When Data Becomes Judgment: Misaligned Interpretations and Accountability in Food Delivery Platforms

• Users treat spatial and temporal anomalies as evidence of courier intent, while couriers alter trajectories to maintain interface legibility.Real-time tracking thus shapes judgment, accountability, and courier behavior, rather than merely communicating delivery progress.Source: §1 · p. 2
• Users translate apparently objective tracking metrics into moralized judgments of individual negligence.The paper calls this interpretive framework “data-as-behavior.”Source: §1 · p. 2
• Couriers engage in anticipatory data management, prioritizing the appearance of their trajectories over physical efficiency to mitigate complaint risks.This response reflects their awareness of being judged through interface traces and helps manage the data shadow.Source: §1 · p. 2
• A dual-perspective qualitative study of semi-structured interviews with 23 users and 17 couriers examines how both groups interpret the same GPS locations, ETAs, and status nodes.The study addresses how users make accountability judgments, how data awareness shapes courier actions, and how systems might resolve these interpretive misalignments.Source: §1 · p. 2
• Current tracking architectures shift the cognitive and emotional labor of explaining systemic logistical failures onto gig workers, motivating contextual transparency that returns explanatory labor to platforms.Interfaces show granular location data but little context about traffic, batching, route adjustments, restaurant delays, or access constraints.Source: §1 · pp. 2–3

2.1 Location Sharing: From Social Coordination to Asymmetric Purpose-Driven Systems

• Food delivery turns purpose-driven location sharing into compelled tracking within an asymmetric transaction, using location transparency for courier management and customer judgment.Earlier location-sharing research emphasized trusted-group awareness and coordination,, while privacy practices were context-dependent, adaptive, and distinguished social from logistical purposes.Source: §2.1 · p. 3

2.2 Algorithmic Management: Information Asymmetry and Worker Adaptation

• Algorithmic management translates delivery work into metrics that shape couriers’ working conditions, while opaque platform logic leaves them adapting orders and routes amid information asymmetry.Platforms possess system-wide behavioral data but provide workers minimal actionable instructions; couriers continually interpret platform logic and manage both deliveries and their relationship with the system, [8],.Source: §2.2 · pp. 3–4

2.3 Infrastructure Studies: Visibility Without Causal Intelligibility

• Delivery interfaces provide tracking visibility without causal context, so violated point ETAs can make systemic delays appear to be individual courier failures.Users see processed outputs but not conditions such as restaurant delays or dispatch logic; high-granularity data without situational interpretability leaves couriers bearing reputational consequences.Source: §2.3 · p. 4

2.4 The Measure-Target Shift: From Quantitative Metrics to Social Judgment

• When ETAs become both performance targets and user promises, interface ambiguity invites moralized judgments and pushes couriers to adapt their behavior to manage customer perceptions.This recursive dynamic distributes interpretive labor and accountability through ratings, complaints, and reviews, converting shared logistical challenges into localized conflict between users and couriers.Source: §2.4 · pp. 4–5

3.1 Research Design

• The qualitative design juxtaposed users’ and couriers’ concrete experiences to examine how interface data acquires behavioral and moral meaning.The analysis focused on differences arising from their distinct informational and practical positions.Source: §3.1 · p. 5

3.2 Participants Recruitment and Demographics

• The team recruited 40 participants—23 users and 17 couriers—through Chinese social media, seeking diverse backgrounds and at least six months’ platform experience.Users ranged from 19–55 years old and couriers from 24–45; users interpreted interface traces while couriers also navigated real-world constraints.Source: §3.2 · p. 5

3.3 Research Ethics

• Participants gave informed consent after learning the study purpose, data use, and withdrawal rights; researchers anonymized transcripts and minimized interview risks.Data were stored on encrypted devices with restricted access.Source: §3.3 · pp. 5–6

3.4 Data Collection Procedures

• Role-tailored semi-structured interviews lasted 45–65 minutes and were audio-recorded with consent.Interviews followed participants’ respective roles in the delivery ecosystem.Source: §3.4 · p. 6

3.4.1 User Interviews.

• User interviews elicited routines and memorable deliveries, then probed noticed cues, judgment formation, data-reality mismatches, responses to uncertainty, and desired future information.Researchers avoided predefined data dimensions in questions to reduce priming.Source: §3.4.1 · p. 6

