Route optimization: Google Maps, GPS and their limits

Can you do route optimization with Google Maps? For one vehicle and a few stops, partly: the app accepts 9 stops, final destination included, in the order you choose. Google’s APIs can reorder up to 25 stops. As soon as you need to meet delivery slots (time windows) or a load capacity, or to split stops between several vehicles, GPS is no longer enough. The limits below were checked in Google’s documentation on 18 October 2026, and a worked example shows the gap.

At a glance

  • Google’s limits: the app accepts 9 stops, final destination included, and the order is changed by hand (Google help, desktop); the APIs go up to 25 intermediate waypoints, 98 with an experimental feature enabled on request.
  • Worked example: with 9 stops and one vehicle, the shortest order saves 12.6% of distance compared with an order chosen “by eye”. Meeting three time windows costs 9.6 km more (+22%).
  • Beyond that: time windows, capacity, several vehicles, driving time and breaks call for a route planning engine (software that calculates routes under constraints).

What Google Maps allows for a multi-stop route

You need to tell the consumer app apart from the APIs Google offers to developers. As found in Google’s documentation on 18 October 2026:

ToolNumber of stopsStop orderConstraints described
Google Maps app (desktop)Up to 9, final destination included; help section written for desktop, nothing on Android and iPhoneChanged by hand (drag and drop); no automatic reordering describedNone
Routes API (route calculation)Up to 25 intermediate waypointsoptimizeWaypointOrder option: reorders to reduce the cost of the route, favouring travel time and taking other factors into account, such as distance and number of turnsNo time windows or capacity in the option’s documentation; via waypoints are not accepted
Experimental “Routes Preferred” feature98 points as coordinates, 25 if a point is a place IDSame principlePreview (“pre-GA”), not enabled by default, activated through Google support; beyond 25 points: straight-line distance under 1,000 km and driving mode

Google’s help states that directions to several destinations exist for all modes of transport except public transit and flights.

Two points:

  • The old Directions API, which many tutorials still cite, has “Legacy” status: since October 2026, it can no longer be enabled on a new project (documentation). Google recommends the Routes API.
  • Requests with reordering are billed at a higher rate than a simple route calculation.

Finally, Google offers developers a route planning service separate from these two APIs. Its documentation describes support for several vehicles, time windows, vehicle capacity, and drivers’ working hours and breaks. It is used through API requests, with a Google Cloud billing account, and it is billed per shipment. The pages we read do not state a cap on shipments or vehicles.

GPS or route optimization: two different problems

A GPS looks for the best path between two points, taking traffic into account. Route optimization decides the order of the stops and how they are split between vehicles, under constraints: time windows, capacity, driving time. This is known as the Vehicle Routing Problem (VRP). The number of combinations grows very fast with the number of stops: the page Route optimization: why and how much you save gives the orders of magnitude.

Route optimization worked example: 9 stops, one vehicle

This example is illustrative: the stops are fictitious and distances are straight-line. It shows a method and an order of magnitude, not the gain you will get on your own routes. The data below is enough to redo the calculation.

Assumptions and method

  • Square area of 20 km by 20 km, depot at the centre at (10, 10), coordinates in km.
  • Stops: 1 (12, 9), 2 (17, 13.4), 3 (3.5, 10.5), 4 (5.5, 12.5), 5 (17.3, 9.4), 6 (5, 17.6), 7 (14.9, 6.1), 8 (3.9, 10), 9 (13.7, 12.2).
  • One vehicle, return to the depot, 30 km/h, 5 minutes per stop, departure at 8:00.
  • All 362,880 possible orders are calculated and compared: the result is the exact optimum, not an estimate.
  • Order “by eye”: you always go to the nearest unvisited stop (nearest-neighbour method).
  • Differences are calculated before rounding the distances.

Stop order, without constraints

OrderDistance
Entry order (1 to 9)93.7 km
Nearest neighbour (1-9-2-5-7-4-3-8-6)49.1 km
Shortest order (1-7-5-2-9-6-4-3-8)42.9 km

The entry order is only a starting point: any arbitrary order is poor on this area, and it does not represent a good planner. The useful comparison is with the nearest neighbour: the shortest order saves 6.2 km out of 49.1, or 12.6%. For a single vehicle with no constraints, the reordering option of Google’s APIs aims at a similar result (it mainly optimizes travel time). The gap lies elsewhere.

With three time windows

Let us add three time windows: stop 2 between 8:00 and 9:00, stop 8 between 8:30 and 9:30, stop 5 between 10:00 and 11:00. The vehicle waits if it arrives before opening time.

OrderDistanceWindows missedBack at the depot
Entry order (1 to 9)93.7 km1 of 312:17
Shortest order (1-7-5-2-9-6-4-3-8)42.9 km2 of 311:40
Shortest that meets all 3 windows (9-2-6-4-3-8-7-5-1)52.4 km0 of 310:30

The shortest order on the map misses two windows and gets back at 11:40, waiting time included. The best order that meets all three windows is 52.4 km long, which is 9.6 km (22%) more, and gets back at 10:30. The figures depend on the assumptions; they show that minimum distance and meeting time windows are two distinct objectives. A route planning engine weighs one against the other, across dozens of stops and several vehicles.

When Google Maps is enough, and when it no longer is

  • Up to 9 stops, one vehicle, no imposed schedule: the app is enough, with an order chosen by hand.
  • From 10 to 25 stops, one vehicle, no imposed schedule: the app is no longer enough. The reordering option of the APIs makes it possible, at the price of development work and usage-based billing.
  • Delivery time windows, vehicle capacity, several drivers, driving time and breaks: neither the app’s help nor the documentation of the ordering option describes them. You need a route planning engine, whether it is Google’s planning service or ready-to-use software such as OCTAVE-ENGINE. Load capacity is a topic of its own: see Bin packing and truck loading: the problem costing millions.

FAQ

How many stops can you add in Google Maps?

On desktop, Google’s help indicates up to 9 stops, final destination included. This section of the help only covers desktop; it says nothing about Android or iPhone. The APIs go up to 25 intermediate waypoints for route calculation.

Does Google Maps optimize the order of stops?

In the app, Google’s help only describes reordering by hand, by drag and drop. In the Routes API, the optimizeWaypointOrder option reorders intermediate waypoints to reduce the cost of the route.

Does Google Maps handle delivery time windows?

The app’s help and the documentation of the reordering option do not describe time windows. The planning service that Google reserves for developers supports them.

Can you split stops between several drivers with Google Maps?

The app and the reordering option handle one route for one vehicle. Splitting 50 stops between 5 drivers calls for a multi-vehicle route planning engine.

Sources

Google pages read on 18 October 2026: Google Maps help, Routes API, route calculation, optimizing the order of stops, Routes Preferred, Directions (Legacy), planning service: overview and quotas and billing.


Still planning your routes with Google Maps? Contact the OCTAVE-ENGINE team to put a figure on the gap for your own routes.