1. Introduction

The global food-delivery industry is still growing rapidly: a study forecasts that the global transaction volume of food-delivery platforms will increase from US$293.6 billion in 2021 to US$ 466.4 billion in 2026 [1]. In China, local life service retail rebounded strongly. In 2023, catering revenue exceeded 5 trillion yuan for the first time, a year-on-year increase of 20.4%, and service retail sales increased by 20.0% overall [2]. According to the latest data from the National Bureau of Statistics, in 2024, the national catering revenue reached 5,571.8 billion yuan, a year-on-year increase of 5.3% [3]. As the traffic dividend peaks and the cost of acquiring customers is high, "acquiring users" has given way to "retaining users", and customer stickiness has become the core indicator of sustainable growth of the platform. The feedback mechanism was originally regarded as a key tool to improve stickiness. However, the conventional "stars-and-text" mechanism generally has the problems of weak incentives and weak interactivity. It lacks substantial rewards and subsequent dialogue space. Users are unwilling to invest time to produce in-depth evaluations, and it is difficult to aggregate scattered opinions into assets that can be used for governance.

To break this bottleneck, Meituan launched the "Xiaomei Jury" in 2020, embedding gamification elements such as points, badges, and rankings into the dispute arbitration process and attempted to improve willingness to interact and information quality through user co-governance; by the end of 2022, the number of users participating in consumer dispute reviews had exceeded 5.4 million [4]. During the same period, the U.S. market likewise shifted into a zero-sum landscape: almost half of American consumers now place repeat delivery orders at least once a week, underscoring the strategic priority of retaining existing users for sustained platform value [5].

Research on how "gamification + crowd arbitration" feedback can improve customer stickiness through psychological pathways (fairness, interactive fun, trust, etc.), the evidence is fragmented. Although Meituan's Xiaomei Jury already engages millions of real users, publicly available quantitative evaluations are confined almost exclusively to the company's CSR reports [4]. In addition, as an emerging governance tool, Crowd-Driven Online Dispute Resolution (CODR) is still in the stage of "practice before theory" [6]. Therefore, systematically integrating interdisciplinary evidence and building a testable conceptual model are of urgent value for improving platform governance theory and guiding practice. Cross-national platform governance comparisons show that the Crowd-Driven Online Dispute Resolution (CODR) mechanism based on a crowdsourcing jury can reduce governance costs while improving platform transparency and trust [6]. However, the effect and boundaries of CODR in the food-delivery scenario still lack systematic testing. This article focuses on the "gamification+crowd arbitration" model, selects Meituan's gamified crowd arbitration and DoorDash, which uses the traditional star rating system, as a comparison, and explores its role and boundaries in improving user stickiness and platform governance effectiveness.

2. Case description

Next, to illustrate how differing feedback architectures operate in practice, this chapter selects Meituan and DoorDash, two representative platforms in terms of market size, business model, and user volume, as comparative cases. The two platforms respectively embody the two mainstream paths of "gamification + crowd judgment" and "non-gamification one-way scoring".

2.1. Meituan

Since 2020, Meituan has embedded the "Xiaomei Jury" in the food delivery complaint workflow- a Crowd-Driven Online Dispute Resolution (CODR) practice with user co-governance as the core [4]. The platform opens a registration channel to customers who have completed at least 5 orders in the last 12 months after real-name authentication; after being selected as a "juror", users need to view screenshots of the order, statements of both parties, and photos of evidence, supporting photos, and then cast a verdict.

To drive continuous participation, Meituan links points, tiered badges, and leaderboards into a closed loop: each ruling earns points that both unlock five badge tiers ("Bronze→Diamond") and can be redeemed for coupons, combining immediate and delayed rewards. Meituan's 2022 Corporate Social Responsibility Report states that the Xiaomei Jury has attracted 5.4 million registered jurors, and the average case-handling time is kept within a few hours [4]. This design achieves decentralized governance and positive incentives in parallel at the process level: the platform still retains the final execution right, but uses gamified rewards in exchange for users' investment in information screening and responsibility allocation, thereby reducing customer service costs and accumulating high-quality dispute information.

