6 ArcGIS-to-QGIS migration cases: real savings and limitations
A comparison of six public-sector GIS migrations: which savings were reported, which were projected, what was reinvested, and which figures remain unknown.
How much does an organization actually save by migrating from ArcGIS to QGIS, PostGIS or an open geospatial architecture? Public evidence does not provide one universal answer. Across six documented cases, we found recurring license savings, first-year savings, a projection based on future users, recurring investment in open-source development and a Colombian migration whose savings were never quantified.
Those categories matter. A result reported after a transition is not equivalent to a forecast. A managed platform does not have the same cost structure as an internally operated QGIS/PostGIS stack. And spending on development must not be presented as savings.
This article compares cases from Cornare, St. Joseph County, Barrow, Surrey Heath, Malaysia and Solothurn under a simple rule: preserve the source’s currency and scope, and state what each document does not allow us to conclude.
Independence note: these projects were published by the organizations cited below. They are not GeoSAT clients or projects delivered by GeoSAT. We analyze them as public evidence for better migration decisions in Colombia.
Summary: six figures with different meanings
| Organization | Documented change | Evidence type | Published value | Correct interpretation |
|---|---|---|---|---|
| St. Joseph County | ArcGIS → QGIS as standard GIS | Reported annual savings | USD 7,500–10,000/year | License savings, not complete TCO |
| Barrow Borough Council | GeoServer + GeoNetwork publishing | Reported annual savings | GBP 26,000/year | Publishing-layer software and licenses, not desktop migration |
| Surrey Heath | On-premise/cloud systems → XMAP Cloud GIS | Reported first-year savings | GBP 13,000 | Managed annual subscription, not a free self-hosted stack |
| Malaysia federal department | Planned QGIS + GRASS adoption | Calculated potential savings | RM 3.2 million | Projection: 320 installations × assumed RM 10,000 |
| Cornare | ArcGIS → QGIS and databases → PostgreSQL | Reported migration | No quantified savings | The source states an expectation, not a measured financial result |
| Canton of Solothurn | ArcInfo/ArcView3 → QGIS, GRASS and PostGIS | Reported annual investment | ≈ $30,000/year | QGIS reinvestment, not savings or complete TCO |
The table rules out a tempting but invalid conclusion: converting and averaging all six figures to produce a “typical savings percentage.” These cases involve different years, currencies, products, architectures and scopes. Their value is showing how an economic case can be built and validated, not promising a universal percentage.
1. St. Joseph County: reported annual license savings
The St. Joseph County GIS Department’s 2021 annual summary says QGIS became the county’s standard GIS software and that the switch saved USD 7,500–10,000 annually in licensing fees compared with its previous ArcGIS software.
This is the most direct of the six cases: the organization identifies the replaced product, its new standard and a recurring annual range. Its financial scope is nevertheless limited to licensing. The document does not provide seat count, retired Esri products, training cost, project conversion, QGIS support or any specialist licenses that might have remained.
The same report listed continued transition to QGIS as primary GIS software among its upcoming work. That does not invalidate the reported savings; it shows that an institutional decision can produce results before every workflow has completed migration.
Read the full St. Joseph County analysis.
What a Colombian organization can replicate: compare its actual renewal with the seats it can verifiably retire, without multiplying another county’s result by its own user count.
2. Barrow: publishing and catalogue savings, not QGIS Desktop
The Barrow Borough Council case published by the Local Government Association documents an open solution for publishing INSPIRE-compliant metadata and services. GeoNetwork managed the catalogue and GeoServer delivered map services. The source reported GBP 26,000 in annual software and licensing savings.
Architectural precision matters: Barrow was not describing a council-wide replacement of ArcGIS Pro or ArcMap with QGIS. It replaced proprietary publishing and discovery components. The server was already available, the GIS team implemented the software and IT supported the infrastructure.
