15 Questions to Ask When Choosing a Backup Solution
Nanda Teknoloji Engineering Team · · 8 min read
In brief
The right backup solution is not the one with the most features; it is the one that meets your business objectives for RTO and RPO, ransomware resilience, verifiable restore, and operational load. These 15 questions let you compare at the quote and demo stage.
When choosing a backup solution, the right starting point is not a product list but business objectives: what do you need to bring back, how fast, and with how much data loss? That answer, the RTO and RPO objectives, is the core of what you expect from a solution. Next come scope, ransomware resilience, verification, security, compliance, operational load, and contract topics. The 15 questions below offer a framework for comparing different solutions on the same scale during quote collection, demos, and pilots. Next to each question there is a “what to look for” and a “red flag” note. This article is not a product comparison; it describes the questions to ask whichever provider you choose.
A. Business objectives and scope
1. Which systems, applications, and data types are we protecting? Virtual machines, physical servers, databases, file shares, SaaS data, containers, workstations, and configurations should be listed. Red flag: Scope described vaguely, such as “most” or “standard workloads.”
2. What are our RTO and RPO objectives for each tier, and does the solution support them? A daily backup is suitable only for an RPO at the level of hours; a tighter objective requires frequent backup or continuous protection. Red flag: Choosing a solution without objectives.
3. Is application consistency ensured? For databases and multi-tier applications, the ability to take a consistent (application-aware) backup is sought. Red flag: Only disk-level copies that are unaware of the application.
B. Ransomware and data security
4. Is there an immutable or offline copy, and how is it implemented? Object lock, a hardened repository, or an air-gapped environment; the retention period must not be shortenable by administrators. For details, see our article on the 3-2-1-1-0 backup rule.
5. Is backup management separated from the production identity plane? Separate accounts, multi-factor authentication, and role-based authorization are sought. Red flag: The backup console being managed with a domain account.
6. Is data encrypted in transit and at rest, and who manages the keys? TLS for transfer, strong encryption at rest, and clear responsibility for key management are needed.
7. Is there anomaly detection and an alerting mechanism for backups? You should be able to get alerts for unexpected large changes, bulk deletion attempts, and retention period changes.
C. Restore and verification
8. What are the restore options? Single file, full virtual machine, restore to different hardware or a different location, and booting in an isolated environment. The options should cover your scenarios.
9. Is the restorability of backups verified automatically? Post-backup integrity checks and boot tests in an isolated environment are sought under this heading. A successful backup job does not mean a restorable backup.
10. How is the restore exercise carried out and who is responsible? The test frequency, records, and reporting should be clear. For the method, the disaster recovery test checklist gives a useful framework.
D. Architecture, compliance, and contract
11. Where are the copies kept, and does it meet data residency requirements? For backups containing personal data, location and access matter. Review the guidance of the KVKK Authority and internal policies; involve your legal and compliance team in the process.
12. Are retention policies flexible and auditable? Ask about different retention periods for different data types, legal hold, and the handling of deletion requests.
13. Is the growth and cost model predictable? Storage, restore (egress) traffic, and licensing models work differently across solutions. Ask how the total cost of ownership changes when data grows 2-3 times.
E. Operations and relationship
14. Who will run daily operations, and how much effort does it require? It should be clear who is responsible for monitoring backup jobs, resolving errors, patching and updating, capacity planning, and exercises. A managed service model can ease this load.
15. How do support and the takeover process work in a disaster? Support channels, priority levels, shared responsibility, and integration with disaster recovery should be probed. Get in writing which metrics and which solutions the SLA clauses in the contract cover.
Evaluation table
To compare the answers, you can use a scoring table like the one below. The weights are specific to your organization.
| Area | Question no. | Weight (example) | Solution A | Solution B | Note |
|---|---|---|---|---|---|
| Business objectives and scope | 1-3 | High | RTO/RPO fit | ||
| Ransomware resilience | 4-7 | High | Proof of immutability | ||
| Restore and verification | 8-10 | High | Exercise support | ||
| Compliance and architecture | 11-13 | Medium | Location, retention | ||
| Operations and support | 14-15 | Medium | Who does what |
Common mistakes
- Focusing on the feature list. What decides is fit to the objectives, not features.
