ITSM evaluation criteria help procurement teams compare vendors using evidence instead of polished sales demonstrations. Twelve weighted measures cover process depth, ITIL alignment, usability, configuration, asset relationships, integrations, automation, reporting, deployment, security, ownership cost, and vendor reliability. Buyers can score real workflows under each measure, identify red flags, and use direct questions to challenge vendor claims before contracting.
The approach is ideal for organizations replacing an existing service desk or running a formal platform selection, with operational requirements already established. Evidence should include exported records, agent and employee tests, security controls, pricing schedules, and migration results.
The 12 ITSM Evaluation Criteria at a Glance
How to use this scorecard
Rate vendors from 1 to 5 against the same demonstrable requirement. A score of 1 indicates the capability is absent or demands substantial additional work; 3 indicates the core scenario functions with constraints; 5 indicates the full agreed test succeeds using the proposed subscription and deployment. Enter scores only after capturing evidence, ideally a recording, export, contract passage, or completed test.
Assign a weight to all twelve criteria, then multiply the score by its weight. Add the twelve weighted scores and divide by 100 to obtain a weighted result between 1 and 5. The weights carry greater decision value than individual scores: a vendor receiving 5 for a capability with little relevance to the operating model has delivered little purchasing benefit.
Mid-market and enterprise teams can begin with different weights.
Remember to finalize the weighting profile before inviting demonstrations.
12 criteria and suggested weights
| Criterion | What it measures | Mid-market weight | Enterprise weight | Reason for weighting |
|---|---|---|---|---|
| 1. Process scope | Required incident, request, change, and service workflows operating end to end | 13% | 12% | Daily service coverage determines adoption |
| 2. ITIL depth | Independently verified practices and tested process handling | 6% | 6% | Governance value depends on practice maturity |
| 3. Usability | Agent task completion and employee portal completion | 11% | 7% | Lean desks have fewer resources for user friction |
| 4. Configuration | Administrative changes possible through the interface without code | 11% | 8% | Development overhead weighs heavily on lean teams |
| 5. Asset and CMDB | Tickets linked directly to trusted assets and configuration items | 8% | 9% | Service impact depends on reliable asset context |
| 6. Integrations | Current systems covered by licensed connectors, APIs, and quotas | 9% | 12% | Enterprise environments have larger integration dependencies |
| 7. Automation and AI | Verified results on owned tickets with oversight | 6% | 7% | Automation earns value through measured service outcomes |
| 8. Reporting and SLAs | Live management reports and auditable service-level measures | 8% | 7% | Management requires actionable performance evidence |
| 9. Deployment and residency | Hosting, data location, recovery, and operating requirements | 6% | 8% | Regional and regulated deployment choices carry exposure |
| 10. Security and audit | Access, retention, records, exports, and response controls | 7% | 12% | Large organizations carry wider governance obligations |
| 11. Three-year TCO | Subscriptions, setup, migration, upkeep, and administrative effort | 10% | 4% | Cost predictability is a stronger constraint for lean desks |
| 12. Vendor and migration | References, delivery capacity, contract terms, and exit planning | 5% | 8% | Enterprise replacement programs have greater continuity risk |
| Total | Weighted profile | 100% | 100% | Weights are starting points |
The mid-market profile favors usable workflows, administrator independence, and three-year cost. The enterprise profile assigns extra emphasis to integration depth, security evidence, data controls, and vendor continuity. A twelve-person team with simple procurement needs may increase the weights for implementation effort and cost. A regulated service organization should assign greater weight to deployment, residency, and audit records.

Use the scorecard
Use the ITSM evaluation scorecard to compare three vendors, record evidence, and calculate weighted totals from scores of 1 to 5. Choose the mid-market or enterprise weighting profile before the demonstrations begin. Apply the same tests to all vendors, and approve any weight changes before scores are compared.
Before You Evaluate: Define What You Are Buying

Assess your current ITSM maturity
An ITSM maturity assessment should describe current delivery rather than reward a vendor’s feature list. Begin with the processes that actually run: incident intake, service requests, problem reviews, changes, access approvals, knowledge publishing, and configuration records. Identify their owners and the evidence they generate. Then test how consistently people follow the process.
A short internal audit can uncover the starting point:
- How many service processes have written procedures and named owners?