The lesson is therefore not “GeoServer costs nothing.” Its license has no fee, but administration, security, backups, monitoring, upgrades and specialist expertise remain necessary. The source itself warns that effective service and metadata setup requires expertise.
Read the full Barrow analysis.
What a Colombian organization can replicate: divide the platform into layers. OGC publishing, catalogue, desktop, database and field operations can be evaluated in separate phases, with savings and risk measured by component.
3. Surrey Heath: savings through a managed service
Surrey Heath consolidated on-premise and cloud systems into XMAP Cloud GIS, according to the Local Government Association case. The council reported GBP 13,000 in software-license savings in the first year and a migration in which more than 80% was completed in under two weeks, with the remainder during the next two.
XMAP was not a free, self-managed QGIS/PostGIS installation. It was a cloud product delivered through an annual subscription and built with open-source components. Its value proposition was precisely to transfer some operational and maintenance responsibility to a provider.
The source does not disclose the subscription price, detailed prior cost or a later multi-year cost series. GBP 13,000 is therefore a first-year outcome under that arrangement, not a guaranteed annual amount. Its four-week schedule should not be applied to an organization with ArcPy, complex geodatabases, field applications or mission-critical integrations.
Read the full Surrey Heath analysis.
What a Colombian organization can replicate: compare three routes, not two—continue with Esri, operate open technologies internally or buy a managed platform based on open standards.
4. Malaysia: a transparent projection, not audited savings
The QGIS case about Malaysia’s Federal Department of Town and Country Planning describes gradual QGIS and GRASS adoption. Twelve people received initial training before project offices requested broader instruction.
The case’s most cited figure—RM 3.2 million—comes from an explicit calculation:
320 planned installations × RM 10,000 assumed proprietary-license cost = RM 3,200,000.
Both the year-end 2012 user count and unit price were assumptions when the text was published. The figure is useful as an opportunity estimate, but it does not prove avoided procurement or audited net savings. It also excludes training, support, change management, development and infrastructure.
PostgreSQL and PostGIS appeared as an architecture under study for a future integrated network. Describing all 320 planned users as a completed PostGIS implementation would go beyond the source.
Read the full Malaysia case analysis.
What a Colombian organization can replicate: the formula’s transparency. It should replace both assumptions with its active-user inventory and actual renewal value by profile.
5. Cornare: Colombian evidence without quantified savings
Cornare’s 2024 Management Report documents database migration to PostgreSQL followed by a move from ArcGIS to QGIS within the evolution of its Regional Environmental Information System. The report connects this change to expected lower licensing costs and wider access to GIS tools.
The source does not publish a before-and-after measurement. It does not disclose the value of retired ArcGIS renewals, how many users migrated, which licenses remained or the annual operating cost of the new platform.
The report also mentions COP 291,252,500 in support from HyG Consultores within the SIAR migration. This is not the price of QGIS or a standalone ArcGIS conversion fee. The contract included support, maintenance, system evolution, geoportal integration and changes to other products. Using that amount as a generic migration quote would remove the scope that explains it.
Read the full Cornare analysis.
What a Colombian organization can replicate: institutional use of QGIS and PostgreSQL in an environmental authority. What it must still measure: retired cost, new operating cost and validated functional equivalence.
6. Solothurn: reinvesting in open source still costs money
The QGIS case about the Canton of Solothurn describes a staged migration: web services, data in PostGIS and, finally, replacement of ArcInfo and ArcView3 with QGIS and GRASS.
The source says more than 50 people used QGIS daily while the broader SO!GIS infrastructure served more than 3,500 canton employees. Mixing those populations would inflate a license comparison: consuming a web map is not equivalent to needing an advanced editing and analysis workstation.
Solothurn reported investing approximately $30,000 per year in QGIS since 2007 to develop features and improve the platform. This amount is not savings, a license fee or complete TCO. It supports an important strategy: reinvest part of the budget in plugins, automation and commercial support controlled by the organization.
Read the full Solothurn analysis.
What a Colombian organization can replicate: segment users and fund functional gaps. Sustainable open source does not mean operating without a budget.
