| name | cost-optimization |
| description | Identify and execute cost optimization opportunities — vendor renegotiation, tool consolidation, process automation, efficiency improvements, and structural cost reduction for startup efficiency. |
| version | 2.0.0 |
| author | Crewm8 |
| maintainer | Gokul (github.com/gokulb20) |
| license | MIT |
| homepage | https://crewm8.ai |
| tags | ["cfo","finance","cost-optimization","vendor-negotiation","efficiency","cost-reduction","automation"] |
| related_skills | ["budget-creation-management","profitability-analysis","business-case-modeling"] |
| inputs_required | ["current-spend-data-by-vendor-and-category","saas-subscription-list","vendor-contracts-and-renewal-dates","cloud-infrastructure-spend","contractor-roster"] |
| deliverables | ["spend-diagnostic-by-category","quick-win-list-with-savings-estimates","vendor-negotiation-queue-ranked","structural-improvement-roadmap","savings-tracker-projected-vs-actual"] |
| compatible_agents | ["hermes","claude-code","droid","cursor","windsurf","openclaw","openai","generic"] |
Cost Optimization
Run systematic cost optimization — not knee-jerk cutting, but disciplined, ongoing improvement of the company's cost structure. Renegotiate vendors, consolidate tools, automate processes, and find structural efficiencies. Goal: every dollar of spend is intentional and efficient.
Purpose
Cost optimization in startups is a constant tension: spend enough to grow, but not so much that you run out of runway. Most startups either ignore costs until a crisis (leading to panic cuts that damage the business) or slash indiscriminately (cutting growth investment along with waste). This skill provides a structured, phased approach to cost optimization — quick wins first, then structural improvements, with strategic restructuring only when necessary. The goal is to find the waste without cutting the growth.
When to Use
- "Optimize our costs"
- "Find savings in our OpEx"
- "Renegotiate vendor contracts"
- "Are we overspending anywhere?"
- "Cost reduction sprint"
- "Run a zero-based budget review"
- "Find 10% to cut without hurting growth"
Inputs Required
- Current spend data by vendor and category (from AP and subscription management)
- SaaS subscription list with users, features, and cost
- Vendor contracts and renewal dates
- Cloud infrastructure spend (AWS, GCP, Azure)
- Contractor roster with rates and hours
- Headcount data (salary, benefits, tools per person)
Quick Reference
| Phase | Timeline | Savings Potential | Risk |
|---|
| Quick Wins | 1-4 weeks | 5-15% of OpEx | Low |
| Structural Improvements | 1-3 months | 15-30% of targeted areas | Medium |
| Strategic Restructuring | Crisis only | 20-30%+ of total burn | Severe |
| Cost Category | % of Total Spend | Typical Savings |
|---|
| Headcount (salaries + benefits) | 65-85% | 10-30% (crisis), 0-5% (ongoing) |
| SaaS / Software | 5-15% | 15-30% |
| Cloud Infrastructure | 3-10% | 20-40% |
| Professional Services | 2-8% | 20-50% |
| Marketing / Ads | 2-10% | 20-50% with smarter spend |
| Office / Rent | 2-8% | 20-40% |
| Travel & Entertainment | 1-3% | 50-80% |
Procedure
Phase 1: Diagnostic (Understand the Baseline)
Start with clean data from spend, subscriptions, and budget sources.
Break down every cost into types and assess optimization difficulty.