- How many routine activities have workflow-based approvals or routing?
- How many requests still arrive through email, messaging, or shared spreadsheets?
- Which recurring incidents consume the highest agent hours?
- Which service records can be exported with a reliable ownership and timestamp history?
Document the maturity baseline as a set of findings, such as “seven request categories, three documented approval routes, two manual asset lookups.” That creates measurable requirements. A procurement team seeking to extend service handling into HR or facilities should define those operational scenarios separately from its core IT processes.
Separate must-have from nice-to-have
Write the requirements before the first product demonstration and freeze the approved version. When a vendor showcases a polished feature, reviewers are tempted to add it to the mandatory list. That changes the basis of comparison and can penalize another provider whose strengths address the original operational problems.
A must-have deserves a business reason, owner, measurable acceptance condition, and priority. Procurement teams can transfer the ITSM tool selection criteria into ITSM RFP requirements before inviting formal bids. “Integrates with identity” is vague. “Imports employees from the current directory, preserves department ownership, and deactivates access after a terminated account is disabled” gives two evaluators a common test. Desirable functions can be scored separately, with a lower decision weight.
Compare the requirements with essential ITSM features to identify gaps, while keeping unused capabilities outside mandatory scoring.
Decide who owns the decision
Name the decision roles before vendor contact. The service desk lead should own process tests; IT operations should test monitoring and configuration links; security should validate access, logging, and data controls; finance should own the three-year commercial model; and an employee representative should test the portal. A procurement owner manages contractual and supplier checks.
Publish who scores, who recommends, who approves, and which roles have a veto for control failures. A security objection to missing audit data requires a different treatment from an agent’s preference for a screen layout. Agree on the threshold for elimination, the evidence standard, and the process for resolving disputed scores.
A selection committee lacking this arrangement can reach a high score and still stall at contract approval.
The 12 ITSM Evaluation Criteria in Detail
1. Process scope against your actual workflows

What to score
Calculate the share of mandatory workflows that the vendor can execute end to end using the proposed plan. Test incident, request, problem, change, service catalog, and event intake where they appear in the approved requirement list. Record completion, manual interventions, and unavailable stages.
What good looks like
A score of 5 means the agreed scenarios run through intake, classification, approval or escalation, fulfillment, and closure with assigned ownership and reliable records. The vendor demonstrates the workflow using realistic ticket conditions and configured permissions.
Red flags
- A blanket response that all processes are “configurable” without a working demonstration
- Standard service requests require paid add-ons absent from the commercial offer
- Event alerts, change approvals, or catalog entries depend on manual handoffs
Questions to ask the vendor
- Which of our mandatory workflows can you demonstrate end to end using the quoted plan?
- Which workflow stages require extra products, services, or code?
- Can you export one completed incident and its linked approvals, timestamps, and status history?
2. ITIL alignment and practice depth
What to score
Check current independent ITSM tool accreditation and the exact practices or processes covered. PeopleCert’s ITIL practices provide the process context, while the PeopleCert tool directory and PinkVERIFY list help verify external recognition. Test relevant practice performance separately from certification status.
What good looks like
A top-scoring provider identifies the certified version, independently verified practice names, and current status. Its incident, change, and request demonstrations produce records consistent with the buyer’s operational and audit requirements.
Red flags
- “ITIL-aligned” appears on a slide without an independent verification reference
- A certificate refers to an older product version or different product family
- Certified practices differ from the buyer’s core implementation scope
Questions to ask the vendor
- Which current product version holds PinkVERIFY or PeopleCert accreditation, and where can we verify it?
- Which practices have external verification, and which rely on vendor claims?
- Can you demonstrate a change or problem record using our agreed approval and evidence rules?
3. Usability and end-user adoption
What to score
Measure task completion separately for the agent console and employee portal. Give agents a triage task and employees a common service request; record completion rate, elapsed time, errors, and assistance required.
What good looks like
A score of 5 requires first-time users to complete agreed activities with minimal guidance. Navigation, search, status messages, and mobile responsiveness meet agreed access requirements, while an administrator can configure common forms through the interface.
