Ten-year cost models
The cheapest GIS in year one may be the most expensive over a decade. These planning ranges are not supplier quotes. They are transparent 2026 NZ-dollar models designed to expose the costs that are commonly left out of cloud-versus-local discussions.
Indicative ranges
| Scenario | Assumptions | Indicative direct cost over 10 years |
|---|---|---|
| Individual QGIS project | existing laptop, GeoPackage, two encrypted backup devices, modest online backup, occasional replacement/training | NZ$2,000–15,000, excluding the person's GIS labour |
| Small hapū, managed open cloud | 5–10 editors, 0.5–2 TB, managed PostGIS/web service, MFA/backups, around 100–200 specialist hours/year | NZ$180,000–550,000 |
| Small hapū, commercial GIS SaaS | 5–10 skilled users plus viewers, field/web apps, storage/usage, 80–200 support hours/year | NZ$180,000–650,000 |
| Small hapū, on-premises open stack | server/NAS/UPS, off-site backup, replacement cycle, around 200–350 admin/security hours/year | NZ$300,000–800,000 |
| Small hapū, hybrid | restricted local store plus hosted collaborative/public services | NZ$250,000–700,000 |
| Iwi enterprise environment | multiple business units, enterprise identity, production GIS/database administration, security, support and field/web systems | rough planning order NZ$1 million–5 million+ |
These are models, not market prices
The ranges assume specialist external-equivalent labour of roughly NZ$120–180 per hour where required. That is a modelling assumption, not a statement that every supplier charges this rate or that the work must be outsourced.
Local infrastructure models assume a hardware replacement cycle of roughly five to six years. Cloud examples assume moderate rather than imagery-intensive compute and bandwidth. Large LiDAR, imagery, 3D, digital-twin or real-time sensor environments can cost materially more.
Illustrative midpoint comparison
A comparable small-system model used the following representative values:
| Pattern | Illustrative ten-year value |
|---|---|
| Individual | NZ$8,000 |
| Managed cloud open-source stack | NZ$310,000 |
| Commercial SaaS | NZ$315,000 |
| Hybrid | NZ$430,000 |
| On-premises open-source stack | NZ$490,000 |
The purpose is not to declare managed cloud the winner. Different sensitivity and organisational requirements can justify higher-cost patterns.
Why on-premises can cost more
A local server transfers responsibilities to the organisation:
- operating system and GIS patching
- hardware monitoring
- backup
- off-site recovery
- certificates
- identity
- network security
- hardware replacement
- administrator coverage
If those tasks are omitted, the system may look cheaper because necessary work is simply not being done.
Why SaaS can cost more than expected
Commercial SaaS can add costs through:
- user types or seats
- usage credits
- storage
- premium capabilities
- support
- implementation
- data migration
- application rebuild
- continuing administration
But SaaS can also remove substantial infrastructure labour. Count both sides.
Why the individual project is different
A person doing one mapping project does not need the administration overhead of a shared enterprise system. The main requirements are good data organisation, provenance, independent backup, appropriate sensitivity decisions and a handover path.
Build your own ten-year estimate
For each year estimate:
- platform and licences
- cloud/storage/hosting
- administrator hours
- GIS support hours
- governance and review time
- security and assurance
- training and succession
- backup and recovery
- hardware replacement
- likely migration or exit work
Then add a scenario for growth and one for funding reduction.
Do not use these figures as procurement quotes
Technology prices and service terms change. Before a business case or contract, obtain current supplier pricing and state the date. These models are intended to prevent the common mistake of comparing a monthly cloud bill with only the purchase price of local hardware.
Related pages
Last verified: 16 August 2026