Belfast can turn a public testbed into an exportable product only when a company leaves with usable intellectual property, evidence a buyer recognises, a route through procurement and a proposition that works outside the original site. Access to facilities or data is valuable, but it is not a commercial outcome. The decisive test is whether the company wins a paid contract or repeatable customer after the programme ends.

This distinction matters because city-deal projects and innovation programmes are often described through investment committed, facilities planned and organisations involved. Those figures show inputs. A buyer or investor needs evidence about products deployed, customers won, revenue earned and jobs that remain after public support.

Choose a testbed for the buying problem

A useful testbed reproduces the operating environment of a target customer. A cyber-security company may need realistic network traffic, governance and incident procedures. An advanced-manufacturing supplier may need equipment, production data and safety controls. A digital-twin business may need access to physical assets, geospatial data and staff who understand how decisions are made.

The Belfast Region City Deal describes a portfolio spanning innovation, digital capability, tourism and regeneration. The Northern Ireland Department for the Economy's City and Growth Deals overview explains the wider public framework. These sources identify intended infrastructure and funding, but a company still needs project-level eligibility, dates and access conditions.

Before applying, write down the customer problem, the evidence required to solve it and the next buyer. If the programme cannot provide the necessary environment or a reference that external buyers accept, it may still support research, but it is the wrong commercial testbed.

Set intellectual-property terms before work starts

Companies should identify background IP, project IP, data rights, publication rights and rights to improvements. A vague promise that the business will retain its IP is not enough if a university or public partner can publish results before a patent filing, if software depends on restricted data, or if the company cannot reuse a model outside the test site.

The contract should state which party owns code, designs, training data, annotations, documentation and test results. It should also state who can grant customer licences and whether partners receive royalty-free rights. Where university facilities or researchers are involved, commercial terms need to align with funding conditions and institutional policy.

Data can be more restrictive than the software. The team must know whether it can retain raw data, derived data or only aggregate findings. A product trained on a local dataset may not generalise to another customer, and data-sharing terms may prevent the company from proving why.

Treat procurement as part of the experiment

A successful proof of concept does not automatically become a public contract. The testbed sponsor may not be the budget holder, and the pilot may have been purchased through a route that cannot support a larger deployment.

Companies should identify the procurement route before accepting the pilot. Northern Ireland public opportunities can be researched through the eTendersNI portal, while UK-wide opportunities may also appear through Find a Tender. The buyer should explain whether a later contract will be competed, whether a framework is required and what evidence will be assessed.

The company needs a priced deployment plan, not just a technical report. That plan should cover implementation, security, training, support, data migration, service levels and exit. If the testbed uses exceptional staff effort or free infrastructure, the team must calculate the cost of delivering the same outcome for a normal customer.

For a fuller public-sector sales route, see How a London SME wins its first public-sector contract.

Build evidence that travels beyond Belfast

Exportable evidence answers questions a new buyer will ask. What was the starting condition? What changed? Over what period? Which other factors affected the result? Can the product be deployed with different data, systems and regulation?

A programme completion certificate is weak evidence. A stronger pack includes a documented baseline, technical performance, user adoption, operational cost, security review, limitations and permission to identify the reference customer. If the sponsor cannot approve a public case study, the company should negotiate an anonymised evidence statement or a private reference process.

Northern Ireland has distinct public institutions and legal arrangements. A company selling into Great Britain, Ireland, the European Union or the United States should not assume the original contract, data terms or assurance will transfer. The export plan needs a market-specific buyer, procurement route and regulatory review.

Decide whether a London sales presence earns its cost

London can provide access to UK government departments, large corporate buyers, advisers and international investors. That access may be useful after a Belfast testbed, but it does not require an immediate headquarters move.

Start with scheduled customer visits, relevant industry events and a named commercial owner. Measure qualified meetings, procurement progress and contracts. Add a permanent London presence only when the sales cycle requires frequent local contact or when a critical hiring pool is materially deeper there.

A London office without a repeatable proposition can create activity without revenue. A Belfast product team should retain clear authority over feasibility and delivery so commercial promises remain connected to what the product can do. How regional UK exporters use London sets out the wider decision.

Measure the route from programme to market

Every programme should publish or enable a cohort table that separates applicants, selected companies, prototypes completed, paid follow-on contracts, external customers, exports and private capital. These stages should never be collapsed into one success figure.

The Belfast Region City Deal annual reporting can show programme and project progress. Business outcomes should be cross-checked against procurement notices, company filings and customer confirmation. The NISRA Inter Departmental Business Register provides economic context, but it cannot attribute a company's growth to one testbed.

For each participating business, reviewers should ask whether the product existed before the programme, how much public support it received, what IP changed, who paid next and whether jobs and revenue remained in the region.

Limitations

Many Belfast innovation investments are multi-year projects. Announced funding, construction and programme participation should not be presented as completed commercial impact. Public contract values may include services beyond a testbed, and company filings can lag operational events.

This draft does not claim that a named programme has produced export revenue without company and customer evidence. Publication should follow a cohort of participants for at least two years and offer relevant programmes a right of reply.

Reporting by Shoreditch Talk