3.4.2 Courier Interviews.

• Courier interviews explored everyday workflows, memorable deliveries, risk judgments, customer misunderstandings, data-reality mismatches, and expectations for reducing explanatory burdens.Questions examined how couriers combined platform data with physical conditions when deciding how to act.Source: §3.4.2 · pp. 6–7

3.5 Thematic Analysis and Mechanism Abstraction

• Role-separated thematic coding and mechanism abstraction found users often read anomalies as behavioral flaws while couriers read them as environmental or systemic constraints.User and courier datasets reached saturation after participants 18 and 14, respectively; negative cases helped define the boundaries of these mechanisms.Source: §3.5 · pp. 7–8

4.1 From Spatial Traces to Moralized Attribution: How Users Construct Responsibility

• Users convert location, ETA, route, and staged-status metrics into moralized judgments of courier capability and intent when deliveries deviate from expectations.The analysis draws on semi-structured interviews and identifies these four informational dimensions as users’ primary basis for assigning responsibility.Source: §4.1 · p. 8, §4 · p. 8

4.1.1 Location and Movement Status as Behavioral Proxies.

• Users read stationary or slow map movement as intentional delay, mistaking processed traces for objective behavior while overlooking preparation delays, access barriers, and connectivity gaps.When the interface stops updating, users may infer bad faith because it offers no explanation for stasis; couriers can become targets of blame for signal failures or other hidden constraints.Source: §4.1.1 · p. 8

4.1.2 Algorithmic ETAs as Binding Commitments and Reference Anchors.

• Users treat algorithmic ETAs as binding deadlines, organize routines around them, and blame couriers when deliveries miss the displayed time.Precise minute counts and countdowns make dynamic predictions appear guaranteed, although users lack insight into how ETAs are generated.Source: §4.1.2 · pp. 8–9
• Platform policies can turn ETA deviations into default worker fault and financial penalties after customer complaints [8].These local accountability mechanisms give users’ fixed temporal expectations consequences for couriers.Source: §4.1.2 · p. 8

4.1.3 Route Deviations as Indicators of Opportunistic Detours.

• Users interpret routes that depart from the shortest path as opportunistic detours, although most irregular trajectories reflect logistical constraints rather than bad intentions.Interfaces omit batching, pickup sequencing, dispatch logic, and road obstacles; some winding routes do result from stacking too many orders.Source: §4.1.3 · p. 9

4.1.4 Staged Status Nodes as Interpretive Framework for Delivery Progress.

• Linear status milestones imply smooth progress, so stalled or repeated stages prompt anxiety and blame directed at couriers, even when restaurant or platform bottlenecks cause the delay.Couriers are the most visible and reachable human actors when stage transitions falter.Source: §4.1.4 · p. 9

4.2 Managing the Data Shadow: How Couriers Internalize Interface Risk

• Couriers experience tracking metrics as a continuous risk environment and adjust operational behavior to mitigate complaints and manage their data shadows.They anticipate customer interpretations of interface representations rather than treating tracking as a neutral display.Source: §4.2 · p. 10

4.2.1 Anticipatory Adjustments for Interface Legibility.

• Couriers reshape routing and pacing to avoid suspicious-looking pauses, smooth trajectories, and remain visible, sometimes prioritizing interface legibility over task efficiency.They may shorten necessary stops, message customers during restaurant delays, or follow platform navigation rather than faster shortcuts through signal-dead zones.Source: §4.2.1 · p. 10

4.2.2 Data Anomalies as Preemptive Complaint Triggers.

• Couriers preempt complaints about stuck orders, delayed completion, or ETA changes, sometimes prioritizing riskier-looking orders over physical distance or standard routing.They may move or contact customers near risky tracking states, and finish deadline-sensitive orders first to avert user panic.Source: §4.2.2 · p. 10

4.3 Systemic Expectations: Bridging the Asymmetry Between Users and Couriers

• Users and couriers develop stable expectations about how platforms should operate through repeated misunderstandings, conflicts, and coordination practices.Source: §4.3 · p. 11

