You're In — Full Checklist Below
You've been added to The Product Economist briefing. Here's the complete diagnostic framework.
The Complete R&D Audit Checklist
The 38 questions used in every $7,500 diagnostic engagement. Organized across 6 domains with traffic-light scoring, remediation actions, and benchmark thresholds. This is the same framework used to audit engineering organizations at companies from Series A startups to Fortune 500 enterprises.
📖 How to Use This Checklist
Score each question using the traffic-light rubric. Be honest — this is for your benefit, not anyone else's.
Count your red scores. These are your highest-impact remediation opportunities. Start with the domain with the most red.
Use the action items for each question. Tackle 2-3 red items per quarter. Track progress with the free tools below.
Domain 1: Engineering Velocity & Delivery
How fast and reliably does your engineering organization deliver value?
What percentage of engineering time is spent on maintenance vs. new features?
Why: If maintenance exceeds 40%, you may be approaching Technical Insolvency.
Calculate Innovation Tax: maintenance hours ÷ total hours. Track monthly.
What are your DORA metrics (deploy frequency, lead time, failure rate, MTTR)?
Why: DORA measures delivery speed. Pair with PDI to see if you're shipping fast toward insolvency.
Instrument CI/CD pipeline. Track all 4 metrics weekly.
What is your cycle time from commit to production?
Why: Long cycle times compound delays and reduce feedback speed.
Measure commit-to-production time. Target: <1 hour for elite teams.
How often do deployments cause incidents?
Why: Change failure rate directly measures deployment quality.
Calculate: failed deployments ÷ total deployments.
What is your average sprint completion rate?
Why: Consistently missing sprint commitments signals estimation or capacity problems.
Track: stories completed ÷ stories committed per sprint.
Do you have feature flags for safe rollouts?
Why: Feature flags enable incremental releases, A/B testing, and instant rollback.
Implement feature flag system. Target: all new features behind flags.
What is your code review turnaround time?
Why: Slow reviews create bottlenecks and context-switching costs.
Measure: time from PR open to first review. Target: <4 hours.
Domain 2: Technical Debt & Architecture
What is the health of your technology capital, and where is value being destroyed?
Can you identify your 3 largest sources of technical debt and their financial impact?
Why: Most organizations cannot quantify debt in dollars. Without financial language, leadership ignores it.
Run PDI assessment. Assign dollar values to top debt categories.
What is your Technical Insolvency Date?
Why: The exact quarter when maintenance costs consume 100% of engineering capacity.
Plot Innovation Tax trend. Extrapolate to 100%. That's your insolvency date.
What percentage of your codebase has test coverage?
Why: Low coverage = high change failure rate = slow delivery = more rework costs.
Measure line/branch coverage. Target: >70% for critical paths.
When was your last architecture review?
Why: Architecture debt is the most expensive form of debt — it requires rewrites, not refactors.
Establish quarterly Architecture Review Board. Document all decisions.
How many services or modules have a single maintainer?
Why: Single points of failure. If that person leaves, the knowledge leaves with them.
Audit: map each service to its maintainers. Cross-train where count = 1.
What is the age distribution of your critical dependencies?
Why: Outdated dependencies = security vulnerabilities + compatibility issues + upgrade debt.
Audit dependency ages. Flag anything >2 major versions behind.
Do you have automated security scanning in your CI/CD pipeline?
Why: Manual security reviews don't scale. Automated SAST/DAST catches vulnerabilities before production.
Integrate SAST tool. Block merges with critical vulnerabilities.
A 30-minute Gut-Check call identifies whether you have a real problem — or just technical anxiety.
Domain 3: AI & Emerging Technology Economics
Are your AI investments creating or destroying value?
What is the fully-loaded cost per AI inference request?
Why: AI features often have hidden variable costs that erode gross margins.
Instrument per-request cost tracking: compute + tokens + storage + overhead.
Do you use model routing (different models for different query types)?
Why: Using frontier models for every query costs 10-50x more than necessary.
Classify queries by complexity. Route 70% to smaller, cheaper models.
What percentage of your AI features have positive unit economics?
Why: 40-60% of AI features launch unprofitable. Growth accelerates losses.
Calculate per-feature P&L. Kill or optimize negative-margin features.
How much of your production code was generated by AI, and what's its defect rate?
Why: Vibe-coded applications accumulate hallucination debt — debt no one on the team fully understands.
Track AI-generated code percentage. Measure defect rate vs. human-written code.
Do you have a model right-sizing strategy?
