Fifty suppliers were shut out of G-Cloud by mistake

A UK startup was rejected from G-Cloud 15, the government's main digital marketplace, because it had raised money on a SAFE and the framework's financial assessment read that as debt.

Government was its main customer, so the rejection was close to existential. When it was queried, the answer came back that the whole thing had been a policy oversight. SAFEs should never have been treated that way, and more than fifty other suppliers had been rejected for exactly the same reason.

Nobody set out to exclude them. A requirement was written, it had an effect nobody had thought about, and fifty-odd firms disappeared from a market before anyone noticed.

That story comes from a policy paper the Competition and Markets Authority (CMA) published on 8 September, and it is the paper's whole argument in miniature.

The CMA is the competition regulator, not the procurement regulator, which means it cannot compel any of this and can therefore be blunt in a way the Cabinet Office never is.

After two years looking at scale-ups, civil engineering, defence and cloud, its conclusion is that procurement keeps turning out to be the thing quietly deciding which firms grow.

The cost of bidding is now an official problem

The sharpest recommendation is this. When government introduces a regulation, it is routine to assess what that will cost firms and how it affects their ability to compete. Procurement requirements are not regulation in law but have near-identical economic effects, so they should be tested the same way. Does this raise fixed costs? Does that cost land harder on smaller bidders? Is the benefit worth the suppliers it filters out?

Nobody asks, so requirements pile up. A medical equipment supplier with five employees described holding a Carbon Reduction Plan, an Evergreen assessment, a Modern Slavery assessment and Cyber Essentials to sit on NHS frameworks that guarantee no income at all. A West Midlands consultancy worked out that a dozen bids was an unacceptable gamble and stopped tendering. It did not lose a competition. It left the market.

Note that this is not a rules problem. The Procurement Act 2023 already gives authorities the flexibility. The incentives point the other way, because procurement teams are rewarded for compliance and for not being challenged.

"British" is not one thing

If you are positioning on being a UK supplier, read this bit. The CMA's point is that workforce, production, headquarters, R&D, IP and ownership can all sit in different places, and which one counts depends on what the buyer is trying to achieve. Resilience means where production happens. Strategic autonomy means who owns the IP and controls the outputs. Jobs means domestic employment and training.

So stop claiming Britishness generally and claim the specific dimension the requirement is actually reaching for. The same applies to social value. One case study in the paper: a UK defence tech SME scored badly because it did not have many UK SMEs in its own supply chain, and lost to a US multinational that did.

What to do with it

Three practical things.

Challenge exclusions. The G-Cloud error only came to light after the rejection was queried. If a financial assessment or turnover threshold knocks you out, ask for the reasoning in writing and offer an alternative such as a bank guarantee. A rejection is not always a judgement on your business.

Use it in clarification questions. When a requirement looks disproportionate to the contract value, you now have an independent regulator's published position saying that is a competition problem, with case studies from the CBI, FSB and British Chambers of Commerce attached. That is a citation, not special pleading.

Look where the money is already moving. Sovereign AI's £100m rapid procurement scheme has no turnover minimums, pays upfront and lets suppliers keep their IP. The new Innovation Marketplace lets buyers test solutions before a formal procurement starts. Both are small. Both are built around how smaller firms actually operate, and both are worth more attention than the next framework renewal.

The suppliers knocked off G-Cloud did not lose on price or quality. They lost to a requirement nobody had stress-tested. None of that is news to anyone who bids for a living. What is new is who is saying it.