It sits on top
Participants keep the systems, vendors and reporting obligations they have today. A copy of an agreed set of information flows through one simple connection. The hard work happens above their operations, not inside them.
Data flywheels, the insight layer and the platform
Data that gets smarter every revolution.
A data flywheel is a self-reinforcing system where data usage generates better data, which enables better decisions, which creates more data. This page covers the whole of it: the loop, the shared insight layer that runs it across a market, and the federated platform underneath.
Insight
Collect
Enrich
Apply
Learn
Compound
What it is
Data that compounds. Like interest.
Most organisations treat data as a byproduct, something to store and occasionally report on. A data flywheel is the opposite: a deliberately designed system where every interaction, decision, and output makes the system smarter and more valuable.
The mechanism is self-reinforcing. Better data leads to better models. Better models lead to better decisions. Better decisions generate more and higher-quality data.
We've built data flywheels for e-commerce platforms that personalise at scale, for B2B businesses that predict churn before it happens, for logistics operators that reduce waste through feedback loops, and for SaaS companies that make their product smarter with every user session.
The result isn't just better analytics. It's a structural competitive advantage that grows harder to replicate over time. The flywheel, once spinning, becomes a moat.
How it works
Four stages. One perpetual loop.
01
Collect
Instrument your data sources
We audit existing data assets and instrument new collection points, capturing the signals that matter, cleanly and consistently.
02
Enrich
Build models that learn
Raw data becomes intelligence. We design enrichment pipelines and ML models that improve with each new data point.
03
Apply
Activate insight in workflows
Insights surface where decisions are made: in products, dashboards, and recommendations that act, not reports that sit unread.
04
Learn
Feed outcomes back in
Every decision outcome loops back into the system, refining models, improving collection, and tightening the flywheel with every revolution.
Run inside one company, the loop compounds one company's advantage. Run across a whole market, it becomes the insight layer.
Flagship: the insight layer
Shared intelligence from the data your market already collects.
Every organisation in a regulated market already collects the information that matters. Each collects it differently and keeps it to itself, so nobody can see the market as a whole. Flywheel adds an insight layer above the systems operators already run. No one replaces anything. One connection each, a complete view for the regulator, and useful intelligence back for every participant.
The problem
The whole is invisible.
Picture a few dozen operators answering to the same oversight body. All track broadly the same things, but each defines them its own way, reports on its own schedule, in its own format.
So a pattern spread across two or three operators looks like nothing to each of them individually. The oversight body receives fragments that arrive late and resist comparison. The aggregate view, the only view where the real signals live, is exactly the view nobody has.
Nobody is at fault. Any market whose participants built their systems independently ends up this way. Fixing it takes a layer above them all that speaks one language.
Operator A · fragmentlate · incomparable
Operator B · fragmentlate · incomparable
Operator C · fragmentlate · incomparable
? the aggregate view nobody has ?
How the layer works
One connection. One language. Nothing replaced.
Each participant makes a single connection, comparable to integrations it already maintains. Above the market's existing operations, information is standardised into one set of definitions, and the right view flows back to the right party. The data itself stays where it is.
The market today
Operator A
CSV · monthly
Operator B
XML · quarterly
Operator C
JSON · weekly
Operator D
own schema
Same questions · different definitions · fragments, late
The Flywheel Insight Layer
Sits above, replaces nothing
One connectionOne definitionAnonymised outputs
What comes back
The oversight body
A live, complete view of the whole market, and the evidence its mandate is met.
Each participant
Its own detail set in anonymous market context, and never another participant's data.
Research & public
Population-level knowledge, aggregated and anonymised before it leaves the platform.
Three commitments define it
What the layer promises.
One definition for everyone
Standardisation is the unglamorous source of most of the value. Once every participant's information speaks a common language, the same measure means the same thing everywhere, and comparison becomes possible. No participant can create that alone; no off-the-shelf product supplies it.
Everyone sees what they should, nothing more
The oversight body sees the whole market. Each participant sees its own detail alongside anonymous market context, never another participant's data. Anything released further is aggregated and anonymised before it leaves the platform.
Why combined is worth more
The value compounds. That is the flywheel.
One organisation's data answers operational questions about its own platform. Standardised and combined, the same data answers better ones: where each participant actually stands against the real market, and what is building that no one inside a single company can see yet.
