What the Procurement Act changed for suppliers

Fifteen people in the Government Commercial Function are dedicated to the full-time management of the nineteen largest technology suppliers to the British state. Four more sit in the Government Digital Service. Around them is a commercial network of roughly 6,000 people, most of whom are not technical specialists and who buy everything else government buys.

The Public Accounts Committee put those numbers on the record on 6 June 2025 and called the position not tenable, set against digital procurement of at least £14 billion a year. That is the fact any account of procurement reform has to sit alongside the legislation, because it determines how much of the legislation will be felt.

What actually commenced

On 24 February 2025 the Procurement Act 2023 replaced the regime inherited from the European Union directives. A new National Procurement Policy Statement was laid in Parliament on 13 February 2025 to take effect alongside it.

Three mechanical changes matter to anyone selling technology.

First, transparency extends past the award. Section 52 requires an authority to fix and publish a minimum of three key performance indicators on contracts it expects to exceed £5 million. The section defines an indicator as a factor or measure against which a supplier’s performance can be assessed during the life-cycle of the contract, and it exempts several contract types, frameworks and light touch arrangements among them.

Second, performance against those indicators is not private. Section 71 requires the authority to assess performance at least once in every twelve months of the contract’s life and on termination, and to publish specified information about the assessment.

Third, a single platform now carries the notices. The Cabinet Office introduced a data platform in March 2025 for departments to submit spending and pipeline information, and central publication of procurement data is intended to give the commercial function faster access to what it buys.

Taken together, these convert public procurement from a series of award events into a continuous published record. A supplier’s history with the state becomes, in principle, readable by every other buyer in the state.

What did not change

The buying side did not get bigger, and the market it faces got more concentrated.

DSIT told the committee that technology markets are increasingly dominated by very large suppliers and that the way supply chains want to charge, bill and work with departments has changed. The Cabinet Office described a shift from capital expenditure to resource expenditure that has a considerable effect on how it procures and manages cost. Both are describing subscriptions replacing purchases, which moves the decisive moment of a commercial relationship from the tender to the renewal.

Government’s response is a Digital Commercial Centre of Excellence, announced in the blueprint published in January 2025, with an objective of identifying reforms that help startups, scaleups and small and medium-sized enterprises reach government contracts. The committee reported it will have 24 experts, against the 6,000 general commercial staff across government, and noted that the Cabinet Office and DSIT have far wider ambitions for it than its published aims: harnessing the digital and commercial functions together, improving data on technology spending, leveraging buying power, helping departments optimise cloud use, and upskilling commercial staff.

Twenty-four people cannot do all of that. The question a supplier should be asking is which of those five things the centre actually spends its time on, because that determines whether it meets one at a negotiation or never at all.

The data problem underneath the buying power argument

The £14 billion figure is an estimate produced by third parties. Government has not been able to provide a more precise one, and it holds insufficient information about the pipeline of demand coming from departments.

That is not a statistical inconvenience. It is the reason government cannot use the leverage it obviously has. DSIT explained to the committee that in cloud economics the ability to write a commitment to a provider matters enormously and produces maximum commercial leverage. The commercial function’s own position was that to get the best leverage you would do a single deal, while conceding that lining up multiple customers with differing requirements is very difficult in practice.

So the state has signed agreements with the major cloud providers allowing them to treat the public sector as a single customer for volume discount purposes, while DSIT describes the actual approach to buying cloud as fragmented and, in its own word, not right. The State of digital government review that preceded all this, published on 21 January 2025, found that less than 50% of public sector digital spend with third parties is even covered by Crown Commercial Service frameworks, and gave an example of around 10% in missed volume discounts on Google products across twelve departments. Actual spend with managed service providers exceeded contract value by over 50% in 2023.

A buyer that cannot forecast cannot commit, and a buyer that cannot commit pays close to list. Every supplier reading this already knows that. What is new is that the buyer has now said it in public, to a select committee, on the record.

The next wave is aimed at who wins, not how

The March 2026 government response to the consultation on growing British industry, jobs and skills sets out where this is going. The consultation ran from 26 June to 5 September 2025 and drew 811 responses to the main survey, of which around a third came from suppliers and about 44% from contracting authorities.

Three decisions in it change supplier behaviour more than anything in the Act.

A minimum evaluation weighting for social value on contracts above £5 million, requiring authorities to set an award criterion relating to the quality of the supplier’s contribution to jobs, opportunities or skills. Respondents warned against pushing the weighting higher, arguing it would disadvantage smaller firms.

An exclusion for suppliers on contracts of £5 million and above who cannot demonstrate prompt payment of invoices. The consultation floated 60 days; smaller suppliers and strategic suppliers pushed back and argued for 30. Note the direction of that pushback, which is not the one lobbying usually runs in.

And three-year targets for spending with small businesses and voluntary, community and social enterprises for authorities spending £100 million or more a year, with annual reporting against them.

The response is explicit that legislative changes depend on parliamentary time, and no implementation dates are announced. Anyone building a bid strategy on the assumption that all of this lands on a known date is building on sand.

What a supplier should do differently

Four things follow from the record, and none of them is about bid writing.

Read your own published performance. If you hold contracts above £5 million awarded since February 2025, KPIs exist for them and assessments will be published annually. Those documents will be read by the next authority that shortlists you, or they will not be read by anybody, and you should know which before a competitor tells the story for you.

Treat the renewal as the procurement. Subscription models moved the leverage there, government has acknowledged that its processes have not caught up, and the gap between those two facts is where the margin currently sits. It will not stay open indefinitely.

Assume the buyer is thinly staffed and be legible anyway. Fifteen people covering the nineteen largest suppliers means most commercial conversations happen with someone whose specialism is not technology. Proposals that require the reader to already understand the architecture get scored by someone who does not.

And take the payment and SME provisions seriously earlier than the law requires. Prompt payment becomes a gate rather than a scoring line, and gates do not respond to a persuasive narrative. This is the pattern running through the whole reform: government is gradually converting things it used to score into things it either verifies or publishes.

The framework question nobody has settled

One structural feature survived the reform untouched, and it is the one most likely to determine who wins.

Frameworks are exempt from the section 52 KPI requirement. They are also how a large share of public technology is bought, and less than half of public sector digital spend with third parties runs through Crown Commercial Service frameworks at all, which means the rest sits on arrangements nobody has a central view of.

A framework is a shortlist with a shelf life. It is compiled at a moment in time, against a specification written for the market as it then was, and it then governs buying decisions for several years while that market changes underneath it. Suppliers who were not ready on the day the framework was let are not merely at a disadvantage. They are ineligible, regardless of what they can now do.

That is defensible as an efficiency measure and it works against the stated aim of opening government to startups and scaleups, which is precisely the constituency least likely to have been in a position to bid three years ago. The Centre of Excellence’s published objective is to fix this. Its 24 people will be doing so against an incumbency effect measured in billions.

The thing to watch

Whether the Act changes who wins is an empirical question, and the notices will answer it slowly. Anyone offering a verdict now is guessing, because the first annual assessments under section 71 have only just started to appear and the concentration picture takes years of awards to shift.

What is already clear is that the reform’s centre of gravity is disclosure rather than competition. More is published, at more points in the contract’s life, about more suppliers. That is a genuine change and it is worth almost nothing unless somebody reads it, which is why the interesting institution in British procurement is no longer the framework. It is whichever body first commits to reading the performance notices in bulk and saying what they show