5 Storage pricing for service providers is a balancing act. Price too high and tenants move to a hyperscaler or a competitor with a simpler offer. Price too low and every new terabyte erodes margin, especially once immutability, restores and hardware refresh are counted. The providers that get it right start from a clear cost floor per usable terabyte, choose a billable unit tenants understand, and package extras so that the services that cost more to deliver are paid for. This article is written from the provider’s side: telcos, managed service providers, regional cloud providers and internal IT groups that run storage as a service for other teams. Enterprise buyers looking at cost from the other direction should see storage cost per terabyte. Start with the cost floor Before setting a price, a provider needs an honest number for what one usable terabyte costs per month over the life of the platform. That number should include: Hardware: servers, drives, networking, amortized over the expected service life. Protection overhead: the raw capacity consumed by erasure coding or replication for each usable terabyte. Software licenses and support, whether capacity-based or node-based. Data center costs: power, cooling, rack space and cross-connects. Operations staff: monitoring, patching, capacity management and support. Free space and headroom: capacity reserved for growth, rebuilds and maintenance that cannot be sold. Refresh: the cost of replacing hardware at end of life without disrupting tenants. Divide the total by the usable terabytes actually sold, not the usable terabytes installed. A platform that is only half full costs twice as much per sold terabyte. The framework in total cost of ownership for data storage covers the components in depth, and the scality.com blog explains how object storage sizing affects usable capacity. Pick a billable unit The billable unit decides how tenants perceive the price and how margin behaves as their data changes. Logical protected capacity Tenants pay for the size of the data they protect, before any reduction. This is easy to understand and predictable for the tenant. The provider keeps the benefit of compression and deduplication, which can improve margin, but must estimate reduction ratios carefully. Stored capacity Tenants pay for what actually lands on the platform after reduction. This feels fair and is easy to measure, but it makes revenue sensitive to data reduction and retention changes. Per workload or per user Common for backup services: a fee per virtual machine, server, endpoint or SaaS user, sometimes with a capacity allowance included. Predictable for tenants, but providers must watch for workloads that consume far more capacity than average. Hybrid Many providers combine a per-workload fee with a per-terabyte charge beyond an included allowance. This captures both the management effort and the capacity cost. Tiers that map to cost Not every terabyte costs the same to deliver. Price tiers should follow cost differences: Performance tier for data that needs fast access or frequent restores. Standard tier for typical backup and archive data. Immutable tier where data is locked for a defined period and cannot be deleted early. Long-term retention tier for compliance archives that are rarely read. Geo-replicated tier for data kept in two locations. Immutability deserves explicit pricing. Locked data cannot be cleaned up when a tenant changes its mind, so it carries a minimum commitment by design. Many providers charge a premium for immutable retention or set a minimum retention period that the tenant pays for whether or not the data is still needed. immutable backup capacity explains the capacity effect. Egress, requests and restores Hyperscalers trained the market to expect fees for data leaving the platform and for API requests. Many regional providers compete by removing those charges, which simplifies the offer and makes restores predictable for tenants. If you do include them, be clear: Will a full restore after a ransomware attack trigger fees? Are API request charges likely to surprise tenants running backup software that makes many small requests? Is there a free allowance per month? No-egress pricing is a strong selling point for backup services, because the moment a tenant needs its data is the worst moment to present an unexpected bill. Our explainer on egress vs ingress covers the terms. Commitments, discounts and minimums Commitment pricing helps both sides. Tenants get a lower rate for reserving capacity or term, and the provider gets predictable revenue to plan purchases against. Typical structures include: Reserved capacity for a monthly minimum at a discounted rate, with overage at a standard rate. Term discounts for one-year or three-year agreements. Volume tiers where the rate per terabyte falls as consumption grows. Minimum charges to cover onboarding and support for very small tenants. Reseller and partner channels add another layer. If partners resell your storage, leave room for their margin in the wholesale price. A simple margin model An illustrative way to test a price: Calculate fully loaded cost per usable terabyte per month at expected utilization. Apply expected data reduction if you bill on logical capacity. Add a buffer for unsold headroom and immutable data that cannot be reclaimed. Compare against your target price for each tier and billable unit. Re-run at lower utilization (for example, year one) and higher utilization (year three). The goal is to see where the service breaks even and how sensitive margin is to utilization, data reduction and retention. A platform that grows in small increments, adding nodes only as demand arrives, keeps utilization high and protects margin. A platform bought far ahead of demand carries the cost of idle capacity for years. Metering and billing data A price is only as good as the data behind the invoice. Providers need reliable, per-tenant metering of stored capacity, ideally sampled several times a day and averaged across the month, so that a tenant who uploads a large backup on the last day is not billed as if it had been there all month. Useful metering covers: Stored capacity per tenant and per tier, averaged over the billing period. Objects and buckets, which help spot unusual growth or misconfigured jobs. Restore and egress volume, even if not billed, to understand cost drivers. Immutable capacity, separated from regular data so retention commitments are visible. Feeding this data automatically into billing systems avoids manual reconciliation and disputes. It also gives account managers early warning when a tenant is growing quickly or approaching a quota. Benchmarking against the market Providers rarely set prices in isolation. Tenants compare against public cloud object storage list prices, other regional providers and the cost of running storage themselves. Two cautions apply. First, list price comparisons often ignore egress, request and early deletion fees, so compare total cost for a realistic workload, including at least one full restore. Second, compete on what tenants value beyond price: data location, sovereignty, support in their language and predictable bills. Our article on how to avoid cloud lock-in covers the arguments many tenants are weighing. Pricing for internal storage services Large enterprises, banks and public bodies often run storage as an internal service for business units. The same principles apply, but the goal is cost recovery and fair allocation rather than profit. Showback or chargeback rates should be based on the same cost floor, published in advance and reviewed annually. Checklist: storage pricing for service providers Calculate fully loaded cost per usable terabyte, including refresh and staff. Divide by sold capacity, not installed capacity. Choose a billable unit and test it against real tenant data. Build tiers that follow real cost differences, including immutability. Decide on egress and request fees and state the restore policy clearly. Offer commitment and volume discounts that improve forecasting. Leave room for reseller margin if partners sell the service. Model margin at low, expected and high utilization. Review pricing annually as hardware costs and utilization change. Putting it together Storage pricing for service providers works when it is built from the bottom up: a realistic cost per usable terabyte, a billable unit tenants understand, tiers that follow cost and clear rules for restores and immutability. Keeping utilization high, by growing the platform in step with demand, does more for margin than any pricing trick. Publish simple tiers, avoid surprises at restore time and revisit the model as the platform scales. For the wider service design, see backup as a service for providers. Frequently asked questions How do service providers calculate cost per terabyte? They add hardware, protection overhead, software, data center, staff and refresh costs over the platform’s life, then divide by the usable terabytes actually sold. Should providers charge for egress? Many regional providers do not, especially for backup services, because unpredictable restore fees deter tenants. If you do charge, publish allowances clearly. Should immutable storage cost more? Usually. Locked data cannot be deleted early, which reduces flexibility for the provider, so many charge a premium or a minimum retention period. Is per-VM or per-terabyte pricing better for backup services? Per-VM pricing is predictable for tenants; per-terabyte pricing tracks cost more closely. Many providers combine both, with a capacity allowance per workload. How often should storage prices be reviewed? At least annually, and whenever hardware costs, utilization or the competitive landscape change significantly. Further reading Backup as a service for providers Launching an S3 storage service Storage cost per terabyte Total cost of ownership for data storage
