Startup Business Models

Systematic workflow for choosing revenue models, pricing, and unit economics.

Quick Start (Inputs)

Ask for the smallest set of inputs that makes the decision meaningful:

Business type: SaaS, usage-based/API, marketplace, services, hardware + service
ICP/segment(s): SMB / mid-market / enterprise (and ACV/ARPA bands)
Current pricing and packaging: value metric, tiers, limits, discount policy, billing cadence
Unit economics drivers: fully-loaded CAC, gross margin/COGS (include LLM/infra/third-party), churn/retention, expansion (NRR)
Constraints: sales motion (PLG vs sales-led), implementation constraints (billing metering, proration), gross margin floor, payback target

If numbers are missing, proceed with ranges + explicit assumptions and highlight what to measure next.

Workflow

Classify the model

Subscription, usage-based, freemium, marketplace take-rate, transaction fee, ads, outcome-based, credit-based, hybrid.

Build a segment-level unit economics snapshot

Use references/unit-economics-calculator.md for formulas, benchmarks, and common pitfalls.
Prefer cohort/segment views over blended averages.

Evaluate model fit and risks

Align price metric with value delivered and cost incurred (especially usage + AI compute).
Identify failure modes: margin compression, adverse selection, channel conflict, support cost explosions, metering/overage friction.

Propose pricing + packaging changes

Use references/pricing-research-guide.md for WTP methods and pricing interview scripts.
Use assets/pricing-tier-design.md to draft tiers, limits, upgrade triggers, and enforcement rules.

Define measurement and roll-out

Define success metric + guardrails, evaluation design, and explicit lag windows (conversion now, retention later).

Deliver a decision-ready output

Recommendation, rationale, assumptions, scenarios (base/best/worst), and next experiments.

2026 Heuristics (Context-Dependent)

Prioritize payback and gross margin over a single ratio; LTV:CAC is easiest to game.
Typical SaaS targets (directional, by segment/stage): LTV:CAC 3-5x, payback 6-12 months (PLG) or 12-18 months (sales-led early), NRR >100% (mid-market/enterprise) and gross margin >70% (software-only).
For usage-based / AI products: model contribution margin per unit (token/job/workflow) and set pricing guardrails (rate limits, minimums, commit tiers, credit expiries).

Related Skills (Routing)

Pricing Change Measurement & Experiment Design

Use this when you are changing pricing, packaging, value metric, limits, discounts, or billing cadence.

1) Define success and guardrails (before launch)

| Type | Examples | |------|----------| | Primary success metric | Net revenue retention (NRR), ARPA/ARPU, gross margin %, payback period, upgrade rate, expansion MRR | | Guardrails | New logo conversion, activation rate, refund rate, support load, churn (logo + revenue), sales cycle length |

2) Pick an evaluation design

| Design | Best when | How to read results | |--------|-----------|---------------------| | A/B (randomized) | Self-serve / PLG flows | Compare conversion, ARPA, refunds, and downstream retention by assignment | | Holdout/control cohort | Pricing is hard to randomize | Compare treated vs. holdout cohorts matched on segment, channel, and start month | | Step rollout (time-based) | Enterprise contracts, invoicing cycles | Compare pre/post with a parallel cohort (not exposed yet) to reduce seasonality bias | | Geo/account rollout | Regions/segments are separable | Compare regions/segments; watch for channel mix shifts |

3) Use explicit lag windows (avoid premature conclusions)

Short lag (days to 2 weeks): checkout conversion, activation, sales cycle friction, refund/support spikes.
Medium lag (4 to 8 weeks): upgrades, expansion MRR, usage growth, discounting behavior, proration effects.
Long lag (90 to 180+ days, B2B): churn, net revenue retention, renewal outcomes, contraction risk.

4) Report an "all-in" view (not just conversion)

Revenue quality: net revenue after refunds, discounts, and credits; gross margin impact (including variable compute/COGS).
Segments: break down by plan, seat band, channel, ACV/ARR band, and customer age (new vs. renewal).
Decision rule: write a go/no-go threshold (example: "NRR +2pts with no >0.5pt drop in activation and no >10% increase in support load").

SaaS Metrics (Read When Needed)

Use references/saas-metrics-playbook.md for definitions and templates (MRR/ARR, churn, NRR, Quick Ratio, Magic Number, burn multiple, stage focus).

Resources

| Resource | Purpose | |----------|---------| | unit-economics-calculator.md | LTV, CAC, payback calculations | | pricing-research-guide.md | WTP research methodology | | saas-metrics-playbook.md | SaaS-specific metrics deep dive |

Templates

| Template | Purpose | |----------|---------| | business-model-canvas.md | Full model design | | unit-economics-worksheet.md | Calculate and track metrics | | pricing-tier-design.md | Pricing & packaging worksheet |

Data

| File | Purpose | |------|---------| | sources.json | Business model resources |

Do / Avoid (Jan 2026)

Do

Define your value metric (seat/usage/outcome) and validate willingness-to-pay early.
Include COGS drivers in pricing decisions (especially usage-based).
Use discount guardrails and renewal logic (avoid ad-hoc deals).

Avoid

Pricing as an afterthought (“we’ll figure it out later”).
Margin blindness (shipping usage growth that destroys gross margin).
Misleading LTV calculations from immature cohorts.

What Good Looks Like

Packaging: a clear value metric, tier logic, and discount policy (with enforcement rules).
Unit economics: CAC, gross margin, churn, payback, and retention defined and tied to cohorts.
Assumptions: one inputs sheet, ranges/sensitivities, and scenarios (base/best/worst).
Experiments: pricing changes tested with decision rules (not “gut feel” rollouts).
Risks: margin compression, adverse selection, channel conflict, and support cost modeled.

Optional: AI / Automation

Use only when explicitly requested and policy-compliant.

Summarize pricing research and competitor snapshots; verify manually before acting.
Draft pricing page copy; humans verify claims and consistency with contracts.