IntakeStructured intake
We get what your operating leaders already know out of their heads and into a form a model can use. Most models are wrong because this step gets skipped.
Every decision has consequences inside the world your business operates in. We build that world and run it first.
Our Religion is a decision simulation platform. Your customers, your rivals, your market, the forces acting on all of them — built into a world that runs, and that your team puts decisions through before committing budget to any of them.
STRATA.
The platform layer that turns a business into a running world, and turns what that world does back into a decision. Six stages, executed in order. Nothing gets built before the question is clear — the first three exist to make it clear, and they are where most simulation projects quietly fail.
TranslationBusiness into world
Your business restated as a world: who the actors are, what forces act on them, and the one question worth answering. In plain English, not a diagram.
ReadinessScored out of thirty
Six lenses, scored. It comes after translation rather than before, so we’re scoring the model and not the meeting. Below sixteen we tell you, and we stop.
ArchitectureThe world design
Five layers specified to the point an engineering team could build them. Most of the effort goes into the rules engine, because everything downstream rests on it.
TestingThe runs
A baseline, then three to five interventions, then a stress test to find what breaks. The path worth taking falls out of the comparison.
AccountabilityThe ledger
Every call recorded with its date, its confidence band, and what actually happened. This stage never closes. It’s the only honest answer to how well the model reads.
STRATA is the platform. It runs on BASIN, the cognition engine licensed from Worldlines. BASIN makes a simulated world run; STRATA makes it answer the question you actually have — about a product you haven’t launched, a market that hasn’t responded yet, a change of direction you can’t test any other way. STRATA is what you get access to, and it does not stop running when a project would have ended.
Six kinds of world.
Translation ends by naming which of these you have. Open any one — this is the same platform running in six different sectors, with real controls and live output.
Customer journey
Sector
Consumer product
A household care brand, $180m at retail, sold through grocery and its own direct channel.
The question
Trial sells brilliantly. The second purchase does not. Where are we losing them?
The simulated solution
Ten thousand buyers are dropped into a purchase vortex and followed for ninety days. Three gates: they try it, they use it properly, they come back. The width of each ring is the live pass rate — what falls past a ring is lost for good.
Launch discount
How much we take off to win the first purchase.
After they buy
Whether anyone shows a new buyer how to use it properly.
—Repeat rate
—Annual revenue from new buyers
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Market entry
Sector
Film
A studio franchise opening in a territory where the studio has never distributed. Eleven cities carry 70% of the box office.
The question
What actually happens when we open here — and does opening wide help, or exhaust the audience before the run begins?
The simulated solution
The territory as a rotating world. Every marker is a city, its column height is admissions that week. Release ignites the map, then word of mouth carries it or it goes dark. Twelve weeks run in front of you.
How we release it
The fans already following it
—Opening week admissions
—Twelve-week box office
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Product development
Sector
Creator economy
A creator platform with 30,000 active creators, currently free, considering a paid tier.
The question
Will the creators this was built for actually pay — and does the price decide it?
The simulated solution
The creator base as a three-shell network: the largest names at the core, working mid-tier around them, emerging creators on the outside. Adoption can only travel along a connection. Twelve months of spread, one month per beat.
Monthly price
What a creator pays to be on the paid tier.
Who we let in first
Proof only travels between creators who can see each other’s economics.
—Paying creators at month 12
—Annual recurring revenue
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Competitive war game
Sector
Sport
A rights holder launching a new competition format next season. Three growth markets carry $84m a year in combined broadcast and sponsorship.
The question
Which rival response actually costs us money — and can we stop it before it happens?
The simulated solution
The season calendar as a board. Three growth markets, each carrying its own rights value, fourteen weeks across. We place our fixture, three rivals answer with their own decision logic, and their reply lands on the board. Whatever it knocks out is money off the table — the running total on the left is the answer.
What we announce
Our slot in the calendar
Rivals answer with whatever costs them least.
—Rights value at risk
—What that reply costs the rival
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Cultural trajectory
Sector
Intellectual property
A character IP, forty years old, licensed across nine categories and still commercially healthy.
The question
It performs today. Is anyone under thirty going to bring it up in five years?
The simulated solution
The character at the centre, four audience cohorts in orbit around it. Distance is indifference: as affection decays a cohort drifts outward and dims. The inheritors are the outermost shell — the ones who decide whether this survives.
What we spend the next five years doing
Combine any of them. Cost markers are relative.
—Licensing revenue, year five
—Affection index, 61 today
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Crisis and resilience
Sector
Brand
A premium consumer brand across four channels and three loyalty segments.
The question
What actually breaks us — and which of those do we not recover from?