- Only taking backups in the pilot. The pilot’s criterion is the restore of a real application.
- Confusing backup with disaster recovery. For the difference between the two, see the article on the disaster recovery center and DRaaS.
- Calculating cost only on storage. Restore traffic, licensing, and operating effort are included too.
- Not thinking about the exit strategy. The ease of moving data to another solution should be asked about from the start.
- Not reading the liability limits in the contract. Which part is the provider’s responsibility and which is yours must be in writing.
Who should you ask the questions with?
Backup is not the decision of a single team. Different parts of the questions belong to different stakeholders, and the answers should pass through their approval.
| Stakeholder | Questions they own | Why needed |
|---|---|---|
| Business unit owners | 1, 2, 12 | The definition of critical data and acceptable loss comes from them |
| Infrastructure and operations team | 3, 8, 9, 14 | Technical fit and daily load |
| Information security | 4, 5, 6, 7 | Ransomware resilience and access management |
| Legal and compliance | 11, 12 | Retention, location, and personal data obligations |
| Procurement and finance | 13, 15 | Total cost and contract terms |
Sharing this distribution in the evaluation meeting prevents the decision from turning into the preference of a single team.
Example: Comparing two quotes on the same scale
Suppose an organization receives two different quotes. The first has a low starting cost and fast setup, but the immutable copy is offered as an add-on module and the restore exercise is out of scope. The second costs somewhat more; immutability and automatic verification come as standard, and regular exercises are included in the operations service.
A team looking only at first-year cost might pick the first quote. Yet once the immutable module, the effort to be spent on exercises, and possible restore traffic are added, the gap may close; moreover, the difference in resilience in a ransomware scenario carries a value that is not measured in numbers. That is the purpose of the questions: to make hidden costs and missing capabilities visible at the quote stage.
Pilot (PoC) plan: Verifying the questions in the field
Answers can look good on presentation slides; the real information emerges in the pilot. Design the pilot with the following steps:
- Choose representative workloads. A database, a file share, a group of virtual machines, and, if possible, some SaaS data.
- Write a list of criteria. Backup window, achieved RPO, restore time, resource consumption, and ease of use of the interface.
- Run a failure scenario. Deliberately delete or corrupt a virtual machine, roll a database back to an earlier point, and simulate an unauthorized access attempt on the backup account.
- Test immutability. With administrator privileges, try to delete a copy or shorten the retention period and record the result.
- Observe operations. Are the alerts meaningful, is the daily report sufficient, how many steps does incident resolution take?
- Test support quality. Open a real support ticket during the pilot and evaluate the response time and content.
At the end of the pilot, fill in the scoring table above and go into the decision meeting together with the business owners.
Contract and exit topics
Once technical fit is established, review these headings in the contract:
- The scope of service level commitments and how they are measured
- Data ownership, data deletion, and return of data at the end of the contract
- The list of subcontractors and data locations
- Security incident notification times and shared responsibility
- Pricing variables: storage, restore traffic, licensing, support levels
- Support for data migration at the end of the contract
These items are often addressed after the technical evaluation, yet they are what you turn to most in a crisis. Involve the legal and procurement teams early.
Decision summary
- Write the objectives (RTO, RPO, scope) first; derive the questions from these objectives.
- Ask for evidence of ransomware resilience and restorability; do not settle for statements.
- Calculate total cost including storage, traffic, licensing, and operating effort.
- End the pilot with a real recovery scenario.
Moving forward with Nanda Teknoloji
On projects like this, Nanda Teknoloji carries out the discovery, design, migration, and operations steps: clarifying requirements, designing the backup architecture, and operating it with migration and regular restore tests. To get started, take a look at the backup solutions page.