Why: Using a Ferrari for the mailbox. Right-sizing cuts AI costs 60-80%.
Benchmark: test smaller models against quality thresholds. Document findings.
What guardrails exist for AI output quality?
Why: Without guardrails: hallucinations, bias, and harmful outputs reach users.
Implement output validation, safety filters, and quality monitoring.
Domain 4: Product & Revenue Alignment
Is engineering investment aligned with revenue generation?
What is your Revenue Per Engineer (RPE), and how does it trend?
Why: Declining RPE signals engineering capital misallocation.
Calculate: ARR ÷ engineering headcount. Track quarterly. Use APER calculator.
Can you identify which features generate revenue and which are zombie features?
Why: Most organizations maintain features that destroy value. 30-50% of features have <5% usage.
Instrument feature usage. Identify features with <5% MAU. Run Kill Switch Protocol.
Do your PMs own a P&L, or just a backlog?
Why: PMs who don't understand their P&L make uninformed capital allocation decisions every sprint.
Create per-product P&L. Train PMs on unit economics.
Can you calculate the gross margin of each product line?
Why: AI features introduce variable COGS. Without margin visibility, you may be scaling losses.
Allocate engineering + infrastructure costs per product. Calculate margins.
What would happen if you removed your 10 least-used features tomorrow?
Why: The Kill Switch Protocol typically recovers 20-40% of engineering capacity from zombie features.
List 10 lowest-usage features. Calculate maintenance cost of each. Draft removal plan.
What is your time-to-revenue for new features?
Why: Long time-to-revenue means engineering investment isn't generating returns fast enough.
Track: feature release date → first revenue attribution. Target: <30 days.
The Insolvency Diagnostic quantifies your exposure and delivers a written Risk Report with prioritized remediation.
Domain 5: Organization & People
Is your team structured for sustainable, scalable delivery?
What is your engineering attrition rate over the last 12 months?
Why: Each departure costs $150K-250K (recruiting + onboarding + lost productivity).
Calculate: departures ÷ average headcount × 100.
What is the average tenure on your engineering team?
Why: Low tenure means constant knowledge loss and ramp-up costs.
Track average tenure. Flag teams with <18 month average.
Is your engineering org structured around products or projects?
Why: Project-based teams ship and move on. Product teams own outcomes.
Evaluate: do teams own products end-to-end, or get assigned projects?
What is your span of control (direct reports per manager)?
Why: Below 5: manager overhead too high. Above 8: insufficient coaching.
Audit: count direct reports per manager. Restructure outliers.
How many key-person dependencies exist?
Why: If one person's departure would halt a project, that's a critical risk.
Map: for each critical system, who are the only people who understand it?
Do you have a documented career ladder with clear levels?
Why: Without clear progression, top engineers leave for companies that offer it.
Publish engineering career ladder. Review annually.
Domain 6: Strategic & Financial
Is your R&D investment being valued, reported, and optimized at the board level?
What percentage of your "R&D spend" is actually maintenance OpEx?
Why: The Innovation Tax — many companies report 50% R&D investment when 80% is actually maintenance.
Audit: categorize every engineering hour as innovation vs. maintenance.
If a PE firm audited your engineering organization today, what would they find?
Why: Technical Due Diligence reveals hidden liabilities. Better to find them yourself.
Conduct an internal pre-diligence audit using this checklist + PDI tool.
What is the accuracy-cost curve for your critical AI features?
Why: Going from 80% to 95% accuracy often costs 10x more. The Cost of Predictivity must be modeled.
For each AI feature: plot accuracy vs. cost. Find the diminishing returns inflection point.
Can your engineering investment survive a 30% budget cut?
Why: Knowing your critical path vs. nice-to-have helps make tough decisions before they're forced.
Create a tiered investment plan: must-have (70%), should-have (20%), nice-to-have (10%).
Do you report engineering health metrics to the board?
Why: Boards that see engineering metrics make better investment decisions.
Create quarterly technology capital report: PDI, APER, DORA, Innovation Tax.
What is the total cost of ownership for your technology stack?
Why: Most companies underestimate TCO by 40-60%. Hidden costs: maintenance, integration, training, migration.
Map TCO for each major platform: license + integration + maintenance + opportunity cost.
🛠️ Answer These Questions With Free Tools
Don't guess — use our calculators to get accurate scores for the metrics above.
Want These Questions Answered Professionally?
Book a diagnostic engagement and get a written executive summary with quantified findings, benchmarks, and a prioritized remediation roadmap.