20.11.2026

HubSpot Implementations

How to Run an Enterprise CRM Evaluation (HubSpot, Salesforce, Dynamics 365)

19 min read

Matthew

How to Run an Enterprise CRM Evaluation (HubSpot, Salesforce, Dynamics 365)

An enterprise CRM evaluation needs shared requirements, comparable demonstrations and a documented decision rule. Without those controls, a buying committee can struggle to distinguish a convincing presentation from evidence that a platform fits its operating model.

This eight-step CRM selection process gives you working examples of a requirements register, scorecard, demo scenario, ownership-cost model and reference-check questions.

Huble provides CRM strategy and integration services; the process below applies the same evidence standards to HubSpot, Salesforce and Microsoft Dynamics 365.

 

Quick Guide: How to Run an Enterprise CRM Evaluation in 8 Steps

  1. Agree on outcomes, scope, and decision ownership: Define business problems and assign evaluation committee roles.
  2. Turn processes into testable requirements: Convert each team's workflows into observable, scorable statements.
  3. Set scoring before demos begin: Agree mandatory gates, weighted criteria, and supporting evidence rules.
  4. Give every vendor the same demo script: Run buyer-led sessions with identical scenarios and tasks.
  5. Validate uncertain claims through a focused trial: Separate what was shown from what needs building.
  6. Build a comparable TCO model: Model all cost categories over the same horizon per finalist.
  7. Run structured reference checks: Interview comparable organisations using a standard question set.
  8. Assemble the recommendation with approval conditions: Compile scores, risks, and conditions into a decision pack.

 

New call-to-action

 

How to organise an evidence-led CRM selection process

Run your evaluation through eight documented stages, approving the evidence at each stage before progressing. For background on platform capabilities and enterprise requirements, use Huble's enterprise CRM software guide; this article focuses on how to conduct the selection.

1. Agree on outcomes, scope, and decision ownership

Start by defining the business outcomes your CRM must deliver, not the features it must have. Outcomes are measurable changes: shorter sales cycles, higher forecast accuracy, unified reporting across regions, or reduced manual data entry between systems.

Scope decisions follow. Determine which teams, regions, and business units are included in the initial deployment. A phased rollout reduces risk, but the evaluation still needs to account for the full eventual footprint so that architecture decisions made now do not create constraints later.

Assign clear decision roles. Your evaluation committee should include representatives from IT (security, integration, infrastructure), RevOps or sales operations (process, data, reporting), sales leadership (adoption, pipeline management), marketing (automation, attribution), finance (budget, TCO validation), and procurement (contract terms, vendor risk).

Each role needs a named individual, a defined vote or veto, and agreed escalation paths. Appoint one person as the internal accountable owner for the evaluation: the individual responsible for keeping the process on schedule, resolving procedural disputes, and presenting the final recommendation to the executive sponsor.

Document these decisions in a one-page evaluation charter that the executive sponsor signs. Use the charter to manage scope changes and involve missing stakeholders before finalising the criteria.

2. Turn processes into testable requirements

Requirement documents that list features ("must have workflow automation") rather than observable behaviours produce scores that do not differentiate platforms. Each platform will claim the feature exists; the question is how it behaves in your context.

A stronger requirement reads: "when a record moves to 'Proposal Sent,' the system must assign a task to the solutions engineer with a due date calculated from the expected close date." This is specific enough to verify during a demo or trial.

Work with each team to map their current workflows, identify pain points, and translate those into testable statements. A useful format for each requirement is: Given [context], when [trigger], then [expected behaviour]. This format forces specificity and gives evaluators a pass/fail criterion.

Compile requirements into a register with columns for requirement ID, business process, requirement description, priority (mandatory, important, desirable), responsible team, and evidence method (demo, trial, vendor response, reference). The right number of requirements depends on your organisation's complexity. Too few and you lack differentiation; too many and scoring becomes unmanageable.

Example requirements register structure:

ID Process Requirement Priority Evidence Method
REQ-01 Enquiry routing Auto-assign inbound enquiries to the territory owner within a defined SLA based on region and product interest Mandatory Demo
REQ-02 Pipeline reporting Produce a weighted pipeline forecast by region and product line without manual export Mandatory Demo + trial
REQ-03 Approval routing Route discount approvals above a defined threshold to the regional director with audit trail Important Trial

 

3. Set scoring before demos begin

Scoring criteria must be frozen before the first vendor demo. Adjusting weights after presentations introduces bias toward whichever platform showed well on a dimension the committee decided to value after the fact.

