Case Study: Scaling Field Ops | Sharanya Chaturvedi

The Context

Managing 2,000 monthly installs
on "hope and spreadsheets."

Technicians ran on WhatsApp. Dispatchers juggled Google Sheets. Management had zero visibility into the last mile.

0%

Real-time Visibility

8–12hrs

Manual Data Entry

35%

Job Redo Rate

High

Manager Burnout

3+

Calls per Install

₹700

Cost per Install

015 people, 2–3 hours every morning, assigning jobs from Excel — one row at a time.

02Technicians worked off WhatsApp addresses and called from personal numbers.

03Consumers got zero information — no confirmation, no ETA, no tracking.

04Every install generated 3+ support calls the agent couldn't answer.

05Every install cost ₹700. No platform, no tracking, no data.

Research across the ecosystem

"I went on 10 installations before opening Figma."

Installs, support shifts, ops meetings — four people to build for:

engineering

The Technician

"Address from WhatsApp, plan my own day, prove completion on WhatsApp again. No tool."

Key Outcome: Offline-First Mobile App

support_agent

The Support Agent

"I piece together job status from WhatsApp threads. No real answer for the customer."

Key Outcome: Real-time Job Tracking

dashboard_customize

The Ops Manager

"2 hours every morning assigning jobs across WhatsApp, Excel, and tickets."

Key Outcome: Auto-routing Dashboard

home

The Customer

"Took a half-day off. No confirmation, no ETA. Support doesn't know either."

Key Outcome: Uber-like Tracking Web-app

Product Ecosystem

From MVP to Engine

Technician App

Today's Jobs — order, time, location

01

The Last-Mile Tool

A React Native app that replaced 12 WhatsApp groups with one structured workflow. Built for variable tech-literacy — colour-coded CTAs, no reading required.

check_circle

Offline Data Sync

80% of data submitted offline, synced later — works in zero-signal basements.

check_circle

Mandatory Geo-capture

No GPS lock, no completion — kills ghost jobs.

check_circle

One-tap Call Masking

Central number — personal numbers never exposed.

Ops Dashboard

All Jobs Dashboard — the central ops view

02

Admin Control Tower

The brain of the operation. A custom distance matrix assigns 200+ jobs daily — one click replaces the morning Excel session.

check_circle

Proximity Auto-assignment

Nearest qualified tech, assigned in one click.

check_circle

12 Actionable Statuses

Each triggers automatic WhatsApp/SMS to the consumer.

check_circle

3-Tier Prioritisation

TAT-expiring → Premium (₹50k+) → Future. 3 failures auto-escalate.

Engineering Highlight

Dynamic Routing — best technician in one click

03

The Routing Engine

A→B distance caching, Haversine fallback, Ola Maps backup. Scales to 400 technicians inside Google's free tier — ₹2.5L/month to ₹0.

check_circle

Permanent Distance Cache

No A→B pair calculated twice — 80% fewer API calls.

check_circle

Inventory-aware Blocking

No parts, no assignment — no wasted trips.

check_circle

Mileage & Fuel Tracking

Routes log timestamps and reimbursement automatically.

PM Decision Log

Hard Decisions, Scalable Wins

Strategic

Build vs Buy for Routing

▸Market tools: ₹2–3L upfront + per-tech fees that explode at scale
▸No vendor could own our IoT-linked routing

settings_suggestBuilt in-house in 8 weeks

Crisis

The ₹2.5L API Bill Anomaly

▸25 techs → ₹2.5L Google Maps bill in one month
▸Cause: zero caching — same A→B recalculated every call
▸Fix: permanent cache + Haversine + Ola Maps backup

trending_down80% reduction in API costs

Process

Radical Transparency

▸Ops wanted to hide failures — fewer updates, fewer escalations
▸Data showed the opposite: hiding failures drove repeat calls

groupReal ETAs + real updates

Prioritisation

3-Tier Job Ordering

1TAT-expiring — committed today, always first
2Premium (₹50k+) — go second
3Future/rescheduled — negotiable, go last
▸3 failures → Needs Intervention queue

Scoping

Saying No to Everyone

Occurrence × Impact ÷ Build cost

✕Auto-replacement orders — deferred to Phase 2
✕Custom filter views — too complex for MVP
✕Consumer self-rescheduling — deferred entirely
✕Holiday calendar UI — Excel upload saved weeks

stars

Owning the Launch

I don't just ship it — I own the launch. On the floor post-go-live, training before every feature, an owner on every escalation. Then, as the product matured, I built the Product Support function.

507

Users supported

5 roles

Trained & onboarded

0

Dropped threads

The Impact

Numbers that moved the business.

Nov 2024

Conception & build

Dec 2024

MVP live — 15 technicians

May 2025

Phase 2 — routing v2, scale

Oct 2025

100 in-house technicians

Now

17–18K jobs/month, ₹0

assignment_turned_in

1,50,000+

Total jobs processed

across the platform, end to end

trending_down

64%

Reduction in install cost

₹700 arrow_forward ₹250

savings

₹0

Platform running cost

scales to 400 technicians, free tier

verified

96.4%

Job completion rate

96.4%

bar_chart

20%

Before

→

45%

After

Installable portfolio share of revenue — single biggest growth driver

trending_up

8–10×

Volume Growth

2K → 15–20K jobs/mo

engineering

10×

Technician Scale

10 → 100 in-house

group

507

Active Users

across 5 roles

speed

17–18K

Jobs / Month

at ₹0 platform cost

confirmation_number

83

Jira tickets shipped
across 2 phases

swap_vert

5 people · 10 techs

Before

10 people · 100 techs

After — 10× scale, 2× headcount

currency_rupee

₹0

Platform running cost
until 400 technicians

Scaling Field Ops
from WhatsApp to
150k Jobs @ ₹0 Cost

Managing 2,000 monthly installs
on "hope and spreadsheets."