Red flags
- The demonstration shows the agent console but skips the employee request portal
- Employees require agent assistance to find a routine catalog request
- Accessibility or mobile issues appear in ordinary completion tests
Questions to ask the vendor
- Can five employees unfamiliar with the product submit our top three requests unaided?
- How many clicks and minutes does a new agent need to classify and route a common incident?
- Can we test the portal on our managed browsers, phones, and accessibility settings?
4. Configuration versus customization
What to score
Count the approved workflow changes an authorized administrator can complete through graphical configuration, without writing code. Measure time to update categories, SLAs, forms, routing, permissions, and approval rules; classify remaining work as scripting or external development.
What good looks like
A score of 5 means administrators can complete the selected changes through documented controls, with rollback, permissions, and an upgrade-safe change history. The vendor identifies any custom code and its ongoing maintenance owner.
Red flags
- Routine approval changes require external developers or proprietary scripts
- Customizations lack a documented testing and upgrade procedure
- The implementation quote excludes future maintenance of bespoke workflows
Questions to ask the vendor
- Can our administrator build a new three-stage approval route without code while we observe?
- Which requirements need scripts, plug-ins, or professional services?
- How are configured and coded changes tested after a major product release?
5. Asset and CMDB linkage

What to score
Select ten incidents involving identifiable devices or services. Count how many reveal the correct asset or configuration item from the ticket, including ownership, service relationships, and recent changes. Verify if the CMDB and asset records share a native data model or rely on external synchronization.
What good looks like
A score of 5 means agents can trace the ticket to a trusted asset and relevant relationships without a separate manual lookup. Update ownership or asset status, then verify that the connected incident context remains accurate.
Red flags
- Incident records need an unstructured device name and manual asset search
- Asset records and service configuration records disagree after routine updates
- Linking incidents to assets requires another separately licensed product
Questions to ask the vendor
- Can you trace an incident to its asset, owner, related service, and latest approved change?
- What percentage of asset relationships update automatically from our designated systems of record?
- Which data fields are native, synchronized, or maintained manually?
6. Integration coverage and API limits
What to score
Divide the number of required systems served by maintained native connectors by the total number in the buyer’s inventory. Record separately the systems served through APIs or custom middleware. Assess ITSM integrations against the buyer’s actual applications, then test authentication, throughput, failure handling, and the entitlement included in the proposed license.
What good looks like
A score of 5 means mandatory connections work within documented API quotas and security controls, with agreed ownership for failures. The vendor discloses connector fees, data limits, and compatibility requirements up front.
Red flags
- “Integration available” refers to a paid marketplace product absent from the quote
- API throttling blocks realistic alert or ticket volumes
- The vendor lacks a documented response to connector failure or expired credentials
Questions to ask the vendor
- Which of our current systems have maintained native connectors under the quoted subscription?
- What API rate limits, data-volume charges, and authentication restrictions apply?
- Can you demonstrate a failed transfer, retry handling, and an operator alert?
7. Automation and AI capability

What to score
Use an anonymized sample of real incidents and requests. Measure classification accuracy, accepted suggested actions, successful routing, time saved, human override, and erroneous automations. Test access boundaries and the audit history of automated decisions.
What good looks like
A score of 5 requires an agreed improvement on representative tickets with controls for exceptions and approval. Automation follows business rules, exposes its actions to reviewers, and can be disabled without disrupting core ticket handling.
Red flags
- AI demonstrations rely entirely on vendor-created examples
- Automated decisions lack a traceable history or an accessible human override
- AI pricing uses usage units or credit caps absent from the cost model
Questions to ask the vendor
- What percentage of our test tickets are classified and routed correctly without manual correction?
- How can an administrator approve, reverse, or audit an AI-generated action?
- Which usage thresholds, model charges, and data-processing terms govern production AI?
Gartner’s April 2026 I&O survey found that 28% of AI use cases fully succeeded against ROI expectations, while 20% failed outright, among 782 I&O leaders surveyed in late 2025. Gartner also forecasts cancellation of over 40% of agentic AI projects by the end of 2027. Those results favor measured production scenarios over ambitious feature labels.
8. Reporting, analytics and SLA visibility

What to score
List the three reports management currently requests, with definitions for breach, resolution, backlog, and reassignment. Ask the vendor to build one report live from sample data and verify filters, calculations, permissions, and exports.
