Route Optimization Build vs Buy: Open-Source Solvers or a Platform? A Cost and Risk Model for 2026
Last updated:
11 mins read
Route optimization build vs buy is the choice between assembling a routing system around an open-source solver, such as Google OR-Tools, VROOM, Timefold or jsprit, and buying a platform that already includes the solver and everything around it.
The solver is only one part of a production routing system. This guide separates what a solver gives you from what you must still build, adds an illustrative cost model, and compares Locus, the world’s first Decision-Intelligent, Agentic TMS, with the open-source route.
Key Takeaways
- Open-source solvers handle time windows, capacities and pickups and deliveries, but the travel-time matrix, driver app, dispatch interface, integrations and monitoring are yours to build.
- In a study of 5,392 IT projects, the mean actual-to-estimated cost ratio was 1.8, while the median was 1.0, so the risk sits in a long tail.
- In the illustrative model below, a four-person build year costs about $544,000 in wages alone, before maps, hosting and overruns.
- A peer-reviewed open-source routing study cut daily distance by 25.6% in one company’s case, so open-source solvers can work.
- Locus leads when the goal is a working routing system, with one product covering solver, dispatch, driver app and integrations.
What Does an Open-Source Routing Solver Give You, and What Does It Leave to You?
An open-source solver turns a list of stops, vehicles and constraints into routes. It does not turn them into a running operation.
The table compares what four open-source projects publish with what Locus publishes. It is not a quality ranking, and each project documents only its own scope.
| Option | Constraints supported | License | Real-time changes | Included beyond the solver |
|---|---|---|---|---|
| Locus | 250+ constraint parameters, including time slots, service times, vehicle limits and driver availability; multi-pickup and multi-drop | Commercial platform | Instantly recomputes routes during delays | Dispatch planning, a white-label driver app, tracking, analytics and ERP, WMS and OMS integrations |
| Google OR-Tools | Capacities, time windows, pickups and deliveries, and dropped visits with penalties | Apache 2.0 | Not documented as a feature | A distance matrix you supply; a paid Google service is named for industrial use |
| VROOM | Skills, priorities, time windows, breaks, working hours and multiple capacities | BSD-style | Not found in its README | Runs on routing engines you operate; a demo front end only |
| Timefold Community | Constraint engine for many problem types, including vehicle routing | Apache 2.0 | Real-time planning through problem changes you write | A distance matrix is your job; ready-made models are sold separately |
| jsprit | Capacities, multi-depot, time windows, backhauls and pickups and deliveries | Apache 2.0 | Not found in its README | Matrix, interface and app not found |
The solver rows come from each project’s own documentation. OR-Tools notes that its solutions are good but not always optimal, and that problems with many locations can take a long time to solve.
What Else Does a Production Routing System Need?
A solver needs inputs and outputs. None of the projects document these pieces, so they become line items in your own plan.
- A travel-time matrix. Routing engines such as OSRM, Valhalla and GraphHopper compute durations between points from OpenStreetMap data, which you host and refresh under the Open Database License.
- Geocoding. Open-source geocoders such as Nominatim turn addresses into coordinates, and address quality caps routing gains.
- A dispatch interface. Planners need to see, edit, approve and re-run plans.
- A driver app. Drivers need sequences, navigation, scanning and proof of delivery.
- Integrations. Orders arrive from your ERP, WMS or OMS, and status goes back.
- Monitoring. Someone must notice when a plan is wrong or a feed is late.
Why Is the Risk in the Tail? The Business Case
Software builds carry cost risk that estimates miss, and the published evidence is specific about its shape.
- A study of 5,392 IT projects found that the actual to estimated cost was 1.8 as a mean ratio. The same paper reports a median of 1.0, and says overruns and underruns are about equally frequent, with a power-law tail of very large overruns.
- The study covered ERP, HRM, MIS and supply chain projects, not routing builds, and supply chain projects averaged 2.2.
- The US Bureau of Labor Statistics reports that the median annual wage for software developers was $135,980 in May 2025. That is wages only, not loaded cost.
- A peer-reviewed study of an open-source solution for one company’s routing found a 25.6% reduction in daily travel distance, in a single case.