Each participant that connects makes the shared picture sharper for everyone already in it. The more who take part, the more each of them gains.
participants in · picture sharpens
What each party gets
Every seat at the layer sees value.
The oversight body
- A live, complete view of the market it is responsible for
- Consistent, whole-of-market evidence behind every claim
- Signals that only exist in aggregate, found earlier
Participants
- Insight as a service: their own picture, in market context
- Early signals no single operator could generate alone
- A simpler reporting relationship: information flows once
Researchers & public agencies
- Anonymised, population-level knowledge in consistent form
- Evidence for policy built on a shared standard
- Access that improves as the market grows
Where it starts
Player wellbeing: the application where seeing the whole matters most.
Signs of harm pay no attention to the boundaries of a single platform. A pattern that looks minor to each operator can be unmistakable across the market. With a shared layer, the people in a position to help find out earlier, and the market can back its commitments with evidence instead of assurances.
None of this adds work for the operators themselves. Once the layer is in place it extends naturally: to new insight areas for the same participants, and to other jurisdictions ready to adopt a working standard.
Engagement tiers
Three shapes of engagement. Every one priced to the market it serves.
The layer is a managed service, not an infrastructure project, so cost tracks the number of participants, the depth of the insight set and the regulatory obligations in play. That makes every engagement bespoke. Tell us the shape of your market and we will come back with a firm number.
Tier 01
Foundation pilot
A small founding group of operators proving the layer on one insight area.
- Connection built for each founding participant
- One standardised insight set, agreed together
- Private market context returned to every participant
- Whole-market view for the oversight body
- Fixed pilot window with a clear success measure
Bespoke
Scoped on enquiry
Tier 02
Market programme
A jurisdiction moving from pilot to standing operation across most of the market.
- Onboarding for the wider operator base
- Expanded insight sets as coverage grows
- Reporting cadence aligned to regulatory obligations
- Anonymised aggregate access for approved research
- Ongoing stewardship of definitions and standards
Bespoke
Scoped on enquiry
Tier 03
Jurisdiction platform
Regulators and market bodies adopting the layer as durable public infrastructure.
- Multi-jurisdiction deployment on a shared standard
- Governance model, audit trail and access controls
- Custom insight areas beyond the initial application
- Integration with existing supervisory reporting
- Named team and defined service commitments
Bespoke
Scoped on enquiry
No published price list. Pricing depends on participant count, data volume, jurisdiction and reporting obligations, and we would rather quote something accurate than something generic.
How to get started
From first conversation to first signal.
Introductory call
30 minutes
We listen to how your market is structured, who would participate and what question you most need answered.
Scoping and pricing
1 to 2 weeks
We map participants, data availability and obligations, then return a written scope with firm pricing for your situation.
Founding group connected
Weeks, not quarters
Each participant makes one straightforward connection. Nothing changes in their systems and no new work lands on their teams.
First signal live
Ongoing
The layer starts returning market context and whole-market signal, then sharpens with every participant that joins.
The platform
Discharge the obligation. Keep the asset it produces.
A federated data platform for regulated markets. Four components sit inside each member company, an open protocol layer connects them, and a cooperative network turns the evidence they produce into shared intelligence.
Federated by designMember-set trustBuilt for regulated verticals
Why now: the agentic shift behind the platform →The platform is delivered as Attesia, the product arm of Flywheel Ventures Group: three products that turn obligation into evidence and evidence into an asset.
Explore Attesia →Alongside it, the group's M&A advisory arm helps partner companies build and scale inside the ecosystem and advises on transactions from an operator's seat.
Explore M&A Advisory →And the Build & Scale arm co-builds data products and commercial infrastructure alongside portfolio and partner companies, from dormant asset to live revenue line.
Explore Build & Scale →Architecture
Four components per member, one protocol between them.
Stage one runs entirely inside your business: your data never leaves until you decide it should. Stage two connects you to the Commons, where the same components speak a shared language with everyone else in the market.
Stage 01 · Federation, inside the member company
Human interface
Portal
Dashboards, question and answer, reports
Autonomous reasoning
Agent
AI fine-tuned on the member's own operation
Trust enforcement
Gateway
Policy, permissions and a full audit trail
Connectors
Data sources
ERP, CRM, DMS and custom connectors
Protocol layer · agent-native
A semantic and ontology layer acts as the universal translator between members. Trust is negotiated at machine speed, inside bounds the member pre-sets, and revenue settles programmatically.