2.2. DoorDash

DoorDash still uses a traditional five-star system. After an order is completed, customers must rate the order as a whole and may optionally leave a text review or give a "thumb-up" to individual items. The platform aggregates the 100 most recent ratings to generate each merchant's composite score, which in turn determines search ranking and visibility; it offers no follow-up incentives such as points, badges, or leaderboards to reviewers [5]. If a dispute arises, the user can only choose to "refund, redo, or contact customer service", and the dedicated customer service will make a linear judgment based on the delivery track and photo evidence, and the customer does not participate in the crowd arbitration.

In the United States--a comparatively mature market--DoorDash retains its lead largely through high-speed fulfillment and standardized customer support: in 2023, the company held roughly 45 % domestic share, well ahead of rivals such as Uber Eats [7]. However, its feedback chain is still one-way, low-interaction, and de-gamified, focusing more on efficiency and consistency rather than community governance.

It can be seen that these two platforms have three core differences in feedback design, namely incentive structure, dispute participation, and interaction depth. These differences provide a clear comparative frame for investigating how a "gamification + crowd-arbitration" mechanism shapes user psychology and, ultimately, customer stickiness.

3. Resulting impacts and underlying mechanisms of the design differences

3.1. User-psychology pathways

Meituan's "Xiaomei Jury" links voting in crowd arbitration to a closed loop of "points → badges → leaderboards," which corresponds exactly to the three basic psychological needs proposed by self-determination theory - sense of ability, sense of autonomy, and sense of connection [7]. Users can gain experience points and promotion opportunities after handling each order dispute, and users can intuitively perceive that "I am getting stronger"; the user's ruling directly determines the direction of compensation, so that they have quantifiable decision-making power over platform governance; the platform publicly visualizes personal achievements through badges and rankings, forming peer comparisons and identity symbols, and promoting community belonging. This design provides users with continuous positive feedback of the "challenge-skill" balance, promoting a deep immersive experience (flow), thereby transforming one-time dispute resolution into long-term "service community" participation. By the end of 2022, more than 5.4 million users had taken part in consumer-dispute reviews, demonstrating the feasibility of this psychological mechanism at scale [4].

Once psychological motivation is activated, they tend to unfold along an "intrinsic motivation→attitude→habit→stickiness" trajectory that can be observed behaviorally. Large-sample studies have shown that positive attitudes and habits in the use of mobile applications strongly predict continued use behavior, and the formation of habits is inseparable from high-frequency, context-stable repeated interactions [8]. Meituan uses graded badges to package multiple behaviors, such as repeated voting and active ordering, into a "cultivation line", which lowers action thresholds while reinforcing contextual cues, thereby accelerating the formation of automatic usage habits. The same study pointed out that 57% of new users said they were "very likely to use" food delivery apps again after their first experience, offering indirect support for this "habituation→repeat-purchase" chain [8].

At the same time, crowd arbitration itself has significantly improved the sense of procedural fairness. In service scenarios, the right to participate in decision-making has been shown to significantly improve users' evaluation of the fairness of the platform, thereby enhancing satisfaction and trust [9]. Legal research on the crowd arbitration mechanism of online platforms has also found that allowing community members to jointly interpret and adjudicate can reduce perceived bias to about two-thirds of the original level [10]. In contrast, DoorDash still uses a linear customer service adjudication process. Although customers can obtain remedies such as refunds or re-does, they lack control over the "process" and their trust accumulation relies more on brand endorsement rather than community co-governance narratives. Despite this, the one-way star system can still maintain a certain external stickiness in terms of convenience and speed: DoorDash's 2025 annual user survey showed that nearly half of American consumers (47%) re-order the same restaurant at least once a week [5]. This reminds us that external utilitarian convenience can bring "behavioral stickiness", but if it lacks the support of identity and sense of achievement, its long-term retention effect is often weaker than the "intrinsic motivation-emotional identification" path.

