IT Management

Cloud and Infrastructure Vendor Evaluation Form: A Procurement Guide

Nanda Teknoloji Engineering Team · · 8 min read

In brief

The reliable way to compare cloud and infrastructure vendors is to set and weight the criteria in advance and ask every candidate the same questions. Technical competence, security and compliance, sustainability, contract terms, and an exit plan should be scored together.

The most valuable thing to do when choosing a cloud and infrastructure vendor is to put the criteria and weights in writing before looking at price quotes, then ask every candidate the same questions and score them on the same scale. Otherwise the decision drifts toward the most impressive presentation or the offer that looks cheapest; and the continuity, security, and exit problems that surface afterward cannot be solved once the contract is signed. This article offers an evaluation form that technology procurement managers can use directly: criteria groups, a scoring table, questions to ask candidates, document verification steps, and red flags. You can weight the form for your own environment and attach it to your RFP.

Before you start evaluating: write down the need

Define your own need first, not the vendor. A comparison made without the following information stays superficial:

  • The workloads to be migrated and their criticality levels (what is the impact if the business stops?).
  • Continuity targets: acceptable downtime and data loss durations (in the logic of RTO and RPO).
  • Data classes: is there personal data, trade secrets, or regulated data?
  • Growth expectations and usage fluctuation.
  • The internal team’s capacity to operate: what will you do yourselves, and what will you hand over?
  • Integration with existing systems and dependencies.

This summary should not exceed one page and should be communicated to all candidates in the same form. For preparing the internal team and setting priorities, the inventory and risk approach in the CIO roadmap article is useful.

Evaluation criteria and example weights

The table below is a starting template. Change the weights according to your own priorities; the total should be 100. The numbers are examples, not a standard.

Criteria group Example weight What is asked?
Technical competence and architecture 20 Architecture suited to your workload, performance, scaling, networking
Continuity and disaster recovery 20 Redundancy, backup, DR approach, testing practice
Security and compliance 20 Access control, encryption, logging, data residency, KVKK
Service and support model 15 SLA, support channels, working hours, escalation process
Contract and commercial terms 10 Price transparency, change terms, sanctions
Exit and portability 10 Data retrieval, standard formats, dependency
References and sustainability 5 Experience with similar projects, corporate stability

Score each criterion from 1 to 5 and multiply by the weight. Record who gave the score and the rationale; comparing the scores of different evaluators reduces bias.

Questions to ask in each criteria group

Technical competence

  • Can you walk us through the architecture you propose for our workload, including possible single points of failure, with a diagram?
  • How quickly can we scale resources up and down? What are the limits and the cost?
  • What options are there for network connectivity, bandwidth, and latency?
  • What monitoring data can we access when there are performance problems?

Continuity and disaster recovery

  • How is backup performed, where is it stored, and is immutability applied?
  • Who runs restore tests, how often, and do you share the results?
  • How are target times written for a disaster, and how are they proven? For details, also review the scope of the disaster recovery center.

Security and compliance

  • How is administrator access controlled: multi-factor authentication, session recording, least privilege?
  • In which locations is our data kept and processed, and will you commit to this by contract?
  • What is your notification time and process for security incidents?
  • Which independent audits or certifications do you hold? What are their scope and validity date?

For the data residency and KVKK dimension, see the domestic cloud and KVKK article, and for the general set of criteria, see cloud server selection criteria. For official obligations, the KVKK Authority and the opinion of your legal counsel are authoritative.

Service and support

  • Through which channels and during which hours is support provided? Can we reach a technical specialist directly in critical incidents?
  • What is the measurement method in the SLA; what happens in case of a breach?
  • How is planned maintenance announced?

If you are considering handing over part of operations, use the responsibility-sharing table in the managed services (MSP) article as an addition to the form.

Contract and exit

  • Which items in pricing are variable? What are the extra charges (data egress, support, additional capacity)?
  • What are the contract term, renewal, and termination conditions?
  • When the contract ends, in what format, how quickly, and under what conditions do we get our data back?
  • Are subcontractors used, and for which tasks?

