A fintech pilot is evidence that somebody is willing to test a proposition. It is not evidence of repeatable revenue, regulatory readiness or a durable customer relationship. Shoreditch Talk will follow a defined cohort of 25 London fintech companies from a public pilot, sandbox place or first commercial announcement to later evidence of authorisation, paying customers and continued operation.

The study will answer a practical question: which conditions help an early institutional experiment become contracted business, and where does progress stop? It will separate public activity from commercial outcomes and report missing evidence instead of treating silence as success or failure. The finished work will be a cohort study, not a ranking and not a promotion for innovation programmes.

The cohort will be fixed before outcomes are assessed

Selection after the results are known would create a misleading sample. The study will therefore publish its inclusion rules and freeze the cohort before detailed outcome coding begins. Eligible companies must have had a material London operating presence and a public pilot, regulatory sandbox participation or named first-customer announcement during a stated entry period. The announcement must describe an actual test or commercial relationship rather than a general partnership.

The cohort will contain 25 companies because that size permits careful documentary review and direct approaches while still exposing variation. It will not represent every London fintech. The sample will be stratified across activities such as payments, credit, insurance, compliance, data and market infrastructure where enough qualifying cases exist. Companies controlled by the same group will be treated carefully to avoid duplicate outcomes.

The FCA Regulatory Sandbox provides one possible entry source, but sandbox participation will not dominate the sample. Other firms may first appear through a bank pilot, an accelerator demonstration or a named procurement announcement. Each entry event will have a source, date and description of what was actually tested.

Commercial progress requires several distinct measures

The primary outcome is evidence of commercial revenue from the relevant product. Public accounts rarely disclose revenue by pilot, and private contracts are often confidential, so the study will use an evidence hierarchy. Strong evidence includes a customer confirmation, an executed contract shared on a confidential basis, revenue clearly linked to the product in filed accounts, or a named case study with continuing paid use. A supplier announcement without customer confirmation will be coded as an unverified claim.

Other outcomes will be recorded separately. The Financial Services Register can show permissions and status, but authorisation is not revenue. Companies House records can show legal status and filed accounts, but group structures and reporting exemptions can limit interpretation. Follow-on funding may support continued development, but it does not prove a buyer adopted the product.

The study will therefore code five stages: test announced, test completed, relevant permission obtained where required, paid deployment evidenced, and continuing operation evidenced at the review date. A company can progress through some stages in a different order. Products that do not require FCA authorisation will be marked accordingly rather than treated as missing a milestone.

Timelines will expose the work between interest and revenue

For each company, the research will reconstruct a dated sequence from entry event to the latest verified outcome. Interviews will ask founders and buyers about procurement, information security, integration, model validation, legal review, budget ownership and the decision to stop or expand. The same core questions will be used across cases so that one unusually open founder does not define the entire conclusion.

Elapsed time will be reported as a range when exact dates are confidential. The clock will start from a clearly defined public or documented event, not from a founder’s later recollection of when discussions began. Where a pilot involved more than one buyer, each relationship will be separated.

The analysis will look for recurring constraints rather than a single recipe. A regulated permission may be the main dependency for one firm, while a bank’s integration backlog may matter more for another. Some pilots are designed to answer a technical question and were never intended to become a contract. Those cases will be classified by their stated purpose before outcomes are judged.

Missing outcomes will remain visible

Public programmes often announce entrants more consistently than later outcomes. The study will not fill those gaps with assumptions. Each field will carry one of four evidence states: confirmed, supported by one party, contradicted, or not publicly established. Companies will be invited to correct factual records and provide evidence, but participation will not be required for inclusion.

No response will be reported as no response, not as commercial failure. Closure or dissolution will be checked against formal records, but a dormant legal entity may coexist with intellectual property sold elsewhere. Acquisitions, pivots and product withdrawals will receive narrative notes because a binary survival field can hide what happened.

Material criticism will be put to the relevant company or programme before publication. Confidential information will inform a coded outcome only when the team can agree a form of evidence and disclosure that readers can understand. Anonymous assertions that cannot be tested will not carry the same weight as filed or attributable evidence.

The report will connect policy activity to buyer decisions

The final publication will include the cohort definition, a company-level evidence table, aggregate stage counts, a timeline analysis and several documented case studies. It will show how many cases have public evidence at each stage without presenting the proportions as representative of all London fintechs. The downloadable data will state the source and evidence date for every non-confidential field.

The most useful findings will concern decisions. Founders need to know which buyer, regulatory and technical requirements appear before revenue. Buyers need to understand where pilot design creates false momentum. Programme operators need to see whether participation records can be connected to later outcomes. Investors need a clearer distinction between access, permission and repeatable demand.

The FCA Innovation Insights 2025 report offers useful context about how firms use the regulator’s innovation services, but this project will not copy its population or claim to evaluate the FCA itself. Its unit of analysis is the company journey from a named entry event to commercial evidence.

Limits of the cohort study

The sample will be small and deliberately selected, so it cannot produce a London-wide conversion rate. Public information is richer for successful or well-funded companies, which can bias observed outcomes. Contract value, procurement duration and product-level revenue may remain confidential. Company structures and regulatory status can also change after the evidence date.

The research can identify patterns and test specific explanations. It cannot prove that a sandbox, pilot or location caused commercial progress. Any comparison between subgroups will show case counts and uncertainty rather than imply statistical precision. A later edition can extend the observation period, but it must preserve the original cohort to avoid replacing difficult cases with newer success stories.

Reporting by Shoreditch Talk