In summary, the "gamification+crowdsourcing arbitration" model has built a deep user stickiness path driven by intrinsic motivation: satisfying users' basic psychological needs, enhancing procedural fairness perception, and promoting users' habitual use. In contrast, the traditional star rating mechanism mainly relies on functional convenience, generating more superficial user stickiness. These two paths have fundamental differences in psychological trigger points, emotional intensity, and sustainability, thus providing theoretical coordinates for the examination of subsequent behavioral outcomes and governance performance.

3.2. Dispute participation

Under the framework of "gamification+crowdsourcing arbitration", Meituan has significantly improved user engagement. The depth of use directly translates into higher visit frequency and retention rate. A nationwide survey of 610 food delivery app users found that 43.1% of users placed orders 1 to 2 times a week, and 36.4% of users placed orders 3 to 5 times a week; the cross-tabulation further showed that platforms that provided high-interaction functions such as real-time order tracking and arbitration participation had a significantly higher proportion of high-frequency users [8]. In the US market, DoorDash also reported that 47% of consumers repeatedly ordered food delivery at least once a week, but their motivation mainly came from convenience and habit rather than the platform's intrinsic incentives [5].

There are also differences between the two groups of data in terms of the purification of negative word-of-mouth. The sentiment-theme regression analysis of 79,003 platform events showed that when the dispute resolution method changed from a one-way process to a multi-party interaction, the overall negative sentiment coefficient of users decreased by 0.22-0.27 (p < 0.001). This suggests that group arbitration and interactive feedback can guide emotional catharsis to evidence-based rationality, thereby suppressing retaliatory comments [11].

Transaction-side data further confirms the difference between the two paths. Without continuous gamification incentives, DoorDash increased its Marketplace GOV from approximately US$14 billion in the first quarter of 2022 to nearly US$22 billion in the fourth quarter of 2024 through category expansion and subsidies [7]. However, the report did not disclose the simultaneous increase in the conversion rate of monthly active users' orders. In China, the merchant ratings in the catering market as a whole moved to the right - the convergence of low-scoring noise coincided with the launch of the "Xiaomei Jury" [3].

3.3. Platform governance outcomes and long-term value

First, the combination of gamification and group arbitration significantly improves users' sense of procedural fairness. Multi-platform quantitative research shows that when disputing parties are given voting rights, perceived bias decreases by about 33% and negative emotions are reduced accordingly [6]. In Meituan's system, customers transform from passive complainers to co-managers of rules and bring this reciprocal mentality into subsequent purchasing behavior, thus entering a self-reinforcing cycle.

Second, group arbitration replaces centralized proxy audits with distributed voting, thereby reducing response time and governance costs. Simulation experiments show that once the voting group reaches or exceeds 21 people, the gap between group decision-making and expert decision-making narrows to less than 5%, while the average processing time is reduced from days to hours [6]. In contrast, DoorDash still relies on a linear customer service workflow, whose speed is limited by staffing levels, and the process lacks transparency.

Third, the high-resolution data generated during group arbitration decision-making provides the algorithm with fine-grained labels that can be used for iteration and merchant profiling. The report "China Catering Annual Observation and Big Data 2025" shows that in 2024, the proportion of high-quality merchants (score≥4.0) will increase by 13.7 percentage points year-on-year, and 44.8% of merchants will continue to improve their scores, indicating that the catering industry is shifting from price competition to quality competition [3]. Although DoorDash's one-way star rating system is easy to aggregate, it lacks these fine-grained labels and is difficult to achieve accurate traffic allocation.

In short, gamification + group arbitration form a reinforcing cycle with user stickiness by satisfying basic psychological needs, reducing governance costs, and enhancing procedural fairness. Although one-way star rating+centralized customer service provides a standardized workflow, it has inherent limitations in deep incentives, trust building, and data value. This comparison provides hypotheses and benchmarks for future empirical testing. For any platform, a feedback incentive mechanism that can simultaneously stimulate users' needs for capabilities, autonomy, and relevance will bring retention benefits far exceeding one-time discounts. In addition, making the dispute resolution process open and community-driven can not only build trust, but also provide high-value training data for algorithm governance--this highlights the long-term potential of gamification+group arbitration as a sustainable platform governance strategy.

4. Integration framework, practical solutions, and future experiments

4.1. Concept integration framework

Drawing on the empirical findings above, the "gamification + crowd-arbitration" feedback can be viewed as a closed loop of motivation activation → process parallelization → dual benefits.

