Counting direct vendors does not reveal cloud concentration. A fintech must map important services through software providers to hosting regions, identity, DNS, observability, model and data dependencies; identify where supposedly separate suppliers share failure modes; and test whether substitutes and recovery tools remain available during a correlated disruption.
- Key takeawaysMap important business services first, then trace direct and indirect dependencies.
- Decision frameworkA fintech has concentrated supplier risk when several essential services depend on the same underlying infrastructure, organisation or access mechanism.
- Map services before counting vendorsStart with a service customers use: logging in, paying, receiving money or obtaining help.
- Distinguish shared ownership from shared failureTwo products supplied by one group may have genuinely different technical arrangements.
- Test a correlated scenarioImagine the identity provider fails while the support platform and cloud administration both require that identity.
Answer first: Counting direct vendors does not reveal cloud concentration. A fintech must map important services through software providers to hosting regions, identity, DNS, observability, model and data dependencies; identify where supposedly separate suppliers share failure modes; and test whether substitutes and recovery tools remain available during a correlated disruption.
Key takeaways
- Map important business services first, then trace direct and indirect dependencies.
- Treat independent contracts and independent failure modes as different questions.
- Test correlated cloud, identity, DNS and support-channel disruption.
- Prioritise controls where substitution is both difficult and slow.
Decision framework
| Decision area | Evidence to require | Risk it controls |
|---|---|---|
| Important service | Customer outcome and impact tolerance | Wrong resilience target |
| Direct supplier | Contract, service and data dependence | Unowned third party |
| Fourth party | Cloud/identity/DNS/data/model dependency | Hidden common mode |
| Substitution | Time, skills, data and commercial rights | Paper alternative |
| Recovery stack | Recovery tools do not share failed dependency | Circular recovery plan |
A fintech has concentrated supplier risk when several essential services depend on the same underlying infrastructure, organisation or access mechanism. Buying from two vendors does not create meaningful resilience if both require the same cloud region, identity provider or telecommunications route.
The right response is not automatically to duplicate every system. It is to identify correlated failure, understand the customer services affected and invest in the dependencies whose loss would cause the greatest harm.
Map services before counting vendors
Start with a service customers use: logging in, paying, receiving money or obtaining help. List the systems and suppliers required at each step. Then ask those suppliers about their own material dependencies.
The NCSC defines supply-chain mapping as recording and using information about the suppliers involved in a company's chain. Its cloud guidance also makes clear that managed service providers and their chains form part of the customer's supply chain. 12 The map should include cloud hosting, identity, domain-name services, messaging, fraud tools, observability and staff access. An overlooked administrative dependency can disable the people responsible for recovery even when the production platform remains available.
Distinguish shared ownership from shared failure
Two products supplied by one group may have genuinely different technical arrangements.
Two independent companies may use the same region and authentication service. Record the evidence rather than assigning a risk score from brand names.
Ask what is shared: physical facilities, control plane, network, software component, support team or legal entity. A provider's multi-region claim may protect against a regional disruption while leaving a common control-plane failure unresolved.
Do not demand an exhaustive public infrastructure diagram. Seek the information necessary to assess the service, with appropriate confidentiality. If a supplier cannot provide it, record the uncertainty and decide what compensating test or contract protection is possible.
Test a correlated scenario
Imagine the identity provider fails while the support platform and cloud administration both require that identity. Can staff access an emergency route? Can customers receive updates through a channel that does not share the same dependency?
Or assume both the primary payment tool and backup provider use the same underlying connection. The backup contract may still be useful for ordinary provider failure, but it will not solve this particular event.
Document what continues, what stops and which manual actions are feasible at realistic volume. A manual process that handles ten payments is not a credible replacement for a service processing thousands unless the plan explicitly prioritises a smaller essential population.
Understand what the CTP regime does
The UK's Critical Third Party framework addresses systemic risks from designated providers.
The Bank of England explains the designation process and the role of the regulators. 3 It does not mean every important fintech supplier is a designated CTP, nor does designation remove a firm's own dependency-management work.
Keep three concepts separate: your business's critical supplier, an arrangement material under applicable rules and a provider formally designated under the CTP regime. They may overlap, but they are not interchangeable labels.
For a small supplier selling into regulated institutions, clear dependency evidence can be commercially valuable. It helps the buyer assess the arrangement without requiring the startup to claim it has the resilience of a global infrastructure provider.
Spend where substitution is hardest
Prioritise dependencies by service impact, common exposure, recovery time and credible alternatives. Some risks are best addressed through tested failover. Others need preserved data, emergency access, an alternative communication route or a realistic exit plan.
An illustrative scoring exercise can compare a low-impact analytics service with a unique payments connection. The analytics vendor may have more users internally, yet the payments connection can be harder to replace and more consequential for customers. Avoid letting contract value alone set the priority.
Find the recovery dependency inside the dependency map
Ask what staff need to restore a service: administrator access, documentation, source repositories, communication and provider support. These may share infrastructure with the failed service.
In a hypothetical cloud incident, the production system, engineering documentation and emergency contact list all sit behind the same unavailable identity platform. The company owns a recovery plan but cannot reach it. An appropriately protected alternative access and information route may improve resilience more than duplicating a minor application.
Test those arrangements rather than assuming emergency credentials will work. Review access, storage and accountability so the alternative does not create an unmanaged security weakness.
Decide how much uncertainty is acceptable
Supplier information will rarely be complete. Record the unknown dependency, potential impact and evidence sought. Distinguish an unknown that affects a critical service from one with little practical consequence.
Management can then decide whether to accept the exposure, obtain more assurance, reduce dependence or prepare an alternative. A numerical score without that decision can make uncertainty look more precise than it is.
Review the choice after material changes in volume, regulation or customer use. A supplier supporting an experiment may become indispensable without a new procurement event. The map needs to recognise that change before a disruption demonstrates it. Concentration management is a repeated judgement about real services, not an annual count of vendor names.
Make the map a maintained record
Update it when a supplier changes hosting, acquires a competitor or introduces a new subprocessor. Assign owners and evidence dates. An unanswered questionnaire from last year's procurement is not a current dependency assessment.
Keep a short, maintained list of shared failure points and the actions tested against them. If the dependency map does not change a recovery plan, supplier decision or access arrangement, it has yet to earn the time spent maintaining it.
A practical 90-day implementation plan
- Choose the three important services with the greatest customer or market impact.
- Ask direct suppliers for named critical sub-processors and region/service dependencies.
- Overlay identity, DNS, observability, communications and backup paths across the map.
- Run one correlated scenario and fund the highest-impact substitution or recovery gap.
Frequently asked questions
What is a fourth party?
It is a supplier used by your direct supplier. The term matters because your service can depend on it even without a direct contract.
Does regulatory oversight of a Critical Third Party transfer accountability?
No. The regime adds oversight of designated providers; a regulated firm still needs to understand and manage the resilience of its own services and arrangements.
Is multi-vendor procurement enough?
No. Vendors may share the same cloud region, identity provider, network, support channel or data service, leaving correlated failure unchanged.
Reporting by Shoreditch Talk




