Retention windows ran 60 to 90 days where 30 days met the actual compliance requirement, and the per record month meter multiplies every extra day
Backup is bought reactively after an incident and renewed automatically forever. Retention is the one variable nobody revisits.
Prepared by Redress Compliance · August 19, 2026 · Salesforce Backup proposals. 35 to 50 proposals reviewed, 2024 to 2025.
Executive summary
Retention windows ran 60 to 90 days when 30 days met the actual compliance requirement. Retention drives the bill more than data volume does.
Big object and attachment scope was undefined in 4 of 5 proposals, then billed later. Undefined scope is not excluded scope, it is deferred scope.
Restore was priced reactively, at roughly twice the rate a pre paid credit would have set. The cheapest time to price a restore is before you need one.
The question is not whether to buy backup. It is whether to buy the native product or keep a third party stack with the platform as the system of record.
How does the per record month meter work?
It counts one record, stored for one month, across the whole snapshot history. A million record org with daily snapshots over ninety days bills against ninety million record months.
Snapshot count is a multiplier, not an add on
Daily snapshots over a ninety day window produce ninety record copies per record. That is why retention moves the number further than the size of the database does.
The indicative rate band
- List: around ten dollars per thousand record months.
- Volume tier: five to seven dollars in committed estates.
- Multi org discount: negotiable across a global estate.
- Restore service: priced per event for large rollbacks.
The model is set out on the backup product page, with the wider platform shape on the platform pricing page.
Which metering surprises catch buyers out?
The objects nobody named in the proposal. Standard objects behave as expected; the rest is where the later invoice comes from.
- Standard objects: counted as expected.
- Custom objects: counted, often with surprising weight on attachment fields.
- File attachments: counted separately by size.
- Big object backups: often excluded by default, so confirm scope explicitly.
Undefined scope is deferred scope
Big object and attachment scope was undefined in 4 of 5 reviewed proposals and then billed later. A line absent from the proposal is not a line absent from the agreement.
Price the restore before you need it
Restore was priced reactively at roughly twice the rate a pre paid credit would have set. Nobody negotiates well in the week they lost data.
The Salesforce renewal negotiation playbook
The seat and add on posture, the ramp clauses, and the buyer side moves across the whole Salesforce estate.
Get the brief →What 35 to 50 Salesforce Backup proposals showed
Across roughly 35 to 50 Salesforce Backup proposals Morten Andersen reviewed between 2024 and 2025, retention was the bill driver buyers controlled least. Three patterns recur.
- Retention windows ran 60 to 90 days when 30 days met the actual compliance requirement.
- Big object and attachment scope was undefined in 4 of 5 proposals, then billed later.
- Restore was priced reactively, at roughly twice the rate a pre paid credit would have set.
Retention length, not raw data volume, is the lever that moves a per record month backup bill.
- Your agreements decoded into plain English before the auditor interprets them for you
- Coverage grid: liability caps, intellectual property protections and service levels checked in one pass
- A defensible position paper generated in minutes rather than weeks
Native product or third party stack?
Different trade offs on stability, sovereignty and leverage. Neither answer is universally right, and the compliance team often has a stronger view than the architects do.
| Dimension | Native product | Third party stack | Buyer note |
|---|---|---|---|
| Integration | Native, no integration risk | API based, brittle on schema change | Native wins on stability |
| Pricing model | Per record month | Per org or per user | Per record scales with data growth |
| Recovery objective | Granular and reliable | Variable by vendor | Test before deciding |
| Data sovereignty | Platform region only | Customer controlled | Compliance teams often prefer third party |
| Renewal posture | Bundled with the platform | Independent | Independent vendors have less leverage |
The leverage point nobody prices
A backup line bundled into the platform renews with the platform. An independent vendor has less leverage over you, and that is worth something the feature comparison never shows.
The same bundling question runs through the rest of the estate, worked through for the agent platform and the external community licenses.
Watch the briefing · 4:505 Ways to Win Your Salesforce NegotiationWhat to separate, what to cap, and why the early renewal is their trade to pay for rather than yours.
What recovery posture do you actually need?
The one the business needs, not the one the platform supports. The recovery point objective is the snapshot interval, and that interval is the multiplier.
Daily is the default and hourly is expensive
Hourly snapshots are achievable and they multiply the record month count accordingly. The documentation sets out the mechanics on the backup and restore reference.
A restore you have never tested is a plan, not a capability
- Quarterly partial restore test, validating one object family.
- Annual full restore test across the entire schema.
- Post deployment validation after every major release.
- Restore evidence matched to the compliance audit cadence.
Where the common advice on backup pricing is wrong
The standard pitch is that the native product is always the safe choice because it avoids integration risk, so retention should be generous. We disagree.
Generous retention is where the money leaks
In roughly 4 of 5 reviewed proposals the retention window was set to ninety days by default when the governing framework only required thirty, and the meter quietly tripled the bill.
The buyer side move is to size retention to the actual compliance rule, define big object and attachment scope in writing, and pre pay the restore credits. Native is convenient. The commercial posture behind the wider estate sits in the Salesforce pillar.
What the proposals measured, 2024 to 2025
Two cuts of the proposal file, both settled before signature or not at all.
Where big object and attachment coverage was left unstated at signature and then appeared on a later invoice.
Against the rate a pre paid credit would have set, negotiated in the week the data was already lost.
Both are drafting problems rather than pricing problems. Both are free to fix while the proposal is still a proposal.
Your first five moves
- Read the governing compliance rule and size retention to it, because windows ran 60 to 90 days where 30 days was the actual requirement.
- Define big object and attachment scope in writing before signature, since 4 of 5 proposals left it undefined and it was billed later anyway.
- Pre pay the restore credits, which cost roughly half what a reactive restore does and remove the worst negotiating moment from the calendar.
- Set the snapshot interval from the business recovery objective, not from what the platform can do, because the interval is the meter's multiplier.
- Test a restore before the renewal, not after an incident. The Salesforce practice and the utilization calculator size the estate before the quote lands.
Frequently asked questions
How is the product metered?
Per record per month across the snapshot history. A million record org with daily snapshots over ninety days bills against ninety million record months.
What drives the bill most?
Retention length rather than raw data volume. Every extra day of snapshots multiplies the whole record count again, which is why the window is the lever.
How long is retention usually set?
Between 60 and 90 days in the reviewed proposals, where 30 days met the actual compliance requirement in the governing framework.
Which objects get counted?
Standard objects as expected, custom objects with surprising weight on attachment fields, file attachments separately by size, and big objects often excluded by default.
Why does undefined scope matter?
Because undefined is not excluded. Big object and attachment scope was unstated in 4 of 5 proposals and appeared on a later invoice regardless.
What does a restore cost?
Roughly twice as much priced reactively as it would through a pre paid credit. Nobody negotiates well in the week they lost data.
Native or a third party stack?
Native wins on integration stability and granular recovery. A third party stack wins on data sovereignty, separation of duties and renewal leverage.
Does the renewal posture differ?
Yes. A native line bundled with the platform renews with the platform, while an independent vendor holds less leverage over you at renewal.
What sets the snapshot interval?
The business recovery point objective. Daily is the default and hourly is achievable, but each interval multiplies the record month count directly.
How often should a restore be tested?
A quarterly partial test on one object family, an annual full schema test, validation after every major release, and evidence matched to the audit cadence.