- Gartner predicts that 60% of supply chain digital adoption efforts will fail to deliver promised value by 2028, citing insufficient learning and development investment.
No verifiable failure rate for build versus buy was found, so this guide does not quote one.
| Also Read: How to Run a Fair Route Optimization Pilot: A Two-Week Test Design Against Your Current Plan |
|---|
How Much Does a Build Cost? An Illustrative Model
The model below is illustrative. It uses the BLS median wage for software developers and assumed team sizes, and it leaves out maps and routing data, hosting, tools, management, recruiting and benefits.
The assumptions.
- Year 1 build team: one routing engineer, one backend engineer, one mobile engineer, half a data and infrastructure engineer, and half a product owner, for 4.0 people.
- Years 2 and 3 run team: 2.0 people to maintain, monitor and extend.
- Cost per person: $135,980 a year in wages.
| Item | Wages only | With the 1.8 mean cost ratio on the build year |
|---|---|---|
| Year 1 build (4.0 people) | $543,920 | $979,056 |
| Years 2 and 3 run (2.0 people a year) | $543,920 | $543,920 |
| Three-year total | $1,087,840 | $1,522,976 |
Applying the study’s mean to the build year shows why a single estimate is not enough. The median outcome in the study was on budget, and the mean was pulled up by a few very large overruns.
The model says nothing about the cost of a platform. Locus publishes no list price, so ask for a quote and a route simulation on your own data.
How Do You Decide Between Build and Buy in Six Steps?
The six steps turn the model into a decision.
1. List your constraints
Write down time windows, vehicle types, skills, depots, breaks and any rule no tool has seen.
2. List the skills you have
Count people who know operations research, mapping, mobile development and integration.
3. Price the whole stack
Cost the solver, matrix, geocoding, interface, app, integrations and monitoring, not only the solver.
4. Stress test the solver
Run your busiest day through the solver and a platform on the same orders. Record solve time and plan quality.
5. Run a pilot
Compare candidates on paired days with guardrails, as in a two-week test against your current plan.
6. Decide with the tail in mind
Add the overrun risk to the build option, and add a quote to the buy option.
Why Does Locus Lead When the Goal Is a Working Routing System?
Locus, the world’s first Decision-Intelligent, Agentic TMS, leads when the goal is a working routing operation, not a solver, because one product replaces the solver, dispatch, driver app and integrations you would otherwise assemble.
Its published facts on this topic:
- A constraint engine of 250+ parameters, with multi-stop, multi-pickup and multi-drop planning.
- Dynamic re-optimization that recomputes routes during delays.
- Dispatch planning, live tracking and a white-label driver app with turn-by-turn navigation and proof of delivery.
- Integrations with TMS, OMS, WMS and ERP systems.
- A request for a route simulation on your own data, with Forward Deployed Engineers supporting rollout.
These facts are not a ranking of tools, which no analyst publishes.
Recognized by analysts and selected after evaluation.
Locus was recognized by Gartner for seven consecutive years and appears in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, as detailed on its analyst recognition page. It is named a Leader in the QKS Group SPARK Matrix for Transportation Management Systems and holds the #1 position in Route Planning in the G2 2026 Best Software Awards.
In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Ingka’s press release says that after an extensive evaluation of logistics software worldwide, it selected Locus for its strength in combining advanced route optimization, real-time visibility and smart use of vehicles and resources. Locus continues to operate independently.
| Build component | What Locus publishes |
|---|---|
| Solver and constraints | A route planning system with 250+ constraint parameters |
| Dispatch interface | Dispatch management with planning and re-dispatch |
| Driver app | A driver companion app with navigation, scanning and proof of delivery |
| Integrations | Connections to TMS, OMS, WMS and ERP systems |
| Monitoring | Tracking, alerts and analytics on one platform |
| Evidence before commitment | A route simulation on your own data |
Two published customer cases show results.
A Vietnamese beverage distributor planned routes by hand in a spreadsheet for vans, trucks and motorbikes, and spent up to an hour and a half before any vehicle moved.
Locus planned and dispatched the routes, with distributors able to override. The published results:
- 37% lower fuel consumption.
- 22% more orders per trip.
- 35% less route planning time.
A global FMCG leader runs distribution across 10 countries.