What good looks like
A score of 5 means the buyer can reproduce requested management views using configured fields and valid SLA calendars. Scheduled exports preserve the data definition, and authorized managers can trace chart totals back to source records.
Red flags
- The sales demo displays attractive sample dashboards but cannot build a new report
- SLA breaches differ between report totals and individual ticket histories
- Export or report scheduling is licensed separately from the quoted package
Questions to ask the vendor
- Can you build our monthly breach report in the demonstration, using our definition of business hours?
- Can managers filter incident trends by service, entity, priority, and ownership?
- Can an auditor export a reporting period with its source records and calculation rules?
9. Deployment model and data residency
What to score
Measure how many mandatory hosting, data-location, recovery, tenancy, and commercial requirements the proposed deployment meets. Include contractually named regions, backup locations, data export, recovery objectives, and India availability where required.
What good looks like
A score of 5 means the signed proposal identifies data and backup locations, availability commitments, recovery arrangements, and responsibilities for a SaaS, on-premises, or hybrid model. The implementation team can test the agreed recovery and access paths.
Red flags
- The contract names a vendor’s local office while leaving hosting regions undefined
- Backup residency, subprocessors, or exit exports require later negotiation
- A hybrid proposal leaves identity, security ownership, or synchronization duties ambiguous
Questions to ask the vendor
- Which production and backup regions will hold our records, and where is that commitment written?
- Which data categories can leave India through telemetry, AI processing, or third-party services?
- Can we contract and receive invoices in INR, and who handles migration and incidents in India business hours?
India hosting options should be checked per product and subscription: Freshservice, ManageEngine ServiceDesk Plus Cloud, and supported Jira Service Management plans publish India data-location choices; ServiceNow operates India data-center capacity. Regional contracting, tenant eligibility, and backup boundaries still need written confirmation.
10. Security, compliance and audit evidence
What to score
Check mandatory controls using performed tests, including role-based access, segregation of duties, audit history, retention, exports, identity integration, and incident escalation. Count the control tests passed and record any gaps requiring compensating procedures.
What good looks like
A score of 5 means authorization changes, approvals, record edits, and administrative actions can be traced with timestamps and actor identity. Evidence exports meet the buyer’s control requirements, with tested access rights and retention rules.
Red flags
- Administrative actions can alter historical records without traceable evidence
- Retention or deletion rules apply inconsistently to attachments and backups
- Audit exports omit approver identity, timestamps, or changed values
Questions to ask the vendor
- Can you export an incident and related change with a tamper-resistant sequence of actions?
- How are privileged access, approval segregation, and retention policies tested?
- Which security and regulatory artifacts are available for our audit team and procurement review?
11. Total cost of ownership over three years

What to score
Calculate the thirty-six-month cost of subscriptions, deployment, migration, integrations, training, administration, renewals, and exit work. Compare the same number of agents and required functions for all shortlisted vendors, then test headcount and exchange-rate scenarios.
What good looks like
A score of 5 means finance can trace all recurring and initial charges to a written quote, including implementation dependencies and administrator effort. Price increases, usage caps, annual commitments, and exit fees have been quantified or contractually bounded.
Red flags
- Entry license rates exclude mandatory modules and integration charges
- Custom code and routine administration have no staffing cost attached
- Renewal increases or exit charges remain undefined until contract negotiation
Questions to ask the vendor
- What is the full 36-month price for our agent growth, selected functions, and integrations?
- Which implementation tasks and administrator hours remain our responsibility?
- What charges apply to extra environments, data extraction, renewal, and contract exit?
For instance, a comparison for a 20-agent team could include US$11,280 for one year of Standard-tier subscriptions at US$47 per agent per month, plus US$22,000 in setup and migration, US$9,000 in integrations, US$3,000 in training, and US$12,000 annually in internal administration. Assuming flat subscription rates, zero new implementation fees, and the stated annual administration cost, the 36-month total is US$103,840. Contracted prices may differ from the US$103,840 illustrative total.
IDC’s IT skills research projects that skills shortages will affect over 90% of organizations worldwide by 2026, with an estimated US$5.5 trillion in associated losses; the internal staffing line deserves scrutiny.
