, conflicts over price control intensify, exemplified by the 2023 JD vs. Li Jiaqi dispute over a Hauswirt oven and a reserve-price agreement. Although brands typically dominate product pricing, streamers (who command traffic and influence) and platforms (which control orders and subsidies) both seek greater pricing say, creating potential for either beneficial competition or harmful price wars. The paper aims to identify ideal equilibrium states among brands, streamers, and platforms; the conditions under which they arise; and how strategic choices and key factors drive the system’s evolution. Using a "people-goods-scene" lens, the authors model interactions to inform win–win pricing strategies in dual online channels.

The review covers three strands: (1) Dual-channel supply chains: extensive work on channel competition and pricing under various power structures, channel choices (direct online vs. retail vs. platforms), consumer behavior (price sensitivity, reference price effects), and coordination mechanisms. Studies analyze Stackelberg games, online-offline interactions, price discrimination, and sustainability considerations. (2) Live-streaming e-commerce: research on pricing/coordination in live-streaming supply chains, streamer influence and effort, sales model selection, and multinational channel structure. Additional studies examine consumer purchase intentions, trust, and interactivity effects in live commerce. (3) People–goods–scene perspective: originating from new retail theory, it emphasizes integrating users (people), personalized goods, and scenario-based consumption. Prior academic work is largely conceptual or case-based, with limited modeling. The gap identified is a lack of in-depth game-theoretic analysis of pricing power and channel competition between live-streaming and traditional platform channels from the people–goods–scene perspective.

The authors build a three-party evolutionary game model comprising a brand (b), a streamer (a), and a traditional e-commerce platform (p). Each player is boundedly rational and selects between two strategies: Brand—accept low-price negotiation (probability α) or not (1−α); Streamer—bargain/pay out-of-pocket to lower price (β) or not (1−β); Platform—provide subsidies (γ) or not (1−γ). Key assumptions and parameters include fees and costs (streamer pit fee H, commission t; platform drainage r, margin c, selling cost k, fixed cost M), sales and wholesale quantities (K1/K2, V1/V2, Q1/Q2), platform profits (J1/J2), compensation under vicious competition (Ua from streamer, Up from platform), negative utilities of vicious competition (Nb for brand, Na for streamer, Np for platform), positive utilities of healthy competition (Rb, Ra, Rp), probability of vicious competition g, and traffic-related loss/diversion when the brand accepts low-price negotiation without channel competition (μb, ω; λ=μbω). The model specifies payoffs for all eight strategic combinations (payoff matrix) and derives replicator dynamics for each player based on expected and average payoffs. For the brand, streamer, and platform, the authors obtain replicator equations F(α), F(β), and F(γ), analyze first derivatives, and define auxiliary functions to establish monotonicity and thresholds. Propositions characterize best-response dynamics: thresholds for the streamer’s bargaining relative to platform subsidies (γ*), the platform’s subsidy decision relative to brand negotiation (α*), and conditions under which the brand accepts low-price negotiation (β*). The stability of pure-strategy equilibria is examined via the Jacobian and eigenvalues at eight vertices E1–E8. Numerical simulations (MATLAB R2020b) validate theoretical results using two parameter arrays calibrated to realistic settings. Array 1 corresponds to a regime with higher risk of vicious competition and lower incremental benefits, evolving to E1=(0,0,0) {no negotiation, no bargaining, no subsidy}; Array 2 corresponds to a regime with stronger healthy-competition benefits and platform gains, evolving to E8=(1,1,1) {accept negotiation, bargain, subsidize}. Sensitivity analyses investigate effects of compensation (Ua, Up), negative utilities (N•), positive utilities (R•), and initial strategy probabilities on evolutionary paths and convergence speed.

- Two evolutionary stable strategies (ESS) emerge: (1) Brand does not accept low-price negotiation, streamer does not bargain, platform does not subsidize (E1=0,0,0); (2) Brand accepts low-price negotiation, streamer bargains, platform subsidizes (E8=1,1,1). - Brands frequently avoid explicitly rejecting low-price negotiation. Acceptance likelihood increases with higher compensation from streamer/platform under vicious competition (Ua, Up), but decreases with larger losses to the brand (Nb) and higher probability of vicious competition (g). - Streamer decision: there exists a subsidy-related threshold γ*; when γ>γ*, streamers bargain; when γ<γ*, they do not. Higher positive image gains from healthy competition (Ra) encourage bargaining; higher compensation obligations, image losses (Na), and g discourage it. - Platform decision: with low g, platforms tend to subsidize regardless of others’ strategies; with high g, a brand-negotiation threshold α* determines whether to subsidize (α>α* leads to subsidies). - Threshold effects: Sufficiently high positive utility from healthy competition (Rb, Rp, Ra) can shift the system to E8; sufficiently high negative utility from vicious competition (Nb, Np, Na) pushes the system toward conservative E1. - Initial conditions matter for convergence speed and transient paths; under certain initial states, the system may transiently visit unstable equilibria (e.g., E2 or E4) before converging. - Numerical simulations confirm theory: parameter Array 1 evolves to E1; Array 2 evolves to E8. Changes in compensation mainly affect convergence speed rather than the final equilibrium under given regimes.

The study shows that while brands hold primary pricing authority, realized pricing power is endogenous to strategic interaction among brands, streamers, and platforms. Pricing outcomes hinge on whether competition is healthy (expanding market, improving brand value and platform reputation, and enhancing streamer image) or vicious (triggering compensation, reputational losses, and reduced channel profits). The people–goods–scene perspective clarifies why streamer traffic (people), product positioning (goods), and interactive scenarios (scene) jointly shape traffic allocation, negotiation leverage, and ultimately pricing power. The findings imply that brands should time acceptance of low-price negotiations to harness healthy competition’s upside while constraining vicious competition risks; streamers and platforms should calibrate bargaining/subsidy intensity to avoid destructive price wars. In equilibrium, pricing power is shared: brands set the tone, but streamers’ bargaining and platforms’ subsidies can collectively pull prices toward either conservative or aggressive outcomes depending on utilities and risks.

This paper develops a brand-led tripartite evolutionary game model of pricing competition between live-streaming and traditional platform channels through a people–goods–scene lens. Analytical stability results and simulations reveal two ideal equilibria: conservative (no negotiation, no bargaining, no subsidy) and concessive (accept negotiation, bargaining, subsidy). The system’s state depends on compensation levels, risks and costs of vicious competition, and benefits of healthy competition, as well as initial strategic intentions. Managerially, even when brands dominate pricing, win–win pricing strategies require coordinating with streamers and platforms to amplify healthy competition and prevent price wars. The study contributes a modeling framework for channel-pricing power in live commerce and offers actionable guidance for dual-channel pricing decisions.

The model abstracts from direct consumer decision-making and regulatory intervention; these actors may materially affect pricing dynamics. Technology-related trust and quality issues (e.g., via blockchain) are not modeled. Negotiation processes between supply chain tiers are simplified into binary strategies without explicit bargaining protocols. Parameterization and simulations, while illustrative, are not empirically estimated; external validity may be constrained. Future research could incorporate consumers and regulators as strategic agents, explore blockchain-enabled trust and quality mechanisms, and model explicit negotiation procedures.