Phase 2: Quick Wins (1-4 weeks)
These can be done immediately with minimal business disruption:
| Action | Estimated Savings | Time to Implement | Risk |
|---|
| Cancel unused SaaS subscriptions (< 1 active user) | $500-5,000/mo | 1-2 weeks | Low (verify no dependencies) |
| Downgrade over-provisioned SaaS plans (seats, tiers) | $200-2,000/mo | 1-2 weeks | Low |
| Switch SaaS subs to annual billing (save 15-20%) | $1,000-10,000/yr | 2-4 weeks | Low (cash flow timing) |
| Renegotiate top 5 vendors (ask for 10-20% off) | 10-20% of vendor spend | 2-4 weeks | Low-Medium |
| Cut unused or low-ROI marketing channels | $1,000-10,000/mo | 1-2 weeks | Low-Medium (verify attribution) |
| Audit and right-size cloud resources (idle instances, old snapshots) | 20-40% of cloud spend | 2-4 weeks | Medium (engineering time) |
| Cut travel budget, move to virtual events | 50-80% of travel spend | Immediate | Low |
| Audit contractor roster (still needed? full-time cheaper?) | 10-30% of contractor spend | 2-4 weeks | Medium |
Phase 3: Structural Improvements (1-3 months)
Deeper changes that improve efficiency long-term:
| Action | What It Means | Savings | Effort |
|---|
| Tool consolidation | 2-3 tools that do similar things → pick 1 | 20-50% of those tool costs | Medium (migration) |
| Negotiate multi-year contracts | 2-3 year lock with 20-30% discount | 20-30% of vendor spend | Medium (commitment) |
| Insource vs outsource | Build internal capability vs rely on agencies/contractors | Case-by-case | High |
| Process automation | Automate manual finance/ops workflows | Time savings (1-2 FTE) | Medium-High |
| Vendor RFP / competitive bidding | Bid out your top 3 vendor categories | 10-25% of those categories | Medium |
| Office rationalization | Sublease unused space, renegotiate lease | 20-50% of office costs | Medium |
Phase 4: Strategic Restructuring (Only in Crisis)
These hurt but save large amounts:
| Action | Savings | Impact |
|---|
| Hiring freeze | ~5-10% of annual burn (headcount stays flat) | Slows growth |
| Layoffs (10-20% of team) | 10-25% of total burn | Severe — culture, morale, momentum |
| Salary cuts | 10-20% of payroll for cuts applied | High flight risk for top performers |
| Kill a product / feature area | 10-30% of associated team cost | Strategic — refocuses the company |
| Pivot business model | Hard to quantify upfront | Existential — but sometimes necessary |
Vendor Negotiation Playbook
The Framework
1. Know your leverage:
- How much do you spend? (total, growth rate)
- How hard would it be to switch?
- Are you a logo they want? (brand value)
- Is it renewal time? (max leverage is 30-60 days before renewal)
2. Always ask for:
- Better pricing (10-20% off is standard ask)
- Better terms (net-30 → net-45, annual billing discount)
- Added value if they won't move on price (more seats, premium features, training)
3. The conversation:
"Hi [vendor], we're doing our quarterly vendor review. We value the
product, but we're under pressure to optimize our cost structure.
Can we discuss adjusting our plan? We're at $X/mo — would $Y/mo
work with an annual commitment?"
4. If they say no:
"Understood. We may need to evaluate alternatives during our next
budget cycle. Is there anything else you can offer — maybe
additional seats or premium features at our current price?"
5. Always get it in writing: email confirmation or contract amendment.
Top Vendor by Category — Negotiation Targets
| Vendor Category | Typical Discount Achievable | Leverage Point |
|---|
| SaaS tools (< $5k/yr) | 10-20% | "We're consolidating vendors" |
| SaaS tools ($5k-50k/yr) | 15-30% | Annual commitment, competitive alternatives |
| SaaS tools (> $50k/yr) | 20-40% | Multi-year, logo value, formal RFP |
| Cloud (AWS/GCP/Azure) | 5-15% (committed use) | Reserved instances, committed spend discounts |
| Professional services (legal, accounting) | 10-30% | Fixed fee vs hourly, scope bundling |
| Recruiting agencies | 15 → 10% fee | Volume commitment |
| Insurance brokers | 10-20% | Annual re-quote, bundle policies |
Output Format
- Spend diagnostic (spend by category, vendor, department)
- Quick-win list with estimated savings and timeline
- Vendor negotiation queue (ranked, with talking points)
- Structural improvement roadmap
- Savings tracker (projected vs actual)
- Recommendation for CEO: what to do, what NOT to cut
Done Criteria
The skill is complete when:
- A full spend diagnostic is completed (spend by category, vendor, and department, with % of total).
