Backup pricing looks simple because storage is quoted per gigabyte. The bill is rarely that simple. Retention multiplies the copies, a restore pulls data out of storage and through a network, and a credible recovery plan may need servers waiting to take over. A quote that covers only the storage line is not a recovery budget.
This article gives you the cost pools, a formula and the inputs to collect. It does not quote vendor prices, because they differ by region, contract and architecture. The aim is to make two quotes comparable and to expose the cost of the recovery promise you are buying.
Six cost pools
- Protected data: the volume of data in scope, before compression and deduplication.
- Retention: how many versions you keep and for how long.
- Isolation: an immutable or air-gapped copy that ransomware cannot reach.
- Restore: egress, retrieval fees, temporary storage and the labour to run the restore.
- Standby infrastructure: compute, database and network kept ready for a recovery target.
- Operations: monitoring, patching, documentation and restore testing.
Most budgets cover the first two and underfund the rest. That produces a system that stores data faithfully and cannot bring the business back within an acceptable time.
The formula
Monthly cost = storage + platform + restore testing + standby + operations.
Storage is not one line. Calculate it per data class: protected volume × retained copies × the rate for that storage tier. Restore testing is annual test hours divided by twelve, multiplied by the blended labour rate, plus any provider charges for a test restore. Standby is the cost of the smallest environment that can meet the recovery time objective, which is not always the production size.
Keep the rate card outside the formula. When a provider changes a rate, you should be able to update one input rather than rebuild the model.
Inputs to collect before pricing
| Input | How to measure | Common trap |
|---|---|---|
| Protected volume per system | Backup tool report, not disk size | Counting replicas and archives as production data |
| Daily change rate | Compare two recent full backups | Assuming a low rate during a migration or campaign |
| Retention by data class | Policy approved by the data owner | Keeping everything forever because nobody decided |
| Recovery point objective | Business tolerance for lost work | Setting one target for all systems |
| Recovery time objective | Business tolerance for downtime | Promising minutes without standby infrastructure |
| Restore test frequency | Compliance and risk appetite | Testing only the easiest system |
| Egress for a full restore | Provider calculator or contract | Forgetting the network cost of a real recovery |
| Deleted-data retention | Legal and privacy requirements | Backups that outlive the deletion promise |
Tier the data before you price it
Not every system deserves the same recovery target. Put each system into one of four tiers based on business impact: revenue-critical, customer-facing, internal operations, and archive. For each tier, record the acceptable data loss, acceptable downtime, retention period and who approves the target.
Pricing a single tier for everything usually produces one of two bad outcomes. You either pay for fast recovery of archives nobody needs urgently, or you discover that a critical system shares a slow tier with a file share.
Retention multiplication
A daily backup kept for 30 days is 30 copies. Add 12 monthly copies and 7 annual copies and the retained set is 49, not 30. Deduplication and compression change the effective volume, but the provider's billing model may still count logical data. Ask which number appears on the invoice.
Also ask what happens to a deleted record. Many backup systems retain it until the last copy expires. If a customer asks for deletion, the answer must cover backups, exports and archives, not only the live database.
Budget worksheet
| Line | Basis | Owner | Confidence |
|---|---|---|---|
| Primary backup storage | Volume × copies × tier rate | Infrastructure lead | High |
| Immutable or isolated copy | Critical systems only | Security lead | Medium |
| Long-term archive | Annual and legal retention | Records owner | Medium |
| Restore testing | Tests per year × hours × labour rate | Operations lead | Medium |
| Standby environment | Right-sized recovery stack | Infrastructure lead | Low to medium |
| Monitoring and support | Tooling plus contract tier | Service owner | High |
| Contingency | 10–20% for growth and migration | Finance | Hold |
Hidden costs to ask about
- Egress and retrieval: a full restore can be the largest single line in an incident.
- Minimum retention: some tiers charge for a minimum period even if data is deleted early.
- Snapshot sprawl: temporary snapshots that are never expired.
- Support tier: fast-response support is often a separate contract.
- Test environments: restoring into production is not a test.
- Staff time: verification, documentation and incident practice are real costs.
Ten questions for a provider
- How is protected volume measured: source, logical or billed after reduction?
- What does a full restore cost in egress and retrieval, and how is it calculated?
- Which retention tiers have a minimum commitment?
- Can we keep an immutable copy under separate credentials?
- What recovery time can you actually demonstrate for our tier?
- How often can we run a test restore, and what does it cost?
- Where is the data stored, and which laws apply to it?
- What happens to backups when a record is deleted?
- How do you notify us of a failed backup, and to whom?
- What is the exit process and format for returning our data?
A 90-day sequence
Weeks 1–3: inventory systems, volumes and owners. Weeks 4–5: agree tiers, recovery targets and retention. Weeks 6–8: design the storage and isolation model, then price two options. Weeks 9–10: run a real restore test and record the time and cost. Weeks 11–12: right-size, document the runbook and set the review date.
The test is the point. A backup budget that has never produced a successful restore is a storage invoice with optimistic language attached.
Where our team fits
We design, monitor and test backup and recovery environments under Managed Technology & Cybersecurity. Scope and starting points are on the pricing page. To compare architectures before you price them, read cloud backup and recovery options compared.
Sources
- NIST SP 800-34 Rev. 1, Contingency Planning Guide for Federal Information Systems, accessed 17 November 2026.
- AWS Well-Architected Framework, Reliability Pillar, accessed 17 November 2026.
- Microsoft Azure, Backup overview, accessed 17 November 2026.
- Google Cloud, Backup and DR concepts, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.