Route Optimisation in FMCG Distribution — Why 21.64 Crore GPS Points Matter

Field force cost is the second-largest controllable expense for most Indian FMCG and dairy distribution companies, after raw materials and trade promotion. A typical 100-salesperson field force runs ₹2-3 Crore in annual cost across salary, fuel reimbursement, and field allowances. Cutting field time and fuel by 18-25% — what our route optimisation module delivers across pilot deployments — translates to ₹40-75 Lakh of annual savings for that size of operation.

This post covers what's under the hood — how the model works, why the 21.64 Crore GPS data points in the SalesPort system matter, and where the technology beats classical "manually planned beats on Excel" workflows.

## The problem — vehicle routing in Indian distribution

Route optimisation in FMCG distribution is, mathematically, a constrained vehicle-routing problem (VRP). Each salesperson visits 30-60 retailers per day. Each retailer has a service-time window (kirana stores open 9-11 AM and 4-8 PM, restaurants prefer 11 AM-1 PM deliveries). Each visit takes 8-15 minutes depending on order volume. Total field time is bounded (10 hours including travel). Routes must minimise driving time while hitting all assigned retailers.

Classical VRP has been studied for 60+ years. Google's OR-Tools library solves it well. The mathematical structure is well understood. So why isn't every Indian distribution company already on optimised routes?

Two reasons:

1. Real-world constraints are messy. The textbook VRP assumes you know travel time between every pair of stops. In Indian distribution, the actual travel time between Retailer A and Retailer B depends on the time of day (traffic), the salesperson's vehicle (two-wheeler vs four-wheeler), and seasonal road conditions. A naïve VRP solver using map-distance produces routes that look optimal on paper and fall apart in practice.

2. Data depth is required. To know the actual travel time between any two stops in your distribution territory, you need historical GPS traces from your own field force on those exact routes. Most distribution companies don't have this. SalesPort clients do.

## What 21.64 Crore GPS points enable

Every SalesPort SFA app emits a GPS data point every 30-60 seconds while the salesperson is active. Across 132+ deployed apps and 2.3 Lakh daily active users, we've accumulated 21.64 Crore GPS data points over 5+ years.

The dataset is structured: salesperson + timestamp + lat/long + accuracy + speed + vehicle type. We've used it to build a travel-time matrix specific to Indian distribution conditions:

• Time-of-day-varying travel times between major retailer clusters in 200+ Indian cities
• Vehicle-specific speed profiles (two-wheeler vs LCV vs delivery van)
• Seasonal corrections for monsoon, festival traffic, election-period restrictions
• City-specific patterns (Mumbai 11 AM gridlock vs Lucknow 11 AM open roads)

The route optimisation solver uses this empirical travel-time matrix instead of map-based estimates. The output is routes that actually work when the salesperson tries to run them — not routes that look great on a map.

## How the module works in production

Three components:

1. Daily route generation. Each morning at 5 AM, the solver runs for every salesperson scheduled to work that day. Input: assigned retailer list, vehicle type, beat constraints, retailer service-time windows. Output: optimised visit sequence with estimated arrival time per retailer.

2. In-day re-optimisation. If a salesperson is running late (GPS shows behind schedule by 30+ minutes), the solver re-optimises the remaining route. Common patterns: skip lower-priority retailers, swap visit order to catch shorter time windows, re-sequence to end closer to the salesperson's home.

3. Beat-plan-aware constraints. Some retailers must be visited on specific days (Tuesday-only beats are common for slower-moving SKUs). Some retailers have minimum-visit-frequency contracts. Some routes must end at the distributor warehouse for cash deposit. The solver respects all of this.

## The 18-25% improvement — where it comes from

Across pilot deployments, route optimisation delivers two efficiency gains:

• Fuel savings of 18-22% from shorter total drive distance per salesperson per day
• Time savings of 15-25% from reduced idle time waiting for retailer service windows

The time savings convert to either more retailers visited per day (capacity gain) or earlier end-of-day for the salesperson (workforce-satisfaction gain). Most clients run it as a hybrid — slightly higher capacity, slightly shorter days.

The gains are larger in dense urban routes (Mumbai, Bengaluru, Delhi) where travel-time variability is highest. In rural routes (UP tier-3, Bihar tier-4) the gains are smaller because driving time is naturally a smaller share of total field time — most of the day is in retailer interactions, not travel.

## Why standalone route-optimisation vendors are hard to beat at the high end

We're honest about the trade-off. Standalone enterprise route-optimisation vendors (Routific, Locus.sh, FieldAssist's route module, Aforza) have built deeper specialisation than us — better solvers, richer constraint handling, more sophisticated real-time re-optimisation.

For top-tier Indian FMCG (top 50 by revenue), those vendors are often the right choice. They're more expensive (₹2-5 Lakh per month) but the operational depth justifies it.

For SalesPort's segment — mid-market FMCG, regional dairies, agri brands, government cooperatives — the math is different. Our route optimisation is +₹15K/month on existing AMC. The lift is 18-25% rather than the 25-30% of high-end specialists. For a ₹2 Crore field force, 18-25% saved is ₹36-50 Lakh annually — paying for the entire SalesPort platform many times over.

## What ships next

Route optimisation is module 2 of seven AI modules. It's currently in pilot with three SalesPort clients; general availability is Q3 2026. See the AI module suite page for the full roadmap.

The technical foundation — real GPS data from real Indian distribution operations — is something no standalone vendor can replicate cheaply. We've been collecting it for five years. The productisation is the wedge.