12. Vendor viability, support and migration effort
What to score
Rate the vendor against reference availability, regional delivery capacity, contract terms, product roadmap evidence, incident-response commitments, and migration feasibility. Use a reference organization of comparable scale and a sample export from the existing platform. Validate migration cost assumptions against field mapping, data volume, and extraction limits.
What good looks like
A score of 5 means the vendor provides credible customer references, a named delivery team, workable escalation procedures, and a documented exit path. The migration rehearsal preserves agreed records and the customer retains practical access to its data.
Red flags
- No named reference customer of comparable scale in the buyer’s region
- Migration tooling, source-field mapping, or export formats remain undefined
- Contractual escalation or exit assistance depends on verbal commitments
Questions to ask the vendor
- What feedback can we obtain directly from a customer with similar size and governance needs?
- Which references and case studies show successful migration from an environment like ours?
- Which migration charges, service response terms, and escalation contacts will appear in the contract?
The Five-Minute Unified vs Integrated Test
Why the distinction matters
A common login can hide several products linked through synchronization. The operational difference emerges when an agent follows an incident from the initial ticket to the asset record, SLA history, and approved change. If those records refer to different copies of an object, teams may need reconciliation when an owner, device, or service relationship changes.
A genuinely shared data environment should preserve the relevant relationship through normal user actions. A connected suite can also serve the need, provided its integration ownership, latency, and failure handling are explicit. The purchasing question concerns actual information flow and operating costs, rather than a vendor’s category label.
Five questions that reveal the answer
Demonstrate one incident through five connected tasks:
- Can an agent open one incident and reach the associated device record without a separate login?
- Does the same incident expose its running SLA clock and applicable service-level rules?
- Can the agent follow that incident into a related change approval within the same working context?
- When an asset owner or status changes, do the linked records display the same current value without manual reconciliation?
- Can a manager report on assets with open incidents directly from the shared records, without exporting and joining separate files?
Run the demonstration on a real service scenario with a changed asset record rather than a static screen. Record any additional login, synchronization delay, manual lookup, or integration charge. A shared interface is useful; a coherent source of operational truth is the deeper purchasing test.
Scoring the Shortlist
Weighting the criteria for your context
The two suggested profiles are starting points. A regulated institution may transfer weight from usability and advanced AI to deployment, audit evidence, and continuity. A small service team may put higher weight on interface-driven configuration, subscription predictability, and the vendor’s migration workload. Keep the total at 100 and preserve one approved weighting set for all suppliers.
A vendor should also fail qualification when a mandatory legal, data, or security condition remains unmet, even if its aggregate score is high. Weighted rankings are useful only after hard purchasing constraints are handled. Maintain a record of why a criterion received its weight so the committee can defend the decision after staff changes.
Running a proof of concept that tests something
A two-week proof of concept can establish evidence without extending the selection into an open-ended trial.
| Period | Buyer activity | Evidence to retain |
|---|---|---|
| Days 1–2 | Configure access, import anonymized tickets, set acceptance tests | Source-data sample, requirement list, test plan |
| Days 3–5 | Run three frequent workflows and portal tests | Completion times, screenshots, errors, user feedback |
| Days 6–8 | Test integrations, SLA clocks, report building, and audit exports | Logs, records, report definitions, failed-case results |
| Days 9–10 | Review contract assumptions, cost model, and outstanding defects | Signed scoring sheets, commercial clarifications, risk decisions |
Set identical conditions for all participants. End users should submit requests without guidance while experienced agents run harder cases. Include a deliberately broken integration, an overdue SLA, an approval rejection, and a permission boundary. Clean demonstrations reveal potential; failure cases reveal operating burden.
Five ways scoring goes wrong
- Feature-count scoring: awarding points for catalog items that solve no approved operational requirement
- Stakeholder dominance: changing weights because the most senior or vocal reviewer favors one product
- Portal neglect: examining agent screens while employee requests remain difficult to complete
- License-only comparison: ignoring implementation, integration, migration, administration, and exit costs
- Requirements drift: rewriting must-haves after product demonstrations and undermining fair comparison
Resolve disagreements through the same evidence standard, record the decision, and rerun a test where its outcome is disputed.