The simulated solution
Brand health as terrain. Every peak is a segment on a channel; height is how healthy it is. A shock punches a crater into the surface and the damage spreads outward before it heals. Eighteen months run on a loop.
What goes wrong
When someone escalates it
—Peak revenue lost
—Revenue at month 18
What the model returns
—
Illustrative model. Behaviour is representative, figures are not client data.
Simulations under build.
Instances being built now. Partners are named under NDA and their outputs stay with them — what follows is the shape of the deployment.
The Pit WallGlobal motorsport · in build with a media partner
827MPeople who follow Formula 1 worldwidecited — F1 / Nielsen, 2025
438MEngaged, but no brand is tending to themcited — F1 / Nielsen, 2025
100k+Respondents behind the fan modelcited — Global F1 Fan Survey, 2025
186Countries covered by that surveycited — Global F1 Fan Survey, 2025
Can your business be modelled?
This is the readiness score from stage three of STRATA, in short form. Six questions, ninety seconds, and a straight answer, including the one where the answer is not yet.
1. How well documented is the way your brand actually behaves?
2. How many real decisions do your customers make about you?
3. How much of their behaviour can you actually see?
4. How well do you understand what drives your market?
5. How many revenue levers could be modelled?
6. How much rests on the decision in front of you?
The questions we get asked.
The five that come up every time. Answered the way we would answer them on a call.
A platform. You get an instance of your own market: it runs, your team operates it, and it is still there when the next decision arrives. What you acquire is standing capability rather than a finite piece of work — and the recommendations on this page are what the model returns, not what an adviser thinks.
Partly, and that is why the first three stages exist. Assumptions are drawn out of your operating leaders in structured intake, written down in plain English, and scored before anything is built. If they are weak we say so and stop. What we will not do is bury them inside a model and present the output as fact.
No. We work from what your operating leaders know plus whatever behavioural data already exists — no warehouse, no migration, no year of instrumentation. Individual customer data is almost never needed; models run on segments and behaviour patterns.
You own the world built from your business — the design, the rules, the intelligence it produces. A live instance is operated by your team: you change an assumption and re-run it on a Tuesday afternoon without booking anyone. We own STRATA, the layer that makes worlds out of businesses.
Mapping runs four to six weeks and ends with a straight answer about whether your business can be modelled. That answer is useful whichever way it goes — and if the readiness score comes back low, we tell you and stop.
Anything beyond these is specific to your business, and better answered by a person than by a page. Speak to our team.
Built by.
A simulation is only as credible as the people who specified it. Our Religion is a technology platform built by six disciplines working on the same model, because a world specified by any one of them alone produces something confident and wrong.
Data Scientists
Behavioural data, segmentation and the empirical work that grounds a world in what has actually happened. Where the evidence is thin, they are the ones who say so before it hardens into an assumption.
Mathematicians
Calibration, confidence intervals and the honest limits of inference. They are the reason the Ledger records a confidence tier against every call rather than a single claim of accuracy.
AI Engineers
Agent architecture, decision logic and the systems that allow thousands of modelled actors to behave independently rather than collapse into an average.
Product and Marketing Experts
How demand actually behaves — trial, repeat, media response, channel effects — and how an operator uses a model once it exists. A world nobody can operate is a research paper.
Culture Forward Creatives
Practitioners who read where meaning is moving before it appears in the data. Cultural drift is the variable most models omit entirely, which is precisely why brands are surprised by it.
Entrepreneurs and Operators
People who have run the businesses we model — sport, entertainment, consumer, intellectual property. Structured intake yields a usable specification only when the person asking has done the job.
How you get on the platform.
STRATA is how the platform works. This is how you get onto it — three stages, each one a decision point. You can stop after any of them, and nothing is committed until the stage before it has earned it.
MappingIntake · Translation · Readiness
Your business goes into the pipeline and comes out scored for whether it can be modelled at all. Ends with a straight answer either way.
4–6 weeks
World buildArchitecture
Your world is specified and built — landscape, actors, rules, signals, clock — to the point it can be stood up and run.
6–10 weeks
Live instanceTesting · Accountability
Your world goes live and stays live. Your team runs scenarios against it whenever a decision comes up, and the Ledger opens here and never closes.
12–16 weeksthen the Ledger stays open
Access is priced per stage. A live instance is not a project that finishes — it is a world that keeps running, which is the difference between buying an answer and having somewhere to put the next question.
Bring us a decision you’re afraid of.
The ones worth simulating are usually the ones that keep getting deferred because nobody can evidence them. Tell us what is on the table and we will tell you, straight, whether it can be modelled.