Salesforce at $1.2K/Seat/Month: The 90-Day Sunset Plan for Mid-Market

Mid-market companies running Salesforce hit the same wall every renewal cycle. The bill, fully loaded — licences plus required add-ons plus the integration scaffolding plus the in-house Salesforce admin team plus the contractor backfill — is now 2–4% of revenue. The CFO has noticed. The CRO defends it because re-platforming sounds catastrophic.

It doesn’t have to be. We’ve run three Salesforce sunset engagements in the past 18 months. Two replaced Sales Cloud with custom-built CRM-light systems. One consolidated to HubSpot. All three completed inside 14 weeks with no missed sales week.

This is the 90-day plan.

Day 0 — the loaded-cost audit

Most CFOs reviewing Salesforce spend look at the licence line. That’s the visible 50–60% of the bill. The other 40–50% is invisible:

• Required add-ons. Sales Cloud Enterprise sticker is $165/seat/month. The configurations that actually run a sales team — Sales Engagement, Maps, CPQ, sometimes Pardot or Marketing Cloud — land at $1,000–$1,400/seat/month fully loaded.
• Sandboxes and orgs. Full-copy sandboxes start around $10K/year. Production-grade orgs at scale add $30K–$80K.
• Integration scaffolding. Salesforce wants to be the system of record. Wiring it to everything else (data warehouse, marketing automation, finance system, customer support, product analytics) costs an integration platform (Mulesoft, Workato, Tray) plus a team to maintain it. Realistic floor: $200K/year for a 100-seat instance.
• Admin team + contractor backfill. A 200-seat Salesforce org typically requires 1.5–2.5 full-time admins plus contract support during major changes. Loaded cost: $250K–$500K/year.
• Renewal inflation. Salesforce’s standard renewal posture has been 20–30% YoY uplift for the past three years, often justified by “expanded feature set you haven’t yet adopted.”

The honest fully-loaded number for a 200-seat mid-market Salesforce installation is $1.8M to $3.2M/year. The licence line is $400K–$600K of that. The rest is the iceberg.

Day 1–30 — what you actually need

The audit step that matters: shadow a top sales rep for a day. Document every Salesforce screen they touch and every screen they avoid.

The pattern is almost universal:

• 70% of usage is concentrated in 8–12 screens (account view, opportunity view, activity capture, basic reporting, list views).
• The rest of the platform — flows, process builder, marketing journeys, custom objects — is either built and unused, or used by the operations team to compensate for the platform’s limitations.

A custom-built CRM that delivers those 8–12 screens, with the rep’s actual workflow as the design constraint, is genuinely 12–18 weeks of engineering for a mid-market team. Not a year. Not a multi-million-dollar program.

The objection: “but our processes are too complex to rebuild.” On audit, almost never true. The processes are complex because Salesforce’s data model forced them to be. Strip the platform constraint and the underlying business process simplifies by 40–60%.

Day 30–75 — the build

The architecture that works for mid-market CRM-light replacements:

• Postgres for the data model, designed around the actual business entities (account, contact, opportunity, activity), not Salesforce’s abstractions.
• Server-rendered UI for the 8–12 high-traffic screens. Modern stack — React or HTMX, doesn’t matter — but optimised for sales-rep workflow speed.
• OpenTelemetry instrumentation from day one. When the CRM is yours, you own the observability story too.
• Integration as direct webhooks, not a Mulesoft licence. Most mid-market integration needs are 6–12 webhook endpoints, not an enterprise integration bus.
• Reporting on top of the warehouse, not bolted to the CRM. Snowflake, BigQuery, or PostgreSQL itself depending on volume.

Build cost for a 200-seat replacement: $180K–$320K fixed-price. Annualised run cost (hosting + storage): $24K–$60K. Total Year 1: $200K–$380K, vs $1.8M+ on Salesforce.

Day 60–90 — the migration

The migration mistake we see most often is treating it as a big-bang cutover. The pattern that works is parallel-run:

• Weeks 7–10: read-only mirror. New CRM ingests Salesforce’s data in near real-time. Reps can see both. No one writes to the new system yet.
• Weeks 10–13: read-write mirror. Selected pilot reps (typically 10–15% of the team) start writing to the new CRM. Sync back to Salesforce as system of record.
• Week 13–14: cutover. New CRM becomes system of record. Salesforce becomes read-only archive.
• Weeks 14–16: Salesforce contract non-renewal letter. Salesforce stays accessible until contract end for any forensic / historical queries.

This pattern is operationally boring, which is the goal. No drama. No missed quarter.

What changes for the sales team

Honestly: less than they expect. The 8–12 screens they actually use look familiar and load faster. The workflows they were already running stay the same. The data they captured pre-migration is all there.

What changes for the sales-ops team is bigger: they spend more time on actual analysis and less on the Salesforce admin work that used to consume 40–60% of their week.

What changes for the CFO is the line item. A 200-seat install at $1.8M+/year becomes a Year-1 cost of $250K–$420K (build + run) and a Year-2+ run cost of $30K–$60K/year.

What we ship

Fixed-price Salesforce sunset engagements, 12–14 weeks, code in your GitHub from day one, parallel-run period built into the schedule. The Salesforce TCO calculator below estimates honest fully-loaded cost for your specific configuration and the payback months on rebuild.

If your renewal is in the next 90 days, the math is worth running this week.

Read more: /upstream/salesforce-sales-cloud-alternative · /calculators/salesforce-tco · /case-studies/