How We Cut Our B2B SaaS Stack by 35% Without Losing a Single Feature

# How We Cut Our B2B SaaS Stack by 35% Without Losing a Single Feature

The Short Version

Over six months, our team at Spark Werks reduced our active SaaS tools from 47 to 31 --- a 35% reduction --- while maintaining *full* feature coverage across sales, marketing, product, finance, and operations. We didn't just "rip and replace." We mapped every tool to core workflows, surfaced hidden redundancies (including *three* overlapping analytics dashboards and *four* separate contract repositories), and ran head-to-head consolidation tests with real user data. The result? $217,000 in annual savings, 12 fewer vendor logins per employee, and zero workflow degradation --- verified by internal NPS surveys and usage telemetry.

The Problem: How We Ended Up with 47 SaaS Tools

Let me be honest: we weren't reckless --- but we *were* reactive. As Spark Werks scaled from $8M to $24M ARR over three years, our stack grew like kudzu. A new sales rep joined → added Gong + Chorus + Salesloft. Marketing launched ABM → brought in 6sense + Demandbase + ZoomInfo + Clearbit. Engineering adopted a new CI/CD pipeline → added CircleCI, Sentry, Datadog, and New Relic --- *plus* legacy Splunk for compliance logging.

By Q1 2025, our finance team flagged something alarming: we had 47 paid SaaS subscriptions, but only 29 were actively used >3x/week. Eight tools hadn't been logged into in >90 days. Twelve were licensed for 50+ seats but used by <5 people. And --- here's the kicker --- our procurement team discovered *seven* tools billed under different legal entities (subsidiaries, LLCs, old DBAs), meaning we were missing bulk discounts and unified support SLAs.

We weren't unusual. In fact, during our work evaluating stacks for mid-market clients last year, we saw similar patterns: average of 38--44 tools per $10--$50M ARR company, with ~22% overlap in core capabilities (CRM, reporting, document signing, identity management).

The Audit: What We Found When We Actually Looked

We paused all new tool evaluations for 30 days and ran a cross-functional audit:

- Inventory: Cataloged every tool (name, vendor, contract end date, monthly cost, # of licensed seats, # of active users, primary use case)

- Usage: Pulled API logs, SSO login frequency, and feature adoption metrics (via tools like BetterCloud and Torii)

- Ownership: Assigned each tool to a single "Steward" --- not a department, but one named person accountable for ROI

The findings stung:

✅ 11 tools served *identical* functions (e.g., HubSpot, Marketo, and Pardot all used for email campaign orchestration --- with no shared audience or reporting)

⚠️ 9 tools had <15% seat utilization (e.g., a $42K/year project management tool licensed for 120 users, used daily by 14)

❌ 3 tools duplicated *critical* security controls (we had Okta + Azure AD + JumpCloud managing the same SSO flows)

🔍 1 tool was entirely redundant: a $18K/year CPQ platform that sat unused because sales reps defaulted to Excel + Salesforce CPQ instead

The 3-Step Framework We Used

Step 1: Map Every Tool to a Primary Workflow

We refused vague labels like "marketing" or "sales enablement." Instead, we defined *exact* workflows:

- "Lead-to-opportunity handoff" → required CRM update, lead score sync, Slack alert, and calendar invite generation

- "Monthly revenue reconciliation" → needed billing data (Stripe), usage data (Pendo), and financials (NetSuite) in one view

- "Contract renewal tracking" → required e-signature (DocuSign), CLM logic (Juro), and renewal date alerts (Chorus)

Each tool got *one* primary workflow assignment. If it couldn't be tied to a specific, measurable step in an owned process --- it went on the shortlist for sunsetting.

