Stg Calculating Economics

Economics Calculation

Calculate unit economics with range-based estimates, integrated cost structure, and scenario analysis. Every input carries tier label; output inherits highest (weakest) tier. Merged from old str-calculating-economics and str-structuring-costs.

Procedure

Step 1: Gather Inputs with Tier Labels [S]

Read: pricing inputs (from stg-designing-pricing), market sizing (from stg-sizing-markets), competitive data (from stg-analyzing-competition), solution design (growth architecture).

Assemble input table:

Produce: input table with tier labels and source citations.

If channel strategy outputs are available (from stg-designing-channels), use per-channel CAC estimates and investment splits as the primary source for S&M spend and blended CAC inputs. This provides more rigorous CAC decomposition than freestanding S&M estimation.

Gate: inputs_gathered: bool -- all 5 inputs have values, tier labels, and sources.

• Pass: Step 2.
• Fail: For missing inputs, use category benchmarks with T2 label. If no benchmarks available, use SaaS median with T3 label and wide range.

Step 2: Calculate Core Metrics as Ranges [S]

Calculate for optimistic, base, and pessimistic input sets:

LTV = ARPU x Gross Margin x (1 / Monthly Churn Rate)

CAC = (S&M Spend + Allocated Overhead) / New Customers Acquired

LTV:CAC = LTV / CAC

Payback = CAC / (ARPU x Gross Margin)

Output tier = max(input tiers). If any input is T3, output is T3. If all inputs are T2, output is T2.

Produce: metric ranges across three scenarios.

Gate: metrics_calculated: bool -- LTV, CAC, LTV:CAC, and payback calculated for all three scenarios with tier labels.

• Pass: Step 3.
• Fail: If a formula input is missing, use the widest reasonable range and note which input is driving uncertainty.

Step 3: Calculate Cost Structure [S]

Read: solution design (features to build), growth architecture, constraints.

Fixed costs:

• Team: salaries, contractors (from role count x market rate)
• Infrastructure: servers, CI/CD, monitoring
• Software: SaaS tools, licenses
• Operations: legal, accounting, insurance

Variable costs (COGS):

• Hosting/compute per user (if AI inference, model per-interaction cost explicitly)
• Payment processing (2.9% + $0.30 typical)
• Support per ticket
• Onboarding per customer

Derived metrics:

• Gross margin = (revenue - COGS) / revenue
• Burn rate = fixed costs + variable costs at current scale
• Runway = available capital / monthly burn rate

Produce: cost structure with categories and monthly totals [T2].

Gate: cost_structure_calculated: bool -- fixed and variable costs categorized, gross margin calculated, burn rate and runway estimated.

• Pass: Step 4.
• Fail: If cost inputs are unknown, use industry benchmarks. AI inference costs are commonly missed -- always check if product has AI components.

Step 4: Validate Against Benchmarks [R]

Compare metrics against benchmark thresholds:

Flag any metric in warning or critical range with specific values.

Produce: benchmark validation with flags.

Gate: benchmarks_validated: bool -- all four metrics compared against thresholds.

• Pass: Step 5.
• Fail: Critical-range metrics are findings, not blockers. Report them.

Step 5: Add Growth-Model-Specific Metrics [R]

Based on growth architecture from solution design:

All estimates are T2 (from benchmarks) or T3 (predictions). Label each.

Produce: growth-specific metrics with tier labels.

Gate: growth_metrics_added: bool -- at least 3 growth-model-specific metrics estimated with tier labels.

• Pass: Step 6.
• Fail: If no benchmarks for the specific growth model, use general SaaS benchmarks and note the reduced confidence.

Step 6: Build Scenario Analysis [S]

Four scenarios:

• Optimistic: Favorable assumptions resolve (CAC at low end, LTV at high end, churn at low end).
• Base: Estimates as calculated.
• Pessimistic: CAC 3x estimate, segment 40% smaller, churn 2x benchmark.
• Kill: At what point do economics not work at all? State specific threshold values (e.g., "If CAC exceeds $X or churn exceeds Y%, LTV:CAC falls below 1:1").

Produce: four scenarios in register format.

Gate: scenarios_built: bool -- all four scenarios with specific numbers, pessimistic stress-tests assumptions.

• Pass: Step 7.
• Fail: If pessimistic is barely different from base, it is not a real stress test. Apply: 3x CAC, 40% smaller segment, 2x churn. If strategy only works in optimistic scenario, that is a finding.

Step 7: Mode-Specific Validation [R]

If base scenario fails mode thresholds, flag as critical finding.

Produce: mode validation result.

Gate: mode_validated: bool -- all mode-relevant thresholds checked with pass/fail.

• Pass: Step 8.
• Fail: Threshold failure is a finding, not a blocker. Report which threshold failed, by how much, and what would need to change.

Step 8: Write Hypothesis [S]

Write complete unit economics hypothesis in register format:

• Claim: One paragraph stating revenue model, LTV:CAC range, payback range, gross margin trajectory.
• Evidence: All calculations with sources and tier labels.
• Mode Thresholds: Table with required vs estimated for each metric.
• Scenario Analysis: Optimistic, base, pessimistic, kill -- with specific numbers.
• Assumptions: Every input that is T2 or T3 is an assumption (with tier, load-bearing flag, blast radius).
• Kill Condition: At what LTV:CAC ratio or payback period does this stop working? State specific numbers.
• Possibility Space: What alternative revenue models were considered and why eliminated.

Produce: complete unit economics hypothesis in register format.

Gate: hypothesis_written: bool -- all fields populated, no point estimates (all ranges), kill condition has specific thresholds.

• Pass: Done.
• Fail: Fix missing fields or convert point estimates to ranges.

Quality Criteria

• All calculations shown, not just results
• Every input cites source and carries tier label
• Three scenarios presented (optimistic, base, pessimistic) plus kill scenario
• Gross margin accounts for AI inference costs if applicable
• Mode-specific thresholds explicitly checked (pass/fail with numbers)
• No point estimates -- ranges throughout
• Cost structure broken into fixed and variable with category detail

Failure Modes

Boundaries

In scope: LTV/CAC/payback calculation, cost structure (fixed + variable), gross margin analysis, benchmark validation, growth-model-specific metrics, scenario analysis, mode-specific threshold validation, tier propagation.

Out of scope: Pricing design (stg-designing-pricing), market sizing (stg-sizing-markets), competitive analysis (stg-analyzing-competition), solution design (stg-designing-solutions).