4.3.1 User Demands for Contextual Transparency and Evidentiary Clarity.

• Users want platform-verified explanations of delays and operational anomalies so they need not speculate about causes during the wait.Source: §4.3.1 · p. 11
• Users also seek standardized evidence, including clear timeout thresholds and detailed records of when and how long delivery slowed.Source: §4.3.1 · p. 11

4.3.2 Courier Demands for Representational Fidelity and Relief from Explanatory Labor.

• Couriers want tracking to reflect actual delivery conditions accurately, preventing glitches or GPS freezes from appearing to users as misconduct.Source: §4.3.2 · p. 11
• Couriers want platforms to automatically share delay information and record operational milestones, reducing the need to explain routine problems themselves.Source: §4.3.2 · p. 11

4.4 Situating Findings within the Chinese Delivery Context

• The study suggests that rapid-delivery expectations and complaint-linked financial penalties in China may intensify users’ reliance on ETAs and couriers’ vigilance about complaint risk.The authors caution that this influence may differ on platforms with more lenient dispute-resolution mechanisms.Source: §4.4 · pp. 11–12
• Because the study collected no data outside China, whether its findings transfer remains an open empirical question, and its proposed broader relevance is provisional.The authors suggest that unexplained interface outcomes and user-facing tracking may occur elsewhere, while cultural factors may amplify couriers’ trajectory adjustments; these interpretations were not tested comparatively.Source: §4.4 · p. 12

5 Discussion

• Food delivery platforms function as ubiquitous computing infrastructure, continuously translating contingent physical activity into interface representations.Source: §5 · p. 12
• Food-delivery tensions extend beyond interpersonal friction, shaped and amplified by real-time tracking design and the visibility it produces.Source: §5 · p. 12
• The discussion analyzes the burden-shifting mechanism and data-shadow paradox, then considers implications for socially responsible ubiquitous-system design.Source: §5 · p. 12

5.1.1 Visibility Without Causation.

• Tracking interfaces make delivery outcomes highly visible but conceal operational causes, so users read static locations and ETA drift as courier failure, especially amid continuous updates.Static GPS markers, precise ETA countdowns, route visualizations, and status nodes reveal outcomes without explaining delays; real-time updates encourage users to treat stalls or deviations as behavioral signals.Source: §5.1.1 · p. 13

5.1.2 Anticipatory Data Management and the Data Shadow.

• Couriers practice anticipatory data management, reshaping routes, pacing, stops, sequencing, and messaging to make sensor traces appear reasonable to an evaluative user audience.Unlike impression management, emotional labor, or backend algorithmic compliance, this is tactical work on physical actions represented through sensor traces; optimizing for how deliveries look can displace efficient routes.Source: §5.1.2 · pp. 13–14

5.1.3 Platform Design as a Burden-Shifting Mechanism.

• The interface’s burden-shifting mechanism leaves users judging without context and couriers repairing misunderstandings through manual explanations, while platform logistics remain outside user scrutiny.Users asked for system-verified reasons and couriers for platform-led explanations; participants also suggested interval ETAs or coarse status annotations, but interviews cannot establish why context remains absent.Source: §5.1.3 · pp. 14–15

5.2.1 From Bare Visibility to Contextual Transparency.

• Platforms should replace bare tracking with selective, event-triggered context, including interval ETAs and system-verified explanations for route or status anomalies.Possible interfaces include segmented preparation and transit estimates and sensor-informed status icons; inference may be unreliable, while excessive detail can overload or worry users.Source: §5.2.1 · pp. 15–16

5.2.2 Redistributing the Explanatory Burden to Platforms.

• Platforms should automate explanations for system-induced delays and use reconstructable, contestable logs for disputes, rather than relying on couriers’ unpaid explanations.Platform explanations should support worker review and appeal so they do not become another opaque layer of control.Source: §5.2.2 · p. 16
• A proposed report-difficulty feature would corroborate courier obstacle reports with nearby or historical traces and issue customers a verified explanation, but is especially speculative and vulnerable to gaming.Isolated couriers may lack corroboration, and unverified reports could appear suspicious; peer coordination can game thresholds [8].Source: §5.2.2 · pp. 16–17

5.2.3 Data Mirroring and Dual-Sided Design.

• Selective data mirroring could show couriers the customer-facing ETA and movement status, helping them respond when tracking misrepresents their situation.Previews and anomaly alerts should be limited and worker-controlled because mirroring may intensify self-monitoring or be repurposed for performance management; workers need agency to contextualize and challenge data.Source: §5.2.3 · p. 17