Structure your scorecard in three layers. First, mandatory gates: requirements the platform must pass to remain in the evaluation. A failure on any mandatory gate removes the vendor regardless of its total score. Typical gates include data residency compliance, encryption and authentication standards, API capacity above a defined threshold, and support for your primary ERP integration.

For security and supply-chain risk gates, NIST SP 800-161 Rev. 1 provides a structured framework for evaluating vendor security posture as part of the procurement process.

Second, weighted criteria: categories scored on a defined scale, with weights that sum to 100%. Assign weights before scoring. If a genuine requirement changes, document approval and rescore every finalist under the same revised rules. Common categories include usability and adoption risk, reporting and analytics, automation and workflow capability, integration architecture, vendor roadmap and AI trajectory, and partner ecosystem.

Third, evidence-based scoring: every score must link to a specific piece of evidence. A score entered without a reference to a demo observation, trial result, or vendor written response is flagged for review.

Illustrative score anchors for current functional fit:

Score Observed fit
0 Confirmed unable to meet the requirement in the proposed solution.
1 Only a small part of the requirement is met; critical gaps remain.
2 Some required behaviours work, but substantial gaps remain.
3 The core outcome works; defined non-critical gaps need resolution.
4 The requirement works with minor, accepted deviations.
5 The complete requirement and agreed exception cases are verified.

Keep untested requirements marked "unverified" rather than scoring them zero or assuming success. Record evidence confidence and delivery method separately. A verified custom solution can meet a requirement fully, while a native feature can still fail your scenario. Assess its cost and support risk in the relevant criteria.

Roadmap promises do not satisfy a current mandatory gate. Report unverified items as open risks, and require evidence or an explicitly approved decision condition before final selection. Define category sub-weights to total 100% within each category.

Illustrative weighted scoring for a single requirement:

Suppose your committee assigns "Reporting and analytics" a weight of 15%. Within that category, REQ-02 (pipeline forecast) carries a sub-weight of 40%. Vendor A scores 4 on REQ-02. The weighted contribution is: 4 (score) x 0.40 (sub-weight) x 0.15 (category weight) = 0.24 out of a maximum 0.30. This is an illustrative calculation, not a score for a real platform. Repeat for every requirement and sum to produce each vendor's total weighted score out of five. Multiply that total by 20 if you want a percentage. Repeat the calculation with alternative plausible weights to check whether a narrow lead is robust.

Example scoring matrix (illustrative weights, not research-derived):

Category Weight Vendor A (0-5) Vendor B (0-5) Vendor C (0-5)
Usability and adoption risk 25%      
Reporting and analytics 15%      
Automation and workflows 15%      
Integration architecture 15%      
Security and compliance 10%      
Vendor roadmap and AI 10%      
Partner ecosystem 10%      

 

Note: These weights are illustrative examples. Your committee should set weights based on your organisation's priorities before any demos take place. No platform scores are shown because scoring must follow your own demos and trials.

4. Give every vendor the same demo script

Vendor-led demos are designed to showcase strengths and route around weaknesses. Buyer-led demos, built around your own workflows and synthetic data, produce comparable observations across platforms.

Design a demo script that covers four areas: standard workflows (enquiry-to-opportunity, opportunity-to-close, service case resolution), exception handling (approval routing, territory reassignment, imports from non-standard sources), administrative tasks (adding a custom field, modifying an automation rule, changing a permission), and reporting (building a pipeline forecast and an attribution report using data created during the session).

Send the script to each vendor in advance with enough lead time for them to prepare. Include your stage names, synthetic record data (not real customer data), and expected outcomes. Time-box each section so vendors cannot run over on their strongest area and skip the rest.

Example demo scenario: enquiry-to-close with exceptions (synthetic data):

  1. Create an organisation record for "Northgate Engineering Ltd" with address, sector, and headcount. Create two contacts: a Head of IT and a VP of Operations.
  2. Log an inbound enquiry from the Head of IT. Show how the system assigns the enquiry to the correct owner based on region and product interest.
  3. Convert the enquiry to an opportunity. Move the opportunity through your stages (use the stage names from your requirements register). At each stage, show required fields, task creation, and notifications.
  4. Exception: Apply a discount that exceeds the rep's approval authority. Show the routing, the approver's view, and the audit trail.
  5. Admin task: Add a new required field to the opportunity record while the session is live. Show how it appears for the end user and explain any deployment steps required.
  6. Generate a weighted pipeline report filtered by region and product line. Export to a format your finance team uses.