Document and claim verification

Every claim a vendor makes should be provable. Practical steps:

  1. Check the documents: Verify that the scope of the certificate or audit report covers the service and location you will use; check the validity date.
  2. Hold a reference call: Talk to a customer of similar size and workload. Ask how the provider behaved when problems occurred.
  3. Ask for a pilot: Run a short trial with success criteria written in advance (performance, restore time, support response).
  4. Request the draft contract early: Review it at the evaluation stage, not during negotiation.
  5. Run a security questionnaire: Get answers in writing using a standard security question set. Headings from the NIST Cybersecurity Framework can be used as a framework.

Red flags

The signs below are not enough on their own to eliminate a candidate, but they should be questioned carefully:

  • Answering questions with generic marketing language instead of concrete responses.
  • An SLA that promises only “effort,” with no measurement method or sanction.
  • Evasive answers to questions about data location or subcontractors.
  • Exit terms that are absent from the contract altogether.
  • Restore or disaster recovery tests that are never run or never shared.
  • Reluctance to provide references.
  • Pricing made up of opaque line items; total cost of ownership that cannot be calculated.

Example scenario: how does scoring work?

Suppose a mid-sized manufacturing company is going to migrate its ERP servers. The workload is critical, it contains personal data, and the internal team is small. A four-person panel is formed with representatives from procurement, IT, information security, and legal. The weights are adjusted: the continuity and security groups are raised, and the references group is lowered. Each candidate answers the same written question set; panel members score independently, and then the scores are compared in a meeting.

The real benefit of this meeting is discussing the reasons for the score differences. For example, the security representative may have given a candidate a low score for access control while the IT representative gave a high one; the discussion reveals that the answer was ambiguous and that a follow-up question is needed. This way the decision rests on recorded rationale, not on one person’s impression. In the final stage, the two highest-scoring candidates go into a pilot, and the pilot results are reflected in the score.

How is total cost of ownership calculated?

To make offers comparable, the same items must be spread over the same period. Ask for the following items as separate lines for each candidate:

  • Infrastructure usage fee (fixed and variable parts).
  • Setup, migration, and integration work.
  • Separate fees for backup, disaster recovery, and security services.
  • Support and managed service fees; unit prices for out-of-scope work.
  • Network and data egress charges.
  • Exit and migration cost at the end of the contract period.
  • The time the internal team will spend; roughly account for this as well.

Prepare the table with a three-to-five-year horizon and recalculate it with a growth scenario (for example, if capacity doubles). A comparison that looks only at the first year’s unit price is often misleading.

Common mistakes

  • Offers first, criteria later: Writing the criteria after looking at the offers makes the decision biased.
  • Looking only at the unit price: Exit, support, and add-on service costs change the total.
  • A single evaluator: Technical, security, legal, and finance representatives not scoring together.
  • Migrating a critical workload without a pilot.
  • Thinking about the contract annex after signing: SLA and data terms must be part of the main contract.
  • Ignoring vendor diversity: Excessive dependence on a single provider increases risk; if needed, plan a redundant setup for critical workloads.

Decision summary

  1. Write the need summary and the data classes.
  2. Fix the criteria groups and weights before evaluating.
  3. Ask all candidates the same questions and get written answers.
  4. Verify documents, references, and pilot results.
  5. Make the contract and exit terms part of the evaluation.
  6. Give scores with more than one person and record the rationale.
  7. Mark red flags explicitly.

Moving forward with Nanda Teknoloji

Nanda Teknoloji carries out the discovery, design, migration, and operations steps in projects like this. If you are considering Nanda Teknoloji as a candidate in your evaluation process, we recommend clarifying your needs with the questions above; you can review the cloud infrastructure, continuity, and operations scope on the cloud server and managed services pages.

Frequently asked questions

What is the most important criterion in vendor evaluation?

There is no single criterion; the weights vary by workload. For critical systems, continuity and security may weigh more, while for flexible projects, scalability and contract flexibility may weigh more.

Is asking for references and certificates enough?

It is not enough, but it is necessary. Verify the scope and validity date of the documents, and also request a reference call about a project similar to your workload.

Is a pilot (PoC) mandatory?

It is strongly recommended for critical or complex workloads. A short pilot run with predefined success criteria gives more reliable information than presentations.

Why is an exit plan important in the contract?

If the retrieval of data, its format, timing, and cost are unclear when you need to change providers, dependency builds up. Exit terms should be written in from the start.

Related solutions

Related posts

Let's plan your critical systems together

Tell us what you need; our team will work out the right architecture and migration path with you.