Can London turn startup density into durable scale

This draft cannot yet answer whether London turns startup density into durable scale because the planned company cohort has not been built or frozen. It sets out the method needed to answer that question. A positive result would require a meaningful share of young companies to progress from formation and funding to customers, employment and continued operation. Office clusters, accelerator places and financing rounds show activity, but they do not answer that harder question.

London has genuine advantages in finance, talent, research, regulation and customer access. It also has high operating costs, competitive hiring and difficult enterprise procurement. A serious assessment should show which types of company convert those conditions into durable businesses, which move functions elsewhere and which cease operating. A single ecosystem valuation cannot do that work.

Startup density is an input rather than an outcome

A dense market can improve the chances of finding a co-founder, specialist employee, investor, adviser or early customer. It can also produce repeated companies chasing the same talent and capital. The commercial question is whether proximity changes formation and growth enough to outweigh cost and competition.

Published counts often use different populations. The City of London technology SME report examines small and medium-sized technology enterprises in the Square Mile. It should not be described as a count for Greater London. Funding databases may include venture-backed startups across a metropolitan area. Companies House includes legal entities with very different levels of activity.

The first discipline is therefore to state what is being counted. A registered company, an employer, an authorised firm and an operating product are not interchangeable. Density should be reported using a defined geography, date and business population, then connected cautiously to later outcomes.

A cohort reveals progress that a snapshot hides

Shoreditch Talk should define a cohort of London technology and fintech companies founded during a fixed period, such as 2018 to 2021, and follow the same companies through a common observation date. Entry rules would require evidence of a meaningful London operating presence, a technology-led product and an active company at the baseline. The final sample should be frozen before outcome analysis.

The Office for National Statistics business demography methodology tracks births, deaths and survival for businesses registered for VAT or PAYE. It provides a sound reference for understanding official survival measures, but an editorial technology cohort may use narrower inclusion rules. Results from the two populations should not be merged.

Each cohort company needs a persistent identifier so name changes, group structures and acquisitions do not create duplicates. Registered office, operating office and headquarters should be separate fields. A company that remains legally active but has no visible staff or product should not automatically count as a durable operating business.

Durable scale needs more than one measure

Survival is necessary but weak on its own. A dormant or minimal company can remain on the register. The cohort should therefore record continuing operation, employment, revenue evidence, customer adoption, regulatory status, follow-on finance and exit. No single measure needs to be available for every private company, but the evidence state must be visible.

Companies House can provide accounts, confirmation statements, officers and legal status. Smaller companies may file accounts with limited detail, and group revenue may not reveal a London operation. The Financial Services Register can show a firm’s permission and status, but authorisation does not prove customer demand or profitability.

Employment can be estimated from filed accounts, official datasets, company disclosures and professionally collected workforce data, with the source identified. Customer evidence should favour named and continuing use over logo displays. Funding should be divided by stage and type. Acquisitions should record whether the product and team continued rather than treating every sale as an equal success.

London’s contribution should be tested directly

A company’s outcome does not show what London caused. Founder interviews should ask which customers, hires, investors, advisers or institutions were accessed through the city, and which functions were difficult or expensive to keep there. The same questions should be asked of firms that stalled, moved or closed.

Location histories matter. A London-founded company may build engineering elsewhere, open regulated entities in another market or move its headquarters while retaining sales in the city. That is not automatically a failure for London, but it changes the claim. The analysis should show which work the city attracts and which work it retains at different stages.

Matched comparison cohorts in New York, Paris and Berlin could test whether outcomes differ, provided geography, founding years and business definitions are aligned. They cannot turn an observational study into proof of causation. Differences in disclosure, accounting, regulation and funding coverage need to appear beside the result.

Policy should be judged against the bottleneck it targets

An accelerator may improve founder networks without changing later growth capital. A regulatory programme may clarify permission without creating customers. A talent intervention may help one specialist role while housing and salary pressures persist. Programme evaluation should begin with the stated bottleneck and a plausible route to an outcome.

The cohort can help identify where companies stop progressing. If technically credible firms repeatedly fail to convert bank pilots, buyer procurement deserves attention. If follow-on capital is available but senior compliance hiring delays expansion, the labour market matters. If firms remain small by choice and profitable, a growth-only framework may misclassify a durable outcome.

London should not be judged solely by how many companies become very large. Sustainable specialist firms, infrastructure providers and profitable service businesses can create value without venture-scale growth. The report should segment the intended model before comparing outcomes.

A public scorecard should preserve the difficult cases

The strongest output would be a company-level evidence table with clear privacy and fairness rules, plus aggregate measures by founding year and activity. Missing fields should be reported, and companies should have a route to correct facts. Changes to the cohort or coding should appear in a public revision log.

The scorecard should use counts and distributions rather than a composite rank. Combining survival, funding, jobs and authorisation into one number would conceal trade-offs and invite arbitrary weights. Readers need to see whether a conclusion depends on a small group of outliers.

Annual updates should retain closed, acquired and relocated companies. Replacing them with new entrants would turn a cohort into a promotional directory. The method should also record the observation window, because a company founded in 2021 has had less time to scale than one founded in 2018.

Limits of the available evidence

Private-company revenue, customer relationships and employee locations are often incomplete. Companies with funding or strong communications leave more evidence than quieter firms. Legal survival can overstate operating survival, while a dissolution can follow a successful asset sale. Cross-city datasets may use inconsistent sector definitions.

This article therefore sets out the test rather than claiming a result. It should not be presented as an answer to the headline until Shoreditch Talk has built and frozen the cohort, published its inclusion rules and observation date, verified each company and reported missing data. If it is published before that work, it must be labelled as a research-method proposal. The question is not whether London produces many startups. It is how many defined companies build durable operations, what kind of scale they reach and which parts of that journey the city actually supports.