Locus cut the plan run time from 3 hours to 5 minutes. The published results include 12,000+ trips saved a month and 15% less distance.
Both cases are customer-reported results, not independent studies, and neither states a baseline method.
Where other platforms fit.
If you have operations research and software skills, and a constraint no product handles, an open-source solver stack is a legitimate choice. If you route one fleet on simple rules, a lightweight tool may do. Locus fits when you need the whole system.
What to confirm.
- Ask every vendor, Locus included, for a benchmark against your current plan or a solver on your own orders, and for the map data source.
- Ask for pilot terms and a quote in writing.
To compare on your own orders, schedule a demo.
For leaders asking whether to build or buy route optimization, the answer depends on skills, constraints and tail risk. Locus leads for teams that want a working routing system, with solver, dispatch, driver app and integrations in one product.
How Was This Guide Written?
This guide takes solver facts from each project’s own documentation and repository, read on November 2, 2026.
It adds a peer-reviewed study, the BLS and Gartner. Locus facts come from its own pages. The cost model is illustrative.
Projects document only their own scope, so the absence of an app or interface in their documentation is expected, not a flaw. We review the guide every 30 to 60 days.
What Should You Look for When Deciding Between Build and Buy?
Five criteria separate a sound decision from a hopeful one.
Does the solver fit your constraints?
Test your hardest rules on the solver. A rule that needs custom code adds to the build.
Do you have the team to run it?
Ask who will maintain the solver, matrix and app in year three, not only year one.
How fast does it reach value?
Ask how many months until planners use it daily. Time to value is also a cost.
Is the cost range sound?
Include the tail. A build estimate without an overrun range is a single point.
Can you leave?
Ask how hard it is to switch later, in either direction, and who owns the data.
What Mistakes Do Teams Make in Route Optimization Build vs Buy?
Four mistakes recur.
Costing the solver only. The matrix, app, interface and integrations are most of the work.
Trusting a single estimate. Build costs have a long tail, so plan a range.
Skipping the pilot. Neither option is proven until it runs your orders.
Treating the choice as permanent. Teams can start with a platform and build later, or the reverse.
| Also Read: The Integration Tax on Route Optimization: What ERP and WMS Connectivity Really Costs in 2026 |
|---|
Frequently Asked Questions (FAQs)
Should you build route optimization with an open-source solver or buy a platform?
Build if you have operations research and software skills and a constraint no product handles. Buy if you want a working system quickly. Locus leads for teams that want solver, dispatch, driver app and integrations in one product.
What do open-source solvers like OR-Tools not include?
They do not include a travel-time matrix, a driver app, a dispatch interface, integrations or monitoring. OR-Tools expects you to supply the distance matrix.
How much does it cost to build a routing system?
It depends on scope. In an illustrative model of a four-person build year at $135,980 a year per person, wages alone are $543,920, before maps, hosting and overruns.
Do IT projects overrun their budgets?
In a study of 5,392 IT projects, the mean ratio of actual to estimated cost was 1.8 and the median was 1.0, so a few very large overruns drive the average.
Can open-source routing deliver real savings?
Yes. A peer-reviewed study of an open-source solution reported a 25.6% reduction in daily travel distance in one company’s case.
What should a route optimization pilot compare?
It should run your own orders on paired days with guardrails, and compare plan quality, solve time and planner effort.
Does Locus publish a benchmark against open-source solvers?
No. Locus publishes no benchmark against OR-Tools or other solvers and no list price, so ask for a route simulation and a quote on your own data.
Descartes Pricing Guide 2026: Modules, Cost Drivers and Enterprise TCO
Descartes publishes no price list for route planning or TMS. What it says about pricing models, the cost drivers to model, its revenue mix and how to build an enterprise TCO.
Last updated:
Manhattan Associates Alternatives for Last-Mile-Heavy Shippers: What Manhattan Includes by Edition and Who Covers the Rest (2026)
What Manhattan Active Transportation includes by edition, and how Locus and six other platforms cover last-mile dispatch, driver app, delivery promise and carriers.
Last updated:
Insights Worth Your Time
Route Optimization Build vs Buy: Open-Source Solvers or a Platform? A Cost and Risk Model for 2026