Score each section on the same day using a pass/partial/fail scale. Record what was observed, what was stated but not shown, and what was deferred to a follow-up. Do not record assumed outcomes: if the vendor did not show it, it is "not demonstrated."

5. Validate uncertain claims through a focused trial

Demos reveal what a vendor chooses to show. Trials reveal what happens when your own team tries to do their actual work. Scope the trial to the requirements that carry the highest risk or that could not be fully verified during the demo.

Before the trial begins, classify every requirement along two separate dimensions. The first is evidence status: has this requirement been demonstrated (observed working with the agreed synthetic test data), partially demonstrated (shown with caveats), or untested (not yet observed in any session)?

The second is delivery method: how will the vendor actually deliver this capability? The categories are native (available in the platform as shipped), configuration (requires setup using platform tools but no code), custom development (requires code, scripting, or a bespoke integration), third-party (requires a marketplace application or external vendor), or roadmap (capability does not exist today; the vendor has stated intent to build it).

These two dimensions are independent. A requirement can be "demonstrated" in a demo but rely on "custom development" to deliver in production. A roadmap item is by definition not a current capability, regardless of how confidently the vendor presents it. Both dimensions feed into your TCO model and risk register.

During the trial, assign tasks to the people who will use the system daily, not to the evaluation committee. Track task completion rates, support requests to the vendor, and any workarounds the team invents.

6. Build a comparable TCO model

Compare the full ownership cost on a like-for-like basis. Check what each vendor's quote includes rather than assuming it covers implementation or ongoing operations. The actual cost includes implementation, data migration, integration development, internal staff time, ongoing administration and training, and eventual exit or migration.

To produce a fair comparison, model every finalist over the same time horizon using the same assumptions. Include the following cost categories for each vendor:

  • Licences: Per-seat and per-module costs at the required tier, including add-ons for features treated as standard in other platforms.
  • Usage-based costs: AI credits, API capacity charges where applicable, storage tiers, and any consumption-based charges that increase with adoption or data volume.
  • Implementation: Partner fees for configuration, process design, and data model architecture. Include discovery and strategy consulting if applicable.
  • Data migration: Audit, cleansing, field mapping, validation, and cutover, including parallel running costs during the transition period.
  • Integrations: Development, testing, and ongoing maintenance for each connection to ERP, billing, data warehouse, and other business-critical systems.
  • Internal effort: Staff time for project management, requirements gathering, user acceptance testing, and training delivery.
  • Ongoing support and administration: Annual cost of dedicated CRM administrators, vendor support tiers, and partner retainers for optimisation and change requests.
  • Growth scenarios: Model at least two scenarios: a baseline (current headcount and usage) and a growth case (projected seat count, storage, and feature expansion). Track how each vendor's pricing changes under the growth case.
  • Renewal terms: Annual escalation caps, lock-in periods, and contractual protections against mid-term pricing changes.
  • Environment costs: Sandbox, staging, and development environments required for testing and ongoing change management.
  • Exit costs: Data export, contract termination terms, and the estimated effort of migrating to another platform if the relationship ends.

Example TCO structure (illustrative categories, not platform-specific costs):

Cost Category Year 1 Year 2 Year 3 3-Year Total
Licences        
Usage-based costs (AI, API, storage)        
Implementation        
Data migration        
Integrations        
Internal effort        
Support and admin        
Environment costs        
Renewal assumptions (included in relevant charges)        
Exit costs (separate exit scenario only)        
Total (baseline scenario)        
Total (growth scenario)        

 

Note: This is a structural template. Populate it with vendor-specific figures and your organisation's internal cost estimates. Build one copy per vendor and scenario. Include renewal increases in the relevant recurring charges, not as an additional duplicate cost. Treat exit as a separate scenario unless exit is planned within the evaluation horizon; do not add it to a continuing-operation baseline.

7. Run structured reference checks

Vendor-supplied references are curated. To get useful intelligence, ask each vendor for references from organisations of comparable size, industry complexity, and deployment scope. Then ask each reference the same set of questions so you can compare answers across vendors.

Focus your reference questions on five areas:

  • Delivery accuracy: How closely did the actual implementation timeline and cost match the original proposal? What changed, and why?
  • Adoption: What proportion of intended users actively use the system after the initial adoption period? What drove adoption, and what hindered it?
  • Administration burden: How many dedicated administrators does the platform require? How often do you engage the vendor or a partner for configuration changes?
  • Unexpected work: What costs or effort emerged after go-live that were not anticipated during the evaluation? This includes integration maintenance, data quality issues, and feature gaps that required workarounds.
  • Renewal experience: How has the commercial relationship evolved? Were there changes in pricing, licensing terms, or feature availability at renewal?

