Backup is one of the few managed services where delivery cost and client price can drift apart for years without anyone noticing. Storage grows quietly, a few clients hoard retention, someone asks for a full-site restore on a Friday, and the 60% margin line on the quote is suddenly break-even. The fix is not mysterious: pick a billing unit, decide what the base price includes, and check that the money lines up with the promises in your SLA.

This article walks through that framework. Every number in it is a placeholder, chosen to make the arithmetic easy to follow, not a recommendation for your market. Swap in your own wholesale costs, labor rates and client mix before you put anything on a quote.

Three ways to bill, and what each one hides

Most MSPs land on one of three billing units. None is wrong, but each shifts a different risk between you and the client.

In practice the strongest offers are hybrids. A per-endpoint or tiered base price covers labor and tooling, and a per-GB component above an included allowance covers storage growth. That keeps invoices predictable while making sure the client who triples their data also triples the part of the bill that tracks your cost.

What belongs in the base price, and what doesn't

The base price should cover everything the client expects to happen without a conversation: daily monitoring and alert handling, a defined retention window (say 30 daily, 12 weekly, 12 monthly points), periodic restore tests, and a set number of routine restores per month, for example three file or folder restores. That restore allowance is how you demonstrate value before disaster strikes.

Charge separately for anything that consumes real labor or real storage beyond that baseline:

• Bulk and full-system restores. Rebuilding a server or recovering a whole file share is hours of skilled work, often out of hours. Price it per incident or per hour, and say so up front.

• Migrations and onboarding. Seeding a first full backup, moving a client between storage providers or rebuilding after a hardware refresh is project work, not monitoring.

• Long-term archive. Clients who need seven years of monthly points for compliance should pay a separate, lower per-GB archive rate. Keeping that data inside the standard tier means your cheapest storage subsidizes your most demanding requirement.

• Extended retention or extra restore points beyond the standard window, priced per endpoint or per GB.

The rule of thumb: if the client's behavior can change your cost, give it its own line on the price list.

SaaS seats: price Microsoft 365 and Google Workspace per user

Mailboxes, OneDrive, SharePoint, Google Drive and shared drives don't map onto endpoints or raw GB in a way clients understand, so price them the way the platform itself is priced: per user per month. A placeholder of $3–5 per protected user, with a soft storage allowance per seat (for example 50 GB), keeps the quote on the same line as the license the client already pays for.

Two points to get right. First, define a "user" as any account whose data you protect, including shared mailboxes and departed users whose data is retained, otherwise your seat count shrinks while your storage keeps growing. Second, state the retention window for SaaS data explicitly. Clients often assume the platform keeps everything forever; your price is for the retention you actually deliver, and anything longer belongs in the archive add-on above.

The margin math on storage

Storage is where a pricing model either holds or quietly collapses, so run the numbers on a real client. Suppose a client has 2 TB of source data and you buy object storage wholesale at a placeholder $0.006 per GB per month (about $6 per TB). The naive view is that 2 TB costs you $12 a month, so charging $0.10 per GB, or $200, looks like an enormous margin. That ignores retention.

With 30 daily, 12 weekly and 12 monthly restore points and no deduplication, you could be storing many multiples of the source data. Even at a conservative 3x your cost is $36 a month, not $12, and two years of monthlies pushes it higher still. Block-level deduplication changes the picture: each new restore point stores only the blocks that changed, so a typical office data set with a few percent daily change settles at perhaps 1.3–1.6x the source size across the whole retention window. Your 2 TB client then costs you $15–$20 a month.

So bill the client on protected (source) data, which they understand and can verify, while your cost tracks stored (post-dedup) data. The gap between the two is your storage margin, and it widens as dedup ratios improve. Dedup is what lets a per-GB price stay flat for the client while retention grows.

One caution: dedup ratios vary wildly by workload. Databases, media and already-compressed archives dedup poorly, so measure each client's stored-to-source ratio after the first month and adjust the allowance or rate for outliers rather than averaging them away.

Line the price up with the SLA you promise

RPO (how much data the client can afford to lose) and RTO (how long they can afford to be down) are the two promises that cost money to keep. A tighter RPO means more frequent backups and more storage churn. A tighter RTO means standby capacity, rehearsed runbooks and someone available when the call comes. If your tiers don't price these differently, every client will pick the tightest promise at the cheapest rate.

A simple three-tier structure makes the link explicit. Placeholder figures:

The exact numbers matter less than the shape: each step up should cover the extra storage, tooling and, above all, the extra labor you are committing to. If a tier's RTO assumes an engineer at 2 a.m., the price has to pay for that engineer whether or not the call ever comes.

Mistakes that erode backup margin

Underpricing restores. Treating every restore as "included" feels generous, but a full-server rebuild can consume more engineer hours than a year of the client's fee. Include routine restores, price bulk and emergency restores separately, and make the distinction visible before anyone needs it.

Offering unlimited retention. Unlimited anything is a liability on your balance sheet. Retention costs compound every month, and the clients who use it most are the ones you can least afford. Give every tier a defined window and sell longer retention as archive at its own rate.

Billing on stored data instead of protected data. It exposes your dedup ratio to the client, makes invoices jump when a workload changes, and turns every bill into a technical debate. Bill on what they can see.

Pricing the SLA you hope to deliver rather than the one you can. If you can't actually hit a 2-hour RTO for a client at 3 a.m., don't sell it. A tier you can't deliver is a refund waiting to happen.

Never re-measuring. Data grows, dedup ratios shift and wholesale prices move. Review stored-versus-protected ratios and per-client margin at least twice a year, and let the numbers move the price list.

Putting it together

A defensible backup price has four parts: a billing unit the client understands, a base price that covers monitoring, retention and routine restores, separate lines for work that scales with client behaviour, and tiers whose prices rise with the RPO and RTO you commit to. Underneath it all sits the storage math, where deduplication lets you hold per-GB pricing steady while retention grows. Start with placeholders like the ones above, replace them with your real costs, and revisit them twice a year.

If you want the operational side to be as predictable as the pricing, the Duplicati Console's MSP plans give you one place to monitor every client's backups, verify restores and track protected-versus-stored data across tenants, with per-client views that make the margin math above easy to run. See the plans at duplicati.com.