[object Object]

Requirements First

Document before you look at vendors. Without a written requirements baseline, every demo looks compelling and every quote looks reasonable. Capture, at minimum:

  • Team size today and projected at 24 months (seat economics dominate TCO).
  • Industry and regulatory posture (HIPAA, GDPR, SOX, FINRA, sectoral).
  • Process complexity: simple lead-to-cash vs. multi-stage opportunities, channel-partner overlays, multi-currency, multi-entity.
  • Critical integrations with named systems and required directionality (Slack, Outlook, Microsoft Teams, ERP, marketing automation, billing, support, data warehouse).
  • AI needs at three horizons: today (drafts, summaries, scoring), 12 months (agentic workflows), 36 months (autonomous operation in scoped domains).
  • Reporting and analytics depth.
  • Mobile and offline requirements.
  • Data residency.

Two-page document, signed off by sales, ops, IT, and one executive sponsor before any vendor conversation.

3–5 vendors maximum. More dilutes attention; fewer leaves you negotiating against a single bid. Plausible 2026 short-lists:

  • Solo and microbusiness: HubSpot Free/Starter, Pipedrive, Folk, Attio.
  • SMB: HubSpot Pro, Zoho One, Freshworks Customer 360, Pipedrive.
  • Mid-market: HubSpot Enterprise, Salesforce Sales/Service Cloud Enterprise, Microsoft Dynamics 365 Sales Pro, Zoho Plus.
  • Enterprise: Salesforce Sales/Service/Industries Cloud, Microsoft Dynamics 365 Enterprise, Oracle CX, SAP CX, ServiceNow Customer Workflows.
  • Vertical specialists: Veeva (life sciences), nCino (banking), Vlocity-on-Salesforce now Salesforce Industries (insurance, comms, energy), Aptean (manufacturing).

Adjust by industry, geography, and required integrations.

Pilot on Real Data

30 days minimum, real users, real records, real processes. Scripted demos mislead; hands-on usage reveals truth. Configure each candidate with the same dataset, the same five workflows, the same integrations. Have actual users (not the project team) work in each system for two weeks. Survey on usability, speed, accuracy, frustration points.

Pilot cost is real ($25K–$200K loaded for a mid-market evaluation). It’s the cheapest insurance against a wrong multi-year commitment.

Model Total Cost of Ownership

3-year TCO is the right horizon. Sticker price (per-seat license) is 30–50% of TCO for enterprise deployments. Include:

  • Licenses, including platform fees and add-on modules.
  • Implementation services (Salesforce SI partner $200/hr; Big Four consultant $400+/hr; mid-market often $150–$250/hr).
  • Integration build and maintenance.
  • Training and change management.
  • Internal admin headcount (1 admin per 30–50 users at SMB scale; 1 per 75–150 at enterprise).
  • Data migration and cleanup.
  • AI consumption (token costs, agent outcome fees).
  • Storage and API call overages.

Build a side-by-side spreadsheet. The “cheapest” per-seat option is often most expensive at the 3-year TCO line.

Decision Criteria

Weight per your priorities, but skip none:

  • Requirements fit: 50%.
  • Pilot user experience: 30%.
  • 3-year TCO: 15%.
  • Vendor stability and roadmap: 5%.

For regulated industries, add Compliance and Security Posture as a fifth criterion at 15% (re-balancing others).

Common Failure Modes

  • Choosing on demo polish instead of pilot results.
  • Skipping the integration scope (“we’ll figure it out”) — the most common cost overrun cause.
  • Treating AI as a checkbox feature instead of evaluating per-vendor execution.
  • One stakeholder dictates the decision; adoption suffers.
  • Ignoring the SI partner ecosystem — vendors with weak partners cost more to deploy regardless of license cost.

What to Do This Week

If you’re mid-evaluation, write the two-page requirements doc and circulate it. If you’ve already chosen, audit the 3-year TCO against your year-one actuals — variance is usually informative.

[object Object]
Share