- Quick wins are identified with estimated savings and implementation timeline.
- Top 5-10 vendors are ranked for negotiation with talking points and target discounts.
- Structural improvement opportunities are identified with effort vs impact ranking.
- A clear CEO recommendation is provided on what to cut and what NOT to cut.
- Savings tracker is set up to measure projected vs actual savings.
Pitfalls
- Cutting growth to save costs — the #1 startup cost optimization error. Cutting S&M that works, reducing engineering capacity on a growing product, or trimming customer support when NPS is fragile destroys future revenue to save today's dollars.
- Optimizing costs before you have clean data — without a full spend diagnostic, you don't know where the money is going. Optimizing blind means you might cut the wrong things and miss the real savings.
- One-time cost-cutting sprees without ongoing discipline — the company cuts 20% in a crisis, then creeps back to the old spend level within 6 months. Cost optimization is a muscle, not a one-time event. Build the review cadence.
- Killing the wrong SaaS tool — canceling a $200/mo tool that three engineers depend on saves $2,400/yr but costs $50,000 in lost productivity. Before canceling anything, verify usage AND impact.
- Negotiating from weakness — starting vendor negotiations when you're desperate (e.g., "we need 30% off or we'll default") is the worst position. Always negotiate from a position of data and options, not desperation.
Heuristics
- Don't cut growth to save costs: the #1 startup cost optimization error. Cutting S&M that works is self-defeating.
- SaaS sprawl is the easiest win: most startups have $5-20k/yr in unused/underused SaaS. A quarterly audit is worth it.
- Renegotiate annually: vendors expect it. If you haven't asked for a discount in 12 months, you're overpaying.
- Cloud costs are the sneaky budget-killer: they grow with usage and nobody watches them until the bill is shocking. Weekly monitoring.
- Fixed fee > hourly: for legal, accounting, and consulting — push for fixed-fee arrangements. Hourly billing incentivizes the wrong behavior.
Edge Cases
- Hypergrowth mode: optimizing costs aggressively can slow growth. Be selective — cut waste, not investment.
- Pre-product-market-fit: don't optimize. Focus entirely on finding PMF. Cost optimization is for when the model is working.
- Burning cash with strong unit economics: it's okay to spend if LTV/CAC > 3x and payback < 12 months. The constraint is cash, not P&L.
- Vendor dependency risk: if you're deeply integrated with a vendor, price negotiation leverage is limited. Focus on multi-year price locks instead.
Verification
Can you answer "Where is every dollar of OpEx going and which 20% of that spend is waste?" and "What are the top 3 things we should NOT cut?" and "Did we actually save what we projected from last quarter's optimization efforts?" If not, cost optimization is incomplete.
Example
User: "Find 10% savings in our OpEx without hurting growth."
Expected behavior: You run a full spend diagnostic (categorize every vendor, identify unused subscriptions, benchmark cloud costs), identify quick wins (cancel 3 unused tools saving $2k/mo, downgrade over-provisioned plans saving $1.5k/mo, negotiate annual billing on 5 vendors saving $8k/yr), recommend structural improvements (consolidate 3 analytics tools into 1 saving $3k/mo, migrate dev instances to reserved instances saving $1k/mo), and clearly flag what NOT to cut (S&M spend driving 4x ROAS, customer success headcount).
User: "We need to cut $100k/mo from our burn. It's a crisis situation."
Expected behavior: You quickly assess the full cost structure, identify that headcount is 75% of total spend, present a tiered menu of options with human impact (Level 1: hiring freeze saves $15k/mo, Level 2: contractor cuts + marketing hold saves $30k/mo, Level 3: 15% workforce reduction saves $70k/mo), recommend the combination needed to reach $100k/mo, provide a severance and communication plan for the difficult choices, and flag that at $100k/mo savings, runway extends from 9 to 15 months creating a path to the next fundraise.
Linked Skills
- Budget context →
budget-creation-management
- Profitability to know what's worth keeping →
profitability-analysis
- ROI case for bigger changes →
business-case-modeling