Document each reference conversation using a standard form so the evaluation committee can review findings side by side.

8. Assemble the recommendation with approval conditions

The final step produces a decision pack that the executive sponsor and the evaluation committee can review, challenge, and formally approve.

Your decision pack should contain:

  • Evaluation summary: The process followed, the committee members, and the timeline.
  • Scorecard results: Final weighted scores for each vendor, with mandatory gate results and the evidence register.
  • TCO comparison: The complete cost model for each finalist under both baseline and growth scenarios, with assumptions clearly stated.
  • Unresolved risks: Requirements classified as "custom development," "third-party," or "roadmap" that carry delivery risk. Include the vendor's written commitment where available.
  • Commercial assumptions: Contract terms, renewal pricing protections, and any conditions negotiated during the process.
  • Readiness conditions: What must be true before implementation begins, including data quality thresholds, internal resource allocation, and executive sponsorship commitments.
  • Recommendation: The preferred vendor, the rationale tied to scored criteria, and the conditions under which the recommendation changes.

 

Present the pack to the executive sponsor with a clear request for approval. Include dissenting views from committee members where they exist. A decision pack that surfaces disagreement before the contract is signed is more useful than one that papers over it.

New call-to-action

 

Who can help you evaluate enterprise CRM platforms?

Enterprise CRM evaluations involve multiple types of external advisors. Each has structural incentives that you should understand, disclose internally, and factor into the weight you give their input.

Platform vendors (such as HubSpot, Salesforce, and Microsoft) will assign sales engineers and solution architects to support your evaluation. Their expertise on their own platform is deep, and their commercial objective is to win your contract. Use their resources for platform-specific depth. Do not rely on them for cross-platform comparison.

Advisory and analyst firms may describe themselves as platform-agnostic. Before engaging one, ask how it is funded: some earn revenue from vendor briefings, sponsorships, or referral arrangements alongside client fees.

An "independent" label does not guarantee the absence of financial relationships with the vendors you are evaluating. Ask for a written disclosure of any commercial relationships with the platforms on your shortlist.

Platform-specialist partners can contribute configuration knowledge and delivery evidence within their specialism. Ask them to identify the limits of that expertise and how they will validate requirements outside it.

A credible specialist partner will disclose the affiliation clearly and explain where their platform does and does not fit your requirements based on the evidence from your evaluation.

For every type of adviser, ask three questions before the engagement starts: How are you compensated by the platform vendor, directly or indirectly? What methodology do you use to compare platforms you do not implement? Who in your firm is the accountable owner for the quality of the recommendation?

What should your CRM evaluation decision pack contain?

A complete CRM evaluation decision pack gives the executive sponsor everything needed to approve, challenge, or reject the recommendation without repeating the evaluation. Use this checklist to confirm completeness before the final presentation:

  • Signed evaluation charter (outcomes, scope, decision roles, accountable owner)
  • Requirements register with priority classifications
  • Frozen scorecard weights, score anchor definitions, and scoring methodology
  • Completed scorecard for each finalist with evidence references
  • Mandatory gate results (pass/fail per vendor)
  • Demo observation notes and trial results
  • TCO model for each finalist (baseline and growth scenarios)
  • Reference check summaries (comparable format across vendors)
  • Unresolved risk register with vendor commitments and delivery method classifications
  • Negotiated commercial terms and renewal protections
  • Adviser compensation disclosures
  • Readiness conditions for implementation start
  • Formal recommendation with scored rationale and dissenting views

 

How Huble can support CRM evaluation planning

Huble is a HubSpot Elite Solutions Partner and HubSpot's 2024 Global Partner of the Year, certified to ISO/IEC 27001:2022 and ISO 9001:2015 across global operations. Huble's deepest delivery expertise is on HubSpot, and that affiliation is a factor your committee should weigh alongside the evaluation evidence, as with any platform-specialist partner.

Huble provides CRM strategy, integration and migration services aligned with business objectives. Its data migration guidance can help your committee identify discovery, validation and cutover questions to include in the evaluation. Confirm the specific advisory deliverables and commercial scope before commissioning support.

For organisations where HubSpot is already on the shortlist, Huble can contribute platform-specific depth to the evaluation. For organisations still at the longlist stage, the process in this guide applies regardless of which vendor you select.

If your evaluation committee wants to discuss any aspect of the process described in this guide, contact Huble's team for an initial conversation.

