No defensible answer can come from startup counts alone. A city region turns business births into durable scale when new firms survive, add employment or turnover, and continue growing without the result being explained mainly by sector mix, registered-office effects or a few exceptional companies. The scorecard must use consistent geographies, publish its calculations and show how results change when the weights change.
- OverviewNo defensible answer can come from startup counts alone.
- Define every outcome before collecting dataThe Office for National Statistics defines business births and deaths through the population of active enterprises on the Inter-Departmental Business Register.
- Use city-region geographies consistentlyLocal authorities, combined authorities, travel-to-work areas and International Territorial Levels describe different things.
- Build the score from a sequence, not a snapshotThe first measure is formation relative to the active business base.
- Adjust for sector and economic structureProfessional services, hospitality, construction, manufacturing and life sciences have different formation, capital and employment patterns.
No defensible answer can come from startup counts alone. A city region turns business births into durable scale when new firms survive, add employment or turnover, and continue growing without the result being explained mainly by sector mix, registered-office effects or a few exceptional companies. The scorecard must use consistent geographies, publish its calculations and show how results change when the weights change.
This article sets the method Shoreditch Talk will use. It does not publish an illustrative ranking as a real finding. The final table should appear only after the data, matching and independent review are complete.
Define every outcome before collecting data
The Office for National Statistics defines business births and deaths through the population of active enterprises on the Inter-Departmental Business Register. A birth is not simply a company incorporation, and an incorporation is not evidence of trading.
The ONS Business Demography 2024 bulletin provides birth, death and survival measures. Its regional results show why a single measure can mislead. London had the highest regional birth rate in 2024, while provisional death data and sector composition add necessary context.
Define a survivor at one, three and five years. Define high growth using the published statistical definition and eligibility conditions. Define scale through employment, turnover or both, because a capital-intensive manufacturer and a software company can grow differently.
Use city-region geographies consistently
Local authorities, combined authorities, travel-to-work areas and International Territorial Levels describe different things. Select one primary geography and map every dataset to it. Where a nation or city does not have a directly comparable authority, document the constructed area and test alternatives.
Registered offices create another problem. A company can register in London while operating elsewhere, or use an accountant's address outside its main site. ONS local-unit data can improve the operating picture, but not every metric is available at the same level.
Publish a geography file with every local authority included. Do not change boundaries after seeing the result. If official boundaries change, release a bridge table and restate the time series where possible.
Build the score from a sequence, not a snapshot
The first measure is formation relative to the active business base. The second is survival by cohort. The third is employment and turnover growth among survivors. The fourth is the share meeting a consistent high-growth definition.
The ONS publishes high-growth data by local authority and region. The scorecard should retain its published eligibility conditions rather than treating every percentage change in a small company as equivalent scale.
Use multi-year averages to reduce noise. Show the number of eligible firms behind each rate. A small city can move sharply because of a few companies, while a large city can add many firms without a dramatic percentage change.
Adjust for sector and economic structure
Professional services, hospitality, construction, manufacturing and life sciences have different formation, capital and employment patterns. A place dominated by easily registered microbusinesses may have a high birth rate without producing more scale-ups.
Publish the raw score and a sector-adjusted view. The adjusted view should compare each place with the national performance expected from its sector mix. Do not hide the raw result, because sector composition is also part of a region's economy.
Regional output can provide context, but it is not a direct business-scale measure. The ONS regional gross domestic product datasets use methods and revisions that need to be documented separately.
Keep finance as an explanatory measure
Capital availability may help firms scale, but it should not be scored as though money raised were an operating result. The British Business Bank Nations and Regions Tracker shows geographic patterns in debt and equity markets. Its findings can help explain a place's outcome.
Separate seed, venture and growth stages. A few large deals can dominate annual regional equity totals. Report deal count, value, median where available and the concentration represented by the largest deals.
Public investment and grants should be recorded by money committed, money deployed and companies supported. Any estimate of additional jobs or gross value added should be labelled as estimated and kept separate from observed business-demography results.
Publish weights and sensitivity tests
An illustrative score might include formation, three-year survival, five-year survival, high-growth incidence and employment growth. The final weights must be chosen before calculating the ranking and reviewed independently.
Run at least four sensitivity tests. Use equal weights. Double the survival weight. Exclude formation. Compare raw and sector-adjusted results. Also show what happens when London is excluded from normalisation, because its scale and company-registration patterns can affect the distribution.
If cities change position materially, publish that instability. A score that creates false precision is less useful than a set of outcome bands. The data file, code or formula workbook and data dictionary should be downloadable.
Interpret the result as a question, not a trophy
A high score should lead to investigation. Which sectors account for survival and growth? Are companies headquartered and operating locally? Does the region retain employment as companies mature? Which customers, suppliers, universities or infrastructure support the result?
A low score also needs context. It may reflect sector shocks, boundary changes, a large informal economy or missing local-unit data. City leaders should not chase business births if the constraint is customer access, management capability or growth finance.
For London's earlier scale question, see Can London turn startup density into durable scale?. For finance geography, see Can a regional UK technology company raise growth capital without moving to London?. The London Founder Operating Survey shows how primary research can complement administrative data.
Limitations
Business Demography covers enterprises on the statistical register and can miss some economic activity. Death figures may be provisional. High-growth definitions exclude firms below eligibility thresholds. Turnover, employment and registered-office data can refer to different units, and city-region boundaries require choices.
The scorecard described here is not yet a completed result. No city should be described as the winner until the dataset, methodology, review and sensitivity analysis are published together.
Reporting by Shoreditch Talk