5.2.4 Institutional Feasibility, Incentives, and Design Scope.

• The proposals are conditional accountability requirements, not guaranteed fixes: platforms may resist ceding informational power, and design cannot resolve labor precarity or asymmetric business models.Institutional reluctance particularly affects contestable logs and mirroring, whereas corroboration gameability is a distinct mechanism risk that may be addressed through threshold redesign.Source: §5.2.4 · pp. 17–18

Limitations, Transferability, and Future Work

• Retrospective interviews may reconstruct or soften accounts of workarounds, and the study lacks matched user–courier pairs assessing the same incident in real time.The authors treat courier accounts as situated interpretations rather than complete behavioral records and recommend direct observation, platform logs, matched cases, and in-situ event tracing.Source: §5.3 · p. 18
• Findings come from Chinese food delivery and the user–courier dyad, limiting coverage of varying local norms and other actors such as restaurant staff, dispatch teams, and platform designers.Future work should compare countries and platforms, examine broader stakeholder roles, use diary studies or interface-event logging, and include platform engineers’ perspectives on design intent.Source: §5.3 · p. 18
• Simplified tracking traces may invite accountability judgments under uncertainty across contexts, though the mechanisms’ specific manifestations may vary by region.The authors connect this possibility to global tracking systems’ minimal location traces and reliance on situated human interventions that interfaces do not capture.Source: §5.3 · p. 18
• Future comparative and in-situ research should establish which findings generalize and when tracking states trigger judgment or behavioral alteration.Proposed approaches include diary studies and interface-event logging, alongside broader stakeholder inquiry into how misalignments are institutionally produced and sustained.Source: §5.3 · p. 18
• Tracking infrastructures expose outcomes while obscuring causes, prompting user judgments and requiring couriers to manage their data shadows; future systems should distribute interpretive labor equitably.The authors call for an “Infrastructure of Care” and technical prototype verification, including sensor fusion such as barometer data to resolve interpretive misalignments.Source: §5.3 · p. 18

6 Conclusion

• Users turn tracking traces into moralized judgments, while couriers engage in anticipatory data management to navigate risks posed by their data shadows.Together, these dynamics form a burden-shifting mechanism: couriers bear explanatory labor that platform-held logistical data could support, though this may not be the design’s intended purpose.Source: §6 · p. 19
• Participants share a demand for contextual transparency, so platforms take on more of the explanatory burden currently borne by users and couriers.Source: §6 · p. 19
• The authors call for Infrastructures of Care that distribute interpretive labor equitably and ensure data serves coordination rather than judgment.Source: §6 · p. 19

Ethics Statement and AI Usage

• The authors report no ethical issues and used Google Gemini 3.1 Pro to improve writing quality.The prompt asked Gemini to polish supplied text.Source: Ethics Statement and AI Usage · p. 19

Participant Demographics

• The supplement includes detailed participant demographics for the study.The demographics are presented in Table 1.Source: Supplementary Material · p. 23, Table 1 · p. 27

Interview Guide for Users

• The user guide probes mismatches between displayed data and reality, accumulated judgments, uncertainty, records that support judgment, and ideal future systems.It also asks what information users want future systems to explain and retain when something goes wrong.Source: Table 2 · p. 23, Supplementary Material · p. 24

Expectations for Future Systems

• Both guides ask what a more reasonable future delivery system should show or explain to make the process easier to understand and reduce misunderstandings.The courier guide specifically asks which explanations the platform should provide instead of requiring couriers to repeat them themselves.Source: Supplementary Material · pp. 24, 26

Interview Guide for Couriers

• The courier guide asks how couriers interpret platform information, anticipate customers’ views, recognize data-reality mismatches, and respond to possible misunderstandings or complaints.Probes include changing route or speed and contacting the customer.Source: Supplementary Material · p. 25
• The courier guide also explores delivery prioritization, when couriers explain or contact customers, experiences of being understood or misunderstood, and how work experience changes over time.It asks about decisions beyond distance or efficiency and how couriers’ understanding of the process changes with experience.Source: Supplementary Material · p. 25