 

FAQs about enterprise CRM evaluation

How do you build a CRM evaluation shortlist?

Start with your mandatory gates: data residency, security certifications, integration requirements, and budget ceiling. Any platform that fails a mandatory gate is removed before demos.

From there, narrow to a manageable number of vendors based on market position, existing technology partnerships, and alignment with your operating model. Defining gates before researching vendors prevents the shortlist from being shaped by marketing rather than requirements.

What do you do when the evaluation committee disagrees on scoring?

Disagreement surfaces differences in priorities and interpretation. When two scorers rate the same criterion differently, require each to cite the specific evidence behind their score.

If the evidence supports both readings, escalate to the executive sponsor for a tie-breaking decision. Document the disagreement and the resolution in the decision pack so the rationale survives personnel changes.

How long should an enterprise CRM evaluation take?

Timeline depends on scope: the number of platforms being evaluated, the complexity of your requirements, and the availability of your committee members. A compressed timeline risks skipping trials or reference checks. An extended timeline risks stakeholder turnover and evaluation fatigue.

Include the time needed for security review, comparable trials, reference availability and procurement approvals. Consider CRM migration planning alongside selection, while keeping irreversible platform-specific work subject to the final decision.

How should you evaluate an implementation partner alongside the platform?

Treat the partner decision with the same rigour as the platform decision. Score candidates on strategy capability, evidence at your scale, data migration methodology, change management as a defined workstream, and post-go-live support.

A credible partner should be willing to show you redacted examples of discovery deliverables and references that match your headcount and regions. For a detailed framework, see Huble's CRM strategy consulting partner evaluation guide.

When should you bring in an implementation partner during the evaluation?

Consider engaging potential partners during the trial phase (step 5). Partners with platform experience can help your team distinguish between native capability, configuration effort, and custom development during the trial.

This input affects your TCO model and risk register. Starting the partner conversation after the platform decision is made can mean the implementation timeline starts later than necessary.

Transform your front office with HubSpot & Huble.

Considering HubSpot? Discover how we help large businesses implement and optimize HubSpot across marketing, sales, and service to streamline operations, unify data, and drive innovation.

Latest Insights

HubSpot Implementations

19 min read

How to Run an Enterprise CRM Evaluation (HubSpot, Salesforce, Dynamics 365)

Run an enterprise CRM evaluation with testable requirements, shared demo scripts, evidence-led scoring, ownership costs and structured reference checks.

Read more

HubSpot Implementations

8 min read

M&A Tech Due Diligence: Assessing CRM and GTM Stacks Before You Acquire

Standard technology due diligence often skips the CRM and go-to-market stack. Here is what to assess before you acquire and how it should shape the deal.

Read more

7 min read

What Actually Breaks in Year 1 of HubSpot (and How to Prevent It)

Most HubSpot problems after go-live come from gaps in ownership, data and change control. See what breaks in year one and how enterprise teams prevent it.

Read more

AI Solutions

13 min read

HubSpot for the Agentic Era: AI Agents in Your RevOps Stack

Learn where AI agents fit in your HubSpot RevOps stack, which tasks to delegate, and how to retain human accountability across revenue operations.

Read more

9 min read

How to Build a Defensible Revenue Forecast in HubSpot

Build a defensible revenue forecast in HubSpot with seven steps covering shared definitions, buyer evidence, management judgement and accuracy tracking.

Read more

HubSpot Implementations

12 min read

How to Make Your Enterprise CRM AI-Ready: Where to Start and Who to Trust

Learn where enterprise CRM AI-readiness work should start, who needs to own it internally, and what evidence to require from every delivery partner.

Read more

HubSpot Implementations

13 min read

Sales & Marketing Alignment: An Enterprise RevOps Operating Model

Learn how enterprise teams build a RevOps operating model with shared decision rights, lifecycle stages, revenue scorecards, and a 90-day pilot framework.

Read more

HubSpot Implementations

14 min read

HubSpot and Salesforce sync problems: causes and fixes

Diagnose six causes of HubSpot and Salesforce sync problems, learn corrective steps grounded in platform documentation, and know when to redesign.

Read more

HubSpot Implementations

16 min read

Building a Single Customer View After a Merger

Learn how to build a governed single customer view after a merger with a strategic framework for identity resolution, consent, and phased CRM integration.

Read more

AI Solutions

10 min read

What bad CRM data costs you once AI is involved

Understand how poor CRM data affects AI output, forecasts and reporting. Build a cleanup business case that separates staff capacity, cash and risk.

Read more