First, on the incentive-design side, points, tiered badges, and leaderboards align with the three self-determination needs--competence, autonomy, and relatedness--and therefore markedly stimulate users' intrinsic motivation [12,13].

Second, on the governance-process side, distributed voting shifts dispute resolution from a serial to a parallel mode, shortening the average handling cycle from "days" to "hours" while also depositing high-granularity behavioral and evidentiary data [6].

Finally, at the outcome level, intrinsic motivation is converted into repeat purchases and community retention, whereas process parallelization simultaneously cuts customer-service hours and heterogeneous-data processing costs, further reinforcing stickiness by dampening negative word-of-mouth spillover [2,11].

So long as a platform keeps tracking a motivation-activation metric (e.g., level-up rate) and a process-efficiency metric (e.g., average case-closure time), it can fine-tune this loop in real time. When transplanting the loop, the platform can proceed in three gradual phases--incentive design, co-governed dispute resolution, and CSR-trust resonance.

Phase 1 was motivation activation. Roll out low-threshold points and tiered badges to the core user base first, converting the one-off star rating into a repeatable task to spark early feelings of competence and autonomy [12,13]. Second was the co-governed process. Identify the dispute category with the highest complaint volume, launch a small-scale crowd-arbitration pilot, and, once consistency and closure rates have stabilized, extend it to the full workflow--thereby cutting misjudgment risk and sharply reducing customer-service costs [6]. Final trust amplification. Embed quantifiable CSR elements--such as carbon-reduction credits and compensation transparency--into the decision screen, and regularly release arbitration statistics, using emotional value to magnify motivation activation and further bolster trust [11,12].

Implemented progressively, this approach lets a platform boost user stickiness, governance efficiency, and data-asset value within a single framework, while also establishing unified metrics for subsequent causal tests and cross-platform scaling.

4.2. Future experimental agenda

To test the cross-context validity of the framework, it outlines three experimental paths. First, a cross-cultural controlled experiment: randomly deploy the "Little American Jury" logic on European and US platforms to compare how autonomy affects loyalty and test the cultural difference hypothesis [12]. Second, a cost-effective natural experiment: using the phased rollout of crowdsourcing arbitration in cities, tracking the net present value of customer service minute costs and incentive expenditures to quantify governance gains and losses [6]. Third, a longitudinal panel study: using user-level logs to observe the time difference between badge upgrade rhythm and repeat purchases, clarifying the causal direction [8]. Together, these experiments will fill the current gap in cross-sectional evidence and provide platforms and regulators with a quantitative basis for decision-making.

5. Conclusion

This paper explores how gamified arbitration feedback can significantly improve user stickiness on food delivery platforms by stimulating users' intrinsic motivation and parallelizing dispute resolution. Focusing on the mechanism of "gamified incentives + crowdsourcing arbitration co-governance", this paper explains how the activation of psychological drivers and the parallelization of governance can jointly improve user stickiness, efficiency, and data value. Combining research results in the fields of behavioral science, platform governance, and data assets, this paper proposes and verifies a closed "motivation activation → process parallelization → win-win" framework, and outlines the implementation sequence of "incentive design → dispute co-governance → corporate social responsibility-trust resonance". The results show that when points, badges and arbitration voting jointly stimulate users' feelings of ability, autonomy and relevance, the monthly repurchase rate will increase significantly, the dispute handling time and customer service costs will also decrease, and the structured labels generated by arbitration will be fed back into the algorithm optimization, shifting the merchant quality score to the right and suppressing negative emotions, forming a virtuous circle. Future research should use cross-cultural randomized experiments and longitudinal panel tracking to test the robustness of the autonomy effect across different markets, quantify the marginal cost-effectiveness impact of arbitration penetration, and clarify the time difference between incentive upgrades and repurchases. As platform governance enters the era of "user co-governance", achieving a dynamic balance between rights protection, efficiency improvement, and public value creation will remain a key challenge; the framework proposed in this article provides a practical theoretical basis and measurement benchmark for subsequent research and practical applications.