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 the Constraint Engine, the simulator and the Intelligence Report are published, version-controlled in the engine’s source, and explained one by one below. The one thing we refuse to do is dress a modelling choice up as a measurement.
Two kinds of numbers, two standards of proof
- Model constants. K_DRAG, the dimension weights and the recovery factors are calibration choices Caugia made, set conservatively and kept in code under version control. They are published here with the rationale for each value. They are not measurements, and Caugia does not claim they were derived from a named study.
- Peer benchmarks. When the Intelligence Report compares your growth, retention, CAC payback, gross margin, ARR per employee or burn multiple with peers, the value is read directly from a named public edition and the report prints the source, the edition and the peer band next to it. Where no public per-band source exists, the report says so and uses the vertical reference set.
The first kind can be challenged on reasoning. The second kind can be checked against the publication. We label each so you always know which one you are looking at.
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. Revenue drag on a dimension equals revenue × K_DRAG × the dimension’s weight × its weakness (one minus the score divided by 100). A higher K_DRAG turns the same GRIP gap into more euros.
- DIMENSION_WEIGHTS, the relative importance of Guidance, Resources, Implementation and Performance per vertical. Always sums to 1.00.
- RECOVERY_FACTORS, the share of identified drag that is realistically recoverable per dimension over a 12‑month horizon. Recoverable upside equals the drag × the recovery factor of the binding dimension.
Below is the current calibration, exactly as it sits in the engine’s source, with how each set was reasoned.
B2B SaaS
| 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 |
- K_DRAG 0.60 is the reference multiplier the other verticals are tuned against: at full weakness on the heaviest dimension it caps the modelled drag at 21% of ARR, which keeps the leakage figure conservative rather than dramatic.
- Performance carries 0.35 because retention, conversion and pricing discipline convert most directly into ARR; Guidance and Resources carry 0.20 each because their effect travels through the other dimensions first.
- Recovery rises from G to P because pricing and retention fixes land inside a quarter, while strategy and organisation changes take longer to show up in revenue.
Fintech B2B
| Constant | Value |
|---|---|
| K_DRAG | 0.60 |
| DIMENSION_WEIGHTS | G 0.15 · R 0.35 · I 0.25 · P 0.25 |
| RECOVERY_FACTORS | G 0.50 · R 0.55 · I 0.65 · P 0.70 |
- K_DRAG stays at 0.60. Regulated revenue is sticky once a client is live, so drag realises slower than in DTC, but activation gates compound, so it does not realise slower than in SaaS.
- Resources carries 0.35 because pricing, compliance and unit economics are where a payments business wins or loses; Guidance is lighter because the licence shape already sets much of the strategy.
- Recovery is capped lower on Guidance and Resources because licence, compliance and pricing changes move at the regulator’s pace, while activation and retention fixes land within the year.
DTC (published for the method, not a served vertical)
| Constant | Value |
|---|---|
| K_DRAG | 0.55 |
| DIMENSION_WEIGHTS | G 0.18 · R 0.24 · I 0.28 · P 0.30 |
| RECOVERY_FACTORS | G 0.50 · R 0.65 · I 0.70 · P 0.65 |
- K_DRAG 0.55 is marginally tighter than SaaS because DTC revenue is non-recurring: drag realises in the period, not in the next cohort, so the multiplier is smaller.
- Implementation and Performance carry the weight because creative operations, channel mix and retention drive near-term revenue; brand and strategy move slower.
Professional services (published for the method, not a served vertical)
| Constant | Value |
|---|---|
| K_DRAG | 0.58 |
| DIMENSION_WEIGHTS | G 0.22 · R 0.24 · I 0.28 · P 0.26 |
| RECOVERY_FACTORS | G 0.50 · R 0.60 · I 0.68 · P 0.65 |
- K_DRAG 0.58 sits between DTC and SaaS: leverage compounds through headcount, but fee revenue is non-recurring and realises drag in the period.
- Implementation carries the most weight because delivery execution and scope discipline are the realisation engine of a practice; recovery is highest there for the same reason.
Where the named sources come in
The constants above decide how a weakness turns into euros. The peer benchmarks decide what «weak» means for a company of your size. Those come from public editions Caugia has read directly, and every row carries the source, the edition, the band exactly as the source defines it and the page it was read from:
- High Alpha SaaS Benchmarks 2025 (800+ respondents, data as of mid‑2025): year-over-year growth, net and gross revenue retention, CAC payback, software gross margin and ARR per employee, with quartiles, across the bands under $1M, $1‑5M, $5‑20M, $20‑50M and over $50M ARR.