Step 2: Identify Redundancies (Hint: There Were Lots)

We grouped tools by *output*, not category:

- Reporting & Analytics: Looker (used), Power BI (used), Tableau (unused), Metabase (used by 3 engineers) → consolidated into Looker + embedded Power BI for finance-only views

- Contract Lifecycle: DocuSign (signing), Juro (CLM), PandaDoc (proposal drafting), Adobe Sign (legacy HR contracts) → migrated *all* to Juro (with DocuSign embedded for external signers)

- Sales Engagement: Salesloft (active), Apollo (prospecting only), Wingman (call coaching), Gong (call recording) → kept Salesloft + Gong; retired Apollo (replaced by built-in Salesloft prospecting) and Wingman (Gong's AI insights covered 92% of use cases)

Step 3: Test Consolidation Candidates Ruthlessly

No vendor pitch decks. No "free trial optimism." We ran 2-week parallel tests:

- For analytics: Built identical dashboards in Looker and Power BI using *live production data*. Measured load time, query flexibility, and stakeholder approval rate (Looker won: 87% preferred vs. 13%)

- For CPQ: Ran side-by-side quotes for 47 real deals --- comparing accuracy, speed, and configurator UX. Salesforce CPQ matched our legacy CPQ tool on 100% of scenarios

- For identity: Tested Okta vs. Azure AD on MFA enforcement, JIT provisioning, and SSO failover --- Okta delivered 37% faster session recovery

What We Cut, What We Kept, and What We Regret

Cut (16 tools):

- Apollo.io, Wingman, Tableau, Metabase, PandaDoc, Adobe Sign, Clearbit, 6sense, Demandbase, Jira Align, Runway, SmartDraw, Lucidchart (replaced by Miro), Trello, ClickUp (replaced by Asana), Zendesk Guide

Kept (31 tools):

- Salesforce (CRM + CPQ + Revenue Cloud), HubSpot (marketing ops only), Gong (calls), Looker (analytics), Asana (PM), Okta (IDM), Stripe (billing), NetSuite (finance), Pendo (product analytics), Juro (CLM), Miro (collab), Linear (eng), etc.

Regret (1 tool we wish we'd cut earlier):

ZoomInfo. We kept it for sales intelligence --- but after migrating to Lusha + LinkedIn Sales Navigator + native Salesforce Data.com enrichment, we realized 73% of our ZoomInfo usage was for *duplicate* firmographic lookups already handled elsewhere. We sunsetted it in Month 7 --- saving $84K/year. Lesson: if you're paying for *data*, verify it's *unique* data.

The Results: 35% Fewer Tools, Same Feature Coverage

- ✅ 35% reduction: 47 → 31 tools

- ✅ $217,000 annual savings (net of new consolidated tool costs)

- ✅ Zero feature gaps: All 122 core workflows retained full functionality (validated via workflow mapping + user testing)

- ✅ 12 fewer daily logins per employee (avg. drop from 18 → 6)

- ✅ Support tickets down 28% (fewer integrations = fewer failure points)

- ✅ Internal NPS for tool usability up +19 pts (from 32 → 51)

Most importantly: when we surveyed stakeholders *six months post-consolidation*, 94% said their job was *easier* --- not harder --- despite fewer tools. That's the real win.

Key Takeaways for Your Team

1. Start with workflows --- not vendors. If a tool doesn't map to a documented, owned, measurable step in your go-to-market or delivery process, question its existence.

2. **Measure *active* usage --- not licenses**. Seat utilization <30% is a red flag. If it's not used weekly by ≥3 people, it's overhead.

3. **Test with *real* data and *real* tasks** --- not vendor demos. If your team can't complete their top 3 workflows in <90 seconds with the new tool, it's not ready.

4. Assign stewardship, not ownership. One person --- not a committee --- owns ROI, renewal timing, and sunsetting decisions.

5. **Consolidation isn't about "less." It's about *clarity***. Fewer tools means faster onboarding, tighter security, better data lineage, and more time spent selling --- not syncing.

We're still optimizing. Next up: unifying our observability stack (Datadog + Sentry) and rationalizing our dev tools. But for now? We sleep easier knowing every tool on our invoice earns its keep --- and our team spends less time logging in and more time shipping value.

--- Eva Quinn, Strategy Lead at Spark Werks

*We help mid-market B2B SaaS companies optimize their tech stacks --- not just cut costs, but sharpen execution. Learn how we run stack audits: sparkwerks.co/stack-audit*