Generic values teach generic shapes.
Your data has a specific shape.

RoleYou are calculating per-tier conversion values for my B2B SaaS ad-algorithm signal layer. Take my actual closed-won data, ICP rubric tiers, and sales cycle. Derive per-tier values across 4 stages × 4 tiers = 16 cells per ICP segment. Calibrate to my unit economics, not generic placeholders.

My BrandBrand: [your B2B SaaS brand name]
Average ACV (last 12 months closed-won): [e.g. "$25,000 mid-market"]
Average sales cycle (form fill to closed-won): [e.g. "84 days"]
Target CAC payback: [e.g. "12 months for mid-market", see Track 05's strategy foundation for benchmarks]

ICP Rubric Tiers// Pull from Track 05's ICP Scoring Rubric Builder. If not yet run, define tiers based on firmographic + technographic + intent fit.

Tier A definition + score range: [e.g. "ICP score 80+: strict ICP fit, enterprise-size, expansion-ready"]
Tier B definition + score range: [e.g. "ICP score 60-79, fits ICP but with 1-2 gaps; SMB-mid market"]
Tier C definition + score range: [e.g. "ICP score 40-59, partial fit; lower probability buyers"]
Excluded definition: [e.g. "ICP score <40, non-fit; don't optimize for these"]

Closed-Won Data (Last 12 Months)// Pull from HubSpot. If < 12 months data available, use what's available + flag low-confidence calibration.

Total closed-won deals: [e.g. "127 deals"]
Closed-won by tier: [e.g. "Tier A: 78 deals (61%) · Tier B: 38 deals (30%) · Tier C: 11 deals (9%)"]
ACV by tier: [e.g. "Tier A: $32K avg · Tier B: $19K avg · Tier C: $11K avg"]
Sales cycle by tier: [e.g. "Tier A: 72 days · Tier B: 88 days · Tier C: 105 days"]

Funnel Conversion Rates// Cohort calculation: rates from MQL→SQL must use time-lagged cohorts (don't compare same-month MQLs and SQLs for cycles > 30 days).

MQL → SQL conversion rate: [e.g. "Tier A: 38% · Tier B: 22% · Tier C: 9% · blended: 28%"]
SQL → Opportunity rate: [e.g. "Tier A: 70% · Tier B: 52% · Tier C: 31% · blended: 60%"]
Opportunity → Closed-Won rate: [e.g. "Tier A: 35% · Tier B: 22% · Tier C: 12% · blended: 28%"]

TaskCalculate per-tier values across 4 stages:

1. Closed-Won values per tier:
   - Tier A CW value = Tier A average ACV (full deal value)
   - Tier B CW value = Tier B average ACV
   - Tier C CW value = Tier C average ACV
   - Excluded CW value = $0

2. Opportunity values per tier:
   - Tier A Opp value = Tier A CW value × Tier A Opp→CW rate
   - Tier B Opp value = Tier B CW value × Tier B Opp→CW rate
   - Tier C Opp value = Tier C CW value × Tier C Opp→CW rate
   - Excluded Opp value = $0

3. SQL values per tier:
   - Tier A SQL value = Tier A Opp value × Tier A SQL→Opp rate
   - Same formula for Tier B, Tier C
   - Excluded SQL value = $0

4. MQL values per tier:
   - Tier A MQL value = Tier A SQL value × Tier A MQL→SQL rate
   - Same formula for Tier B, Tier C
   - Excluded MQL value = $0
   - Apply MQL noise discount (multiply by 0.6-0.8 to account for MQL signal volatility, see calibration notes)

5. Target CAC payback adjustment:
   - Compare derived MQL value to target MQL CAC
   - If derived value > 2x target CAC → values are too high; algorithm will over-bid
   - If derived value < 0.5x target CAC → values are too low; algorithm will under-bid
   - Adjust derived values within reasonable bounds (typically 0.7-1.3x of pure mathematical derivation) to meet payback target

6. Per-tier value differential check:
   - Tier A:Tier B ratio: should be 1.5-3x (not too aggressive, not too compressed)
   - Tier B:Tier C ratio: should be 5-15x (Tier C accounts close at much lower rates; algorithm should rarely pursue them)
   - Tier C:Excluded: hard cutoff at $0 for Excluded
   - If ratios fall outside these bounds, flag for review

Output format1. Headline: total deals analyzed, average ACV by tier, derived value summary, calibration confidence (high/medium/low based on data volume).

2. 4×4 value matrix: rows = stages (MQL/SQL/Opp/CW), columns = tiers (A/B/C/Excluded). Each cell = derived dollar value.

3. Payback assumption breakdown: target MQL CAC, target SQL CAC, projected CAC payback at derived values, headroom or shortfall.

4. Per-platform deployment specs: Google Ads (offline conversion event values), LinkedIn Ads (CAPI conversion values), HubSpot lifecycle event mapping.

5. Honest calibration:
   - If closed-won data is < 50 deals over 12 months, calibration confidence is LOW. Use derived values directionally but expect re-calibration after more data accumulates.
   - If Tier A:B ratio falls outside 1.5-3x, the ICP rubric may need refinement. Re-check Track 05's ICP Rubric Builder before deploying.
   - If derived MQL values are < $30 across all tiers, the MQL signal is too weak to provide useful algorithm signal. Consider sending only SQL/Opp/CW events.
   - If sales cycle is < 30 days, MQL noise discount should be 0.5-0.6 (faster cycles → noisier MQLs). If > 120 days, discount should be 0.8-0.9.
   - For PLG companies, add a Product Activation event between MQL and SQL valued at 0.5x SQL value. Activation is more reliable than MQL for PLG.
   - If you don't have offline conversions deployed yet (Track 05's Google + LinkedIn CAPI setup), values are theoretical until deployment. Run those workflows first.

// Be specific in calculations. Show the math: "Tier A SQL value = $32K × 35% (Opp→CW) × 70% (SQL→Opp) = $7,840" not just "Tier A SQL = $7,840."
// Apply common-sense rounding to deployable values. $1,047.83 rounds to $1,000. $187.42 rounds to $200. Algorithms don't need precision; round numbers reduce drift over time.
// MQL noise discount is critical. Without it, MQL values are mathematically correct but operationally over-weighted. The 0.6-0.8 discount accounts for signal volatility.
// Don't recommend changing tier definitions based on calculation output. The calculator values existing tiers; rubric refinement is the ICP Rubric Builder's job.