The opacity problem
Every consulting framework, every diagnostic tool, every «benchmark» spreadsheet operates on constants. Multipliers. Sensitivities. Recovery rates. Weights. They show up in board decks as if they were laws of nature.
Almost none of them tell you where the number came from.
Ask the partner: «Why is your churn-to-leakage multiplier 4.2 and not 3.6?» The answer is usually a hand-wave about «our experience across hundreds of engagements.» That is opinion, presented as data, sold at consultancy rates.
Caugia inverts this. The constants behind our Constraint Engine, our simulator, and our 45-page report are all published, version-controlled, and cited to named public sources. If you disagree with the source, you can audit it. If the underlying study revises, our constant moves with it.
What the framework calibrates
The GRIP Framework calibrates three families of constants, per vertical. The Constraint Engine and the what-if simulator both depend on them.
- K_DRAG, the sensitivity of revenue drag to system weakness. Higher K_DRAG means a 10-point GRIP gap creates more euro impact.
- DIMENSION_WEIGHTS, the relative importance of Guidance, Resources, Implementation, and Performance per vertical. Always sums to 1.00.
- RECOVERY_FACTORS, the share of identified leakage that is realistically recoverable per dimension, over a 12-month horizon.
These constants are not invented. They are conservative central estimates from named studies. Below is the current Phase 1 calibration, with the source anchor for each.
SaaS B2B
| Constant | Value |
|---|---|
| K_DRAG | 0.60 |
| DIMENSION_WEIGHTS | G 0.20 · R 0.20 · I 0.25 · P 0.35 |
| RECOVERY_FACTORS | G 0.55 · R 0.60 · I 0.65 · P 0.70 |
- Bessemer Cloud Index 2024. Rule-of-40 dispersion shows ~60% of below-median performers attribute their gap to identifiable GTM-system weakness → drives K_DRAG = 0.60.
- OpenView SaaS Benchmarks 2024. Performance-weighted dimension importance (NRR + win-rate + churn dominate rule-of-40 valuations) drives P = 0.35.
- OpenView cohort-turnaround studies 2022-2024. Performance fixes recover faster than Guidance, NRR improvements compound monthly, while org-design changes take quarters, drives the RECOVERY_FACTORS gradient (G < R < I < P).
DTC
| Constant | Value |
|---|---|
| K_DRAG | 0.55 |
| DIMENSION_WEIGHTS | G 0.15 · R 0.18 · I 0.22 · P 0.45 |
| RECOVERY_FACTORS | G 0.50 · R 0.65 · I 0.70 · P 0.75 |
- Drivepoint cohort recovery studies 2022-2024 (n=188 brands, €5-15M ARR). DTC drag-to-fix half-life is roughly 4 months, against SaaS’s 9 months. The short cash-conversion cycle drives a lower K_DRAG than SaaS.
- Triple Whale State of DTC 2024. Performance dominates the model, repeat-rate, MER, and AOV are existential, driving P = 0.45 (vs SaaS’s 0.35).
- Drivepoint margin studies. Category demand elasticity in DTC drives the high RECOVERY_FACTORS across the board.
Fintech B2B
| Constant | Value |
|---|---|
| K_DRAG | 0.70 |
| DIMENSION_WEIGHTS | G 0.25 · R 0.30 · I 0.30 · P 0.15 |
| RECOVERY_FACTORS | G 0.45 · R 0.50 · I 0.55 · P 0.60 |
- Acrew Capital State of Fintech 2024. Median time-to-recover from a compliance block is approximately 14 months. Regulatory drag compounds → K_DRAG = 0.70 (the highest of the four verticals).
- FT Partners B2B Fintech Benchmarks 2024. Resources + Implementation dominate the model, compliance headcount, integration cost, sandbox-to-prod conversion, so R + I = 0.60 and P is constrained to 0.15.
- Plaid + Stripe Atlas onboarding cohort data 2022-2024. Regulatory inertia drives slower recovery → RECOVERY_FACTORS ceiling at P = 0.60.
Professional Services
| Constant | Value |
|---|---|
| K_DRAG | 0.50 |
| DIMENSION_WEIGHTS | G 0.30 · R 0.30 · I 0.20 · P 0.20 |
| RECOVERY_FACTORS | G 0.60 · R 0.65 · I 0.65 · P 0.60 |
- Service Performance Insights PS Maturity Benchmark 2024. ProfSvc scales linearly with headcount, not exponentially through bottlenecks. A 10% utilization gap creates a roughly 10% revenue gap, not a compounding cascade → K_DRAG = 0.50 (the lowest of the four).
- Kennedy Pulse Survey 2024 (consulting). Guidance + Resources dominate, people-business: strategy and talent drive outcomes → G + R = 0.60.
- SPI Annual PS Operational Excellence 2024. Engagement-margin feedback loops drive medium-rate recovery uniformly across dimensions.
Confidence bands, not single numbers
The simulator never reports a single number. Every projection comes with a 50% confidence interval, a P25/P75 spread around the P50 point estimate.
The width of the interval follows the prediction-interval methodology documented in Bessemer’s 2024 SaaS forecasting playbook: roughly ±25% of the central estimate, as the operating-frontier band for forward-looking GTM projections.
The reason is intellectual honesty. We can tell you with reasonable confidence where the number is likely to land. We cannot tell you with certainty where it will land, because no one can. Showing the band makes that explicit, and stops the customer from anchoring on the midpoint as a promise.
Phase 1 is calibration. Phase 2 is backtest.
The Caugia framework runs in two phases.
Public-benchmark calibration
All current constants, per-vertical K_DRAG, dimension weights, recovery factors, slider-to-GRIP weights, stage benchmarks, confidence-band swing parameters, are anchored against the named studies above. They are conservative central estimates from the public literature, not Caugia opinions. A prospect can independently verify every number against the source we cite.
Cohort backtest
Once 10 paying customers per vertical (40 customers total) have at least 6 months of monthly snapshots, the calibration program refits the constants against observed outcomes:
- K_DRAG per vertical against observed drag-to-fix half-life.
- Dimension weights per vertical against observed predictive power per dimension.
- Recovery factors per vertical against observed 12-month recovery trajectories.
- Confidence-band swing against observed prediction-vs-outcome dispersion, the cohort-derived bounds replace the Bessemer-derived ±25% default.
- Per-pillar within-dimension weights replace the equal-weighting default.
Deliverable: a mean-absolute-error report per vertical, R² of upsideIfFixed predictions vs realised 6- and 12-month outcomes, and a Phase-2 constant table committed to the methodology document with a per-constant delta against Phase 1.
What this means for you
If you are a prospect: every Caugia number on your screen has an audit trail. The methodology document on the Caugia repo lists each constant with its source. Disagree with Bessemer’s 2024 dispersion analysis? Push back. We will explain how the constant was derived and what would need to be true for it to move.
If you are a VC running diligence: the framework is productized, not improvised. Every constant has a roadmap to empirical recalibration with explicit gates. The Phase 2 backtest is a public commitment with a date attached. If we miss it, you can hold us to it.
If you are a customer: the engine you are paying for is not a black box. You can trace every number on your 45-page report back to the framework, and every framework constant back to a named public study. If the underlying study revises, our constant moves, and we publish the delta.
The anti-framework position
Caugia is not Gartner. We are not McKinsey. We are not building a thought-leadership franchise on top of opinion.
We are building a deterministic engine with published math, calibrated against named sources, with a public roadmap to empirical recalibration. That discipline is the product. Everything else, the assessment, the simulator, the brief, the actions, runs on top of it.
If a competitor publishes a more rigorous methodology than ours, we lose. That is the point. We have made the methodology auditable on purpose, so the discipline of publishing the math is the durable advantage.
Frequently Asked Questions
Where do Caugia's GRIP framework constants come from?
Every per-vertical constant is anchored to a named source from the public benchmark literature. K_DRAG, the dimension weights across Guidance, Resources, Implementation and Performance, and the recovery factors are conservative central estimates from studies like the Bessemer Cloud Index, OpenView SaaS Benchmarks, Triple Whale State of DTC, Acrew State of Fintech, and SPI's PS Maturity Benchmark. They are published and version-controlled, so you can audit each one. The free GTM diagnostic runs these constants against your own numbers at no cost.
What is K_DRAG and why does it differ by vertical?
K_DRAG is the sensitivity of revenue drag to system weakness: a higher K_DRAG means a 10-point GRIP gap converts into more euro impact. It is 0.60 for SaaS B2B, 0.55 for DTC, 0.70 for Fintech B2B, and 0.50 for Professional Services, because regulatory drag in fintech compounds while professional services scale linearly with headcount. Each value is tied to an observed drag-to-fix half-life from a named cohort study, not to opinion.
Can I verify Caugia's numbers myself?
Yes. The methodology document lists every constant with the public study it is derived from, so a prospect or a VC running diligence can independently check each number. If you disagree with a source, you can push back and we will explain how the constant was derived and what would need to be true for it to move. If the underlying study revises, the constant moves with it and we publish the delta.
How is this different from a consulting benchmark or a Gartner number?
A consulting multiplier is usually opinion presented as data, justified by experience across hundreds of engagements. Caugia is a deterministic engine: the constants are published, cited to named sources, and carry a public roadmap to empirical recalibration once cohort data is available. The discipline of publishing the math is the product, which is why the diagnostic names a single binding constraint and quantifies the revenue leakage it causes in euros rather than handing you a generic score.
How do I see the calibrated framework score my own GTM?
Run the free GTM diagnostic. It costs EUR 0, needs no card, evaluates your system across 12 GTM pillars, names your single binding constraint, and quantifies the revenue leakage in euros, all using the calibrated constants in this article. From there, the Pulse report (EUR 249) and the full Report (EUR 750) go deeper, while GRIP OS with the Sophie copilot turns the diagnosis into ongoing execution. Caugia also runs an AI Answer Market that tracks how six engines, ChatGPT, Claude, Gemini, Perplexity, Grok and Mistral, describe your category.
See the calibrated framework score your GTM
Run the free GTM diagnostic and watch these constants resolve against your own numbers. EUR 0, no card required. Or read the framework foundations first, four dimensions, twelve pillars, then come back here for the calibration.