MCPA2AOntology
Stage 02 · Commons, the cooperative network
Envoy
Member identity in the network, agent-native
Collective Intelligence Engine
Benchmarks, peer signals, trend detection
Data Marketplace
Consented, anonymised, revenue shared back to members
Target architecture. Deployment model, anonymisation standard and data-rights framework are decided with the founding members.
The trust spectrum
Every dataset sits at a trust level you choose.
Trust modes are set per dataset and can be mixed. Nothing changes without your explicit configuration, and every data point carries a machine-readable consent record from the moment it is ingested.
Mode
Exposure
Typical use
Poll
Lowest
Industry sentiment, yes-no benchmarks
Q&A
Low to medium
Directional intelligence, capability assessment
Query
Medium to high
Benchmark analytics, peer comparisons
Replica
Highest
Deep research, longitudinal analysis
Data dump
Situational
Regulatory reporting, academic research
The cooperative principle
Every data broker keeps the proceeds. We share them back.
Members contribute data they control, receive collective intelligence in return, and earn a revenue share when their consented data creates value for external buyers. The revenue flows back to the members who created it. That difference is the whole point of the cooperative model.
The engagement model
Four phases, from first paid engagement to membership.
You can stop after any phase and still hold something useful. Each one is priced on its own and produces an asset, not a recommendation.
Assessment
Health, Opportunity, Appraisal
A CEO-level paid engagement that produces a scored picture of data-estate quality, regulatory exposure and what the asset is worth to outside buyers. You pay for it whether or not it proceeds, and you keep the finding either way.
Roadmap
Swamp-to-lake architecture plan
Short-term remediation executed by us, long-term architecture contracted alongside it. Every engagement is required to leave behind a reusable platform asset rather than a report.
Federation
Agent and Gateway deployed
A managed service inside your four walls. We own and maintain the connectors into your source systems, so the integration burden stays with us rather than your team.
Commons
Join the industry network
Trust level set per dataset, on your terms. A platform fee plus a share of marketplace revenue generated by data you have explicitly consented to release.
Inside the Assessment
One pass, three sections, one methodology.
The same assessment methodology runs wherever we price a data asset, including the readiness scoring inside the Build & Scale pipeline. Weights are vertical-specific and published, not hidden.
Run the assessmentHealth
Scored
How good the data actually is: completeness, validity, consistency, coverage, freshness, provenance, resilience. Scored per dimension against published, vertical-specific weights, and aggregated so a single fatal weakness cannot hide behind strong dimensions.
Opportunity
Narrative, conditional
The case for joining the Commons, built from your data's fit with the network. It appears only where the fit genuinely qualifies. Where it does not, we say so and the report runs Health and Appraisal only.
Appraisal
Dollar range
What the data is worth to outside buyers as a saleable asset: licensing surface, demand signal, scarcity, linkability, historical depth, jurisdictional limits. Delivered as a range with a central estimate, framed as an asset valuation rather than a revenue forecast.
We build first where the regulatory duty is real, the data is rich, and the operators already know each other. iGaming is one such vertical, and a partner channel for the group. The semantic layer built in a market like that becomes the pattern for every regulated market that follows.
Where we've built
Industries we've deployed flywheels in.
01
E-commerce & Retail
Personalisation engines, demand forecasting, and customer LTV flywheels that improve with every transaction.
02
B2B SaaS
Churn prediction, product usage intelligence, and expansion revenue signals feeding back into a smarter product.
03
Financial Services
Risk scoring, fraud detection, and client behaviour models that get more accurate with every data point.
04
Operations & Logistics
Route optimisation, waste reduction, and supplier intelligence loops that compound efficiency with every cycle.
05
Media & Content
Engagement flywheels that learn what resonates and feed editorial intelligence back upstream.
06
Professional Services
Client intelligence systems and knowledge flywheels that make your firm smarter with every engagement.
Where we are today
A plain statement of stage.
What exists
Active AI-training and data engagements with clients who have since surfaced data-monetisation problems of their own.
What comes next
A beta placed inside a regulated vertical, with clients who already trust us.
How we work
Senior people only, on the architecture already specified. Every engagement leaves a reusable asset behind.
Ready to build?
The sooner the flywheel starts spinning, the faster it compounds.
We start with a data audit: a clear-eyed view of what you have, what you're missing, and what a flywheel could unlock for your business.