ITSM Evaluation Criteria for Indian Organizations
DPDP Act 2023 and audit evidence
Service requests can contain employee names, contact details, identifiers, screenshots, and access records. India’s Digital Personal Data Protection Act, 2023 and notified DPDP Rules, 2025 establish a phased framework, with major substantive obligations scheduled for 2027. Procurement should test the controls needed for lawful processing, retention, access requests, correction, erasure, and demonstrable administrative actions ahead of those commencement dates. Ticket retention and deletion practices also need to account for any separate legal recordkeeping obligations.
CERT-In six-hour incident reporting
India’s CERT-In cyber-security directions and FAQ require qualifying cyber incidents to be reported within six hours of detection or notification. The FAQ clarifies which incident categories meet the reporting requirement and permits a first report using the information then available, followed by later updates. An ITSM evaluation should test detection-to-escalation timestamps, on-call ownership, protected evidence records, and the path from incident identification to regulatory communication.
Ask the vendor: “Can you demonstrate a reportable cyber incident detected at 02:00, escalated outside business hours, with a complete timestamp history and an export ready for the responsible reporting team?” The reporting duty remains with the organization.
Data residency, INR contracting and local service coverage
India hosting, INR billing, and regional service terms deserve a separate line within criterion 9. Freshservice, ManageEngine ServiceDesk Plus Cloud, and eligible Jira Service Management configurations publish India data-location choices. ServiceNow has Indian data-center capacity; the hosting arrangement still needs contract confirmation. Compare primary storage, backup location, AI-data processing, payment currency, local partners, and the exact hours of technical assistance. A local sales office alone gives little evidence about the proposed tenant or its incident-response path.
Applying the Checklist: How Infraon ITSM Scores
Infraon’s ITSM tool scores 4/5 for asset and CMDB linkage and 4/5 for configuration. Moreover, PeopleCert’s Infraon Infinity listing and PinkVERIFY certification provide external accreditation checks. Tata Electronics uses Infraon for managed access covering 18,000+ employees; SMET reports over 75% lower annual IT expenditure.
After You Choose
- Convert approved scoring evidence into an acceptance schedule and a signed statement of work.
- Define owners, deployment phases, test gates, and cutover decisions for ITSM implementation.
- Evaluate migration cost against an agreed data-quality threshold and reconciliation checks.
Teams still refining their shortlist can compare ITSM platforms against the same scoring model.
Frequently Asked Questions
What are the key ITSM evaluation criteria?
The twelve criteria cover process scope, ITIL depth, usability, configuration, asset and CMDB relationships, integrations, automation and AI, reporting, deployment, security, three-year cost, and vendor continuity. Weights should follow operating requirements, while evidence from real workflows establishes the vendor’s scores. Mandatory compliance conditions require a separate pass-or-fail review.
How long should an ITSM evaluation take?
A practical evaluation can run over four to eight weeks once the requirements and decision team are ready. That period can include shortlisting, demonstrations, a two-week proof of concept, security checks, and commercial review. Complex regulated procurement may take longer because internal risk, legal, and sourcing approvals have distinct schedules.
What should be in an ITSM requirements checklist?
An ITSM requirements checklist should name required service processes, ticket volumes, user roles, portal tasks, integrations, SLAs, assets, reporting, security controls, data location, and commercial constraints. Give all mandatory requirements an owner and a test condition. Include growth expectations and migration rules so vendors estimate the same delivery scope.
How do you compare ITSM vendors fairly?
Compare vendors using frozen requirements, identical real-world demonstrations, and one agreed weighting profile. Collect evidence for scores from 1 to 5, then multiply scores by weights and compare totals. Apply mandatory security and contractual gates before selecting the highest-scoring offer. Review a 36-month cost model under common assumptions.
What questions should you ask an ITSM vendor before buying?
Ask which required processes run in the proposed plan, which integrations carry extra fees, what audit evidence can be exported, and where production and backup records reside. Request a sample migration, an administrator-led configuration test, customer references of similar scale, and a written three-year price with renewal and exit provisions.
What are the most common mistakes in ITSM tool selection?
Common mistakes include prioritizing feature counts over tested workflows, changing criteria after demonstrations, ignoring employee portal adoption, trusting AI examples created by the seller, and comparing only monthly license fees. Weak ownership of security checks and migration acceptance can also delay contracts. A shared scorecard and recorded evidence reduce those problems.

