Redirecting to interactive showcase…
Use Case: Pricing & Packaging | CodeLens AI Usage-Based Redesign
A Complete Pricing Strategy with Unit Economics and Migration Planning
Last Updated: March 2026
Executive Summary
CodeLens AI is a code review tool with 8,200 paying customers at a flat $49/month ($4.8M ARR). The flat-rate model is structurally broken: the top 10% of users (500+ AI reviews/month) consume 62% of AI compute at $0.03/review, creating margin-negative accounts, while the bottom 50% (<20 reviews/month) find $49 too expensive for their usage.
This document applies the pricing-packaging skill from the PM Skills Arsenal to design a complete pricing strategy including:
- Pricing Model Selection — Hybrid (base + usage) chosen over flat-rate, per-seat, and pure usage-based
- WTP Assessment — Van Westendorp (n=47): OPP $58, IDP $72, PME $119
- Competitive Pricing Map — 6 competitors mapped; CodeLens positioned as premium AI-native
- Good/Better/Best Packages — Starter $29/mo (50 reviews), Pro $69/mo (250 reviews), Enterprise $149/mo (1,000 reviews)
- Sensitivity Analysis — Revenue robust across +/-30% price variations
- Revenue Impact — Projected $7.4M ARR (+54% uplift) with 8% migration churn
- AI Cost Patterns — Per-review metering, cost trajectory analysis, model quality gating
- Migration Strategy — 5-phase plan: 6-month grandfather, personalized tier assignment, 20% loyalty discount
Key Numbers:
| Metric | Current | Projected |
|---|---|---|
| ARR | $4.8M | $7.4M |
| ARPU | $49/mo | $81.72/mo |
| Gross Margin | 52% | 68% |
| Margin-Negative Accounts | ~200 | 0 |
Frameworks Applied
- Pricing Model Selection — Decision matrix evaluating per-seat, flat-rate, usage-based, and hybrid models against cost structure, value delivery, buyer predictability, and competitive norms
- Van Westendorp Price Sensitivity Meter — Four-question method for finding the acceptable price range, applied to 47 customer interviews
- Competitive Pricing Map — Price-value positioning against SonarQube, Codacy, DeepSource, Snyk Code, GitHub Copilot, and Cursor
- Good/Better/Best Package Architecture — Tier design with feature allocation rationale, upgrade triggers, and margin analysis
- AI/SaaS-Specific Pricing Patterns — Value metric alignment, marginal cost structure, metering approach, GPU cost trajectory
- Sensitivity Analysis — Revenue impact at 7 price variations plus key variable sensitivity
- Revenue Impact Modeling — Churn risk, grandfathering cost, expansion uplift, net 12-month impact
Evidence Quality
- 54 total evidence points (T1-T4: 48; T5: 6; T6: 0)
- All pricing recommendations cite minimum 2 evidence tiers
- Core model decision supported by T1 (usage data), T2 (competitive norms), T3 (customer interviews)
- Revenue projections are T4-T5 (directional) — recommend A/B pricing test for T1 validation
Skill Used
pricing-packaging — Produces pricing and packaging strategy documents with model selection routing, willingness-to-pay assessment, competitive pricing maps, Good/Better/Best package architecture, sensitivity analysis, AI/SaaS-specific pricing patterns, and migration planning.