What the six cases demonstrate together
1. License cost is not total cost
QGIS, PostGIS, GeoServer and GeoNetwork have no software-license fees. An institutional platform still needs architecture, infrastructure, security, backups, monitoring, upgrades, support, training and accountable operators. Net savings appear when those costs are lower than the retired cost and the new solution satisfies critical workflows.
2. Gradual migration is the norm
St. Joseph was still continuing its transition; Barrow changed a specific layer; Solothurn moved in stages; and Malaysia used training to expand adoption. None supports uninstalling an entire proprietary ecosystem on Friday and operating exclusively on open source by Monday.
3. Retaining some licenses can be correct
A team may migrate general editing, databases or publishing while retaining ArcGIS for extensions, integrations or workflows that do not yet have an accepted replacement. A hybrid migration can capture much of the opportunity while reducing operational risk.
4. Equivalence is proven through workflows, not feature lists
Opening a layer in QGIS does not prove that it reproduces domains, subtypes, symbology, layouts, ArcPy scripts, networks, field applications or business rules. Each process needs testing with real data and acceptance criteria before its associated license is retired.
5. Financial evidence needs labels
A review committee should require every figure to be classified as:
- Reported: the organization states an observed result.
- Projected: the figure depends on future adoption or assumed prices.
- Investment: money used to build or operate the alternative.
- Unquantified: the technology changed, but no financial measurement was published.
Without those labels, a “success stories” table can turn expectations into facts.
Estimate your case without guessing open-source cost
GeoSAT’s ArcGIS-to-QGIS/PostGIS migration calculator asks for one number: current annual Esri spend. It does not ask users to invent the cost of a platform that has not yet been designed.
The result uses three visible planning scenarios for initial migration, open-source operations and retained Esri licenses. It provides a preliminary five-year range, not a quote or guaranteed savings. The scenario ratios are GeoSAT planning assumptions; they were not created by averaging these six cases because the evidence is not comparable in that way.
Turning the range into a defensible budget requires a functional inventory:
- Annual Esri invoice or proposal, including extensions, credits and support.
- Actual users by profile and frequency of use.
- Active projects, geodatabases, scripts, services and applications.
- Functions that can migrate, need redesign or require retained Esri products.
- Data, training, development, infrastructure and operating costs.
- Parallel-operation period between platforms.
- Acceptance tests and a retirement criterion for each license.
The useful calculation is not “QGIS costs zero.” It is:
verifiable avoided cost − migration − new operations − retained licenses = net savings.
Frequently asked questions
Are these GeoSAT success stories?
No. They are public cases published by the organizations cited above. GeoSAT does not claim to have designed or delivered these migrations. Our role here is analyzing the evidence, limitations and relevance to Colombian organizations.
Which case demonstrates the largest savings?
There is no valid “largest savings” comparison. Barrow reports recurring publishing savings; St. Joseph, annual desktop-license savings; Surrey Heath, first-year managed-cloud savings; and Malaysia, a projection. Their currencies, years and scopes differ.
Does moving to QGIS eliminate every annual cost?
No. It removes the QGIS license fee, but not infrastructure, support, security, training, development or operations. Surrey Heath paid for a managed subscription and Solothurn reinvested in QGIS development—two different ways to sustain open technology.
Did Cornare prove how much it saved in Colombia?
No. Its report proves the move from ArcGIS to QGIS and the migration to PostgreSQL, and it states an expectation of lower licensing costs. It does not publish realized savings or a before-and-after TCO.
Must an organization abandon all ArcGIS products?
No. A hybrid strategy can retain critical Esri capabilities and first migrate workflows with validated equivalence. The right decision depends on the inventory, cost and risk of each process.
If your organization has its Esri renewal but does not yet know what operating QGIS, PostGIS or GeoServer would cost, start with the number you do know. Calculate a preliminary five-year range, then test it against a technical inventory before making a budget decision.