- Benchmarkit SaaS and AI Metrics Benchmarks 2026 (342 participants, 2025 data): burn multiple across the bands under $5M, $5‑20M, $20‑50M and $50‑100M ARR.
A company is matched on its revenue inside the source’s own band, so the report reads «vs $5‑20M ARR peers: median 31% · High Alpha SaaS Benchmarks 2025», never a blended or interpolated number. The rules for admitting a source: an open publication, figures per ARR band, the latest edition, and a population that matches Caugia’s customers (private B2B SaaS). Sources that fail a rule stay out, and the report says which comparisons still use the vertical reference set: win rate, sales cycle, pipeline coverage, discount and logo churn have no public per-band source today, and no public source publishes per-band figures for B2B payments software, so fintech comparisons stay on the vertical reference set for now.
Confidence bands, not single numbers
The simulator never reports a single number. Every projection comes with a P25/P75 spread around the P50 point estimate, by default about ±25% of the central estimate. That width is a Caugia modelling choice for forward-looking GTM projections, not a published prediction-interval methodology, and it is labelled as such.
The Intelligence Report’s score carries its own confidence band, and that one is computed, not chosen: ±8 points at full coverage, wider when questions were not applicable to the company or when a pillar was covered by only a few answers.
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 reader from anchoring on the midpoint as a promise.
What changes the constants
Versioned, challengeable, refit against outcomes when the data exists
Every constant lives in the engine’s source under version control, so a change is a visible diff with a date. Caugia refits a constant when observed outcomes justify it: once enough Execute engagements carry six months of monthly snapshots, the drag-to-fix half-life, the predictive power per dimension and the 12‑month recovery trajectories get compared with what the model assumed, and any constant that moves is published here with its delta. No date is promised for that refit, because it depends on the cohort existing. Until then the constants stay exactly as printed above.
What this means for you
If you are a prospect: every Caugia number on your screen has an audit trail. This page gives the constants and the formula; the Intelligence Report gives the source and edition behind each peer comparison. Disagree with a weight? Push back. We will explain how it was reasoned and what would need to be true for it to move.
If you sit on the board: the framework is productised, not improvised. A choice is labelled a choice, a measurement is labelled with its edition, and the report says in plain words which comparisons still rest on the vertical reference set.
If you are a customer: the engine you are paying for is not a black box. You can trace every number in your Intelligence Report back to the framework, every constant back to this page, and every peer value back to a publication you can open yourself.
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, with benchmarks read from named editions, and with a standing rule for recalibration against observed outcomes. 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?
They are Caugia's own calibration choices, set conservatively, version-controlled in the engine's source and published on this page with the rationale for each value. K_DRAG, the dimension weights across Guidance, Resources, Implementation and Performance, and the recovery factors are modelling constants, not measurements, and Caugia does not claim they were derived from a named study. The peer benchmarks the Intelligence Report compares a company against are a different kind of number: they are read directly from named public editions (High Alpha SaaS Benchmarks 2025, Benchmarkit SaaS and AI Metrics Benchmarks 2026) and cited next to each value. The free GTM Intelligence Pulse 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: revenue drag equals revenue × K_DRAG × the dimension weight × the weakness of that dimension, so a higher K_DRAG turns the same GRIP gap into more euros. It is 0.60 for B2B SaaS and fintech B2B, 0.55 for DTC and 0.58 for professional services. The differences reflect how fast drag realises in each revenue model: non-recurring DTC revenue realises drag within the period, regulated fintech revenue is sticky once live, and professional-services leverage compounds through headcount. These are Caugia's published modelling choices, open to challenge.
Can I verify Caugia's numbers myself?
Yes, on two levels. The model constants are printed on this page with the formula they enter, so you can recompute any drag or recovery figure in a report. The peer benchmarks are cited with source, edition and peer band next to each comparison line in the Intelligence Report, and its methodology page lists the sources behind that report; every cited edition is an open publication you can read yourself. Where no public per-band source exists, the report says that the comparison uses the vertical reference set.
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 separates the two cleanly: the constants it chose are labelled as choices and published with their rationale, and the benchmarks it reads are labelled with the edition they come from. Both are deterministic, 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 Intelligence Pulse: 31 questions, no card, a binding hypothesis with its estimated revenue impact, delivered by email. The GTM Intelligence Report (EUR 750) runs the full 12-pillar assessment, names the binding constraint and prints every peer comparison with its source. From there Tom Meijer, Caugia's founder, executes the 90-day plan with your team as a fractional operator, working in GRIP OS with Sophie.
See the calibrated framework score your GTM
Run the free GTM Intelligence Pulse 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.