Aditya Textile
TextileGrade

The Problem

Aditya Textile faced a critical operational challenge on their weaving mill floor in Gujarat: a lack of high-fidelity, real-time performance data. Production reporting was run on manual paper logs, machine-level defects went unlogged until post-folding quality checks, and floor supervisor shifts suffered from high latency in identifying throughput bottlenecks.

Without granular operator performance metrics, incentive-based compensation was calculated using subjective estimates. This created two core issues: high-performing weavers were under-incentivized, leading to talent attrition, while operators running older or misaligned looms were unfairly penalized by absolute quotas that ignored machine age and yarn quality variance.

Compounding this was a hard physical constraint: the weaving rooms suffer from severe electromagnetic interference and patchy network coverage. Any cloud-reliant reporting system would inevitably fail during critical shift handovers.

What We Engineered

Sonder developed and deployed TextileGrade — an offline-resilient, floor-level performance tracking and mathematical worker grading system built specifically for Aditya Textile.

TextileGrade shifts Aditya Textile's factory floor from static quotas to a dynamic, cohort-based efficiency model. The system monitors production metrics at every step, tracks equipment faults, and runs a mathematical relative ranking engine. Operators are graded (A through D) daily by comparing their yield and quality metrics against the active average of their peer cohort, automatically normalizing for machine variance, shift conditions, and yarn-grade difficulties.

• Offline-First Room-to-Cloud Sync: Built with an active-local Room/SQLite store and Android's native WorkManager API. The application queues all telemetry, doff logs, and defect events locally, executing a structured, conflict-resolved sync protocol once a stable link is established.
• Single-Screen Supervisor Console: Replaced slow multi-step workflows with a unified Jetpack Compose dashboard. Supervisors can log Worker IDs, select active Looms (Machine IDs), record shift types (A, B, or C), and input precision roll measurements (Grige Number, Taka Number, Roll Number, Doff Meters, and Folding Meters) in under 12 seconds.
• Weighted Defect Scoring System: Integrated localized defect tracking directly into the supervisor input flow. The system maps floor anomalies—including Thick/Thin Places, Kumas, Karli, Reed Marks, and Short Picks—directly to the operator's shift log with predefined penalty weights.
• Dynamic Relative Grading Engine: Ranks performance mathematically. Instead of measuring against fixed targets, the system calculates percentiles dynamically, comparing worker defect points against their peer cohort. A weaver on a legacy loom is graded fairly relative to peers running similar equipment.
• Worker Trust Kiosks: Deployed localized, read-only displays on the mill floor. Operators can check their real-time performance grades and accumulated metrics, resolving payment disputes and establishing transparency.

How We Deployed

Rather than writing speculative specifications from a distance, our engineering pod embedded on the mill floor in Gujarat for the first three weeks. We stood alongside supervisors, observed shift handovers, and mapped the exact ergonomic constraints of glove-wearing operators entering data in loud, high-vibration weaving rooms.

We designed the offline architecture to handle multi-day network blackouts. The synchronization protocol handles data conflicts deterministically through a supervisor-adjudicated review queue for edge cases, ensuring that no metric or shift log is lost.

From initial repository initialization to full production deployment on active tablet terminals took 3 weeks. If you want to know how the system works internally, contact us.

The Outcome

Within 60 days of deployment, Aditya Textile achieved a 28% increase in overall floor throughput, driven by transparent, data-backed performance incentives. High-performing weavers saw a direct correlation between their effort and compensation, leading to a marked drop in attrition.

The weighted defect logging system transformed mill maintenance from reactive repairs to proactive calibration. Real-time logging of Reed Marks and Short Picks enabled the floor team to schedule preemptive loom servicing, reducing unplanned machine downtime by 34%.