Before You Buy

1. Establish the case for change

Start with the operating problem, not the feature list. Define what is slowing execution, fragmenting customer data, weakening measurement, or creating avoidable cost.

Translate each problem into a measurable outcome. “Improve marketing” is not a decision criterion; reduced campaign production time, cleaner handoffs, stronger conversion, or reliable pipeline reporting are.

• Separate requirements from preferences and label each by priority.
• Document what technology will not fix: unclear ownership, poor data, weak process, or missing governance.
• Agree on the decisions the investment must enable after implementation.

2. Assess the environment you already own

Inventory platforms across marketing, sales, service, data, analytics, digital experience, and communications. Include regional purchases, agency tools, contract dates, annual cost, utilization, owners, integrations, and planned retirement.

Before replacing a platform, test whether the current environment is underconfigured, underused, poorly governed, or insufficiently staffed. A new license can make a weak operating model faster and more expensive.

• Identify overlapping capabilities and assign a clear system owner for each.
• Classify what will be optimized, replaced, rebuilt, migrated, archived, or retired.
• Estimate replacement effort separately from first-year licensing cost.

3. Define the work the platform must perform

Build requirements from real use cases. For each one, record the business objective, user, audience, trigger, data, process, channels, decision logic, approvals, exceptions, volume, and reporting outcome.

Ask vendors to demonstrate your work—not a polished version of theirs. The test should reveal the effort required to produce, govern, troubleshoot, and report on the use case.

• Document current and projected contacts, profiles, events, sends, API calls, users, regions, and retention periods.
• Evaluate the operating model required for administration, campaign production, data management, reporting, and support.
• Define the minimum viable capability for launch and the capabilities that can follow.

4. Validate architecture, data, and integrations

A platform’s data model must reflect how the organization actually works: contacts, accounts, opportunities, customers, audiences, events, workspaces, and business units. Terms such as “partition” or “workspace” do not guarantee meaningful separation.

Define systems of record and ownership for identity, consent, customer status, revenue, engagement, and eligibility. Document data quality requirements before assuming the new platform will clean existing records.

• Map every source, identifier, format, frequency, volume, consent status, and retention rule.
• Confirm identity matching, deduplication, normalization, merge, deletion, and suppression behavior.
• Test whether “native” integrations include the required objects, direction, latency, monitoring, support, and license.
• Review APIs, rate limits, webhooks, bulk processing, exports, sandbox access, logging, and versioning.

5. Evaluate governance, reporting, and AI

Governance is part of the operating design. Review roles, approvals, templates, naming standards, shared assets, locked components, audit history, change tracking, workspace separation, archiving, and release processes.

Define the decisions reporting must support and verify that the platform can access the underlying CRM, revenue, cost, campaign, identity, and offline data. Attribution cannot repair missing tracking or conflicting metric definitions.

Treat AI as a workflow capability, not a sales label. Define its actual use case, data use, human review, auditability, access controls, retention, intellectual property terms, and usage pricing.

• Require human review for customer-facing content, eligibility, scoring, compliance language, and sensitive-data personalization.
• Confirm that data can be exported, audited, reprocessed, and retained outside the vendor environment.
• Review consent, deletion, access requests, regional requirements, encryption, SSO, MFA, audit logs, subprocessors, and incident response.

6. Price the operating reality

Compare total cost of ownership across the full contract term—not only the subscription. Model licensing drivers, overages, add-ons, implementation, data migration, integrations, staffing, agencies, support, training, governance, renewal increases, and exit costs.

Clarify what is available today versus beta, limited release, roadmap, or demonstration concept. Contract commitments should be based on available capability and documented assumptions.

• Request a complete list of add-ons for reporting, APIs, sandboxes, channels, permissions, AI, exports, support, and training.
• Estimate the internal time required for strategy, security, legal review, content, testing, and change management.
• Speak with references that resemble your organization in complexity, volume, ecosystem, and regulatory environment.

7. Buy for the operating model you can support

The right technology is not necessarily the platform with the most capabilities. It is the platform that supports the highest-value use cases, fits the ecosystem, can be operated by the available team, provides appropriate data access, and scales without unreasonable cost.

• Name the implementation team and clarify responsibilities across internal teams, vendor, partner, agency, and systems integrator.
• Validate assumptions, dependencies, data preparation, integrations, testing, training, security review, and launch criteria.
• Define how success will be measured after purchase—not just whether the software launched.

Category-specific diligence

• Marketing automation: CRM integration, lifecycle processing, scoring, deliverability, database pricing, and campaign architecture.
• CRM: data model, account relationships, permissions, ownership, workflow, reporting, and cross-functional administration.
• Engagement and journey orchestration: real-time events, re-entry, frequency controls, suppression, channel readiness, and profile or event pricing.
• Data and activation: ingestion, schemas, identity resolution, quality, audience refresh, destinations, governance, residency, and data egress.
• Analytics and attribution: tracking quality, metric definitions, campaign taxonomy, offline activity, cost data, model transparency, and export.
• AI and workflow technologies: defined use cases, model governance, human review, logging, access control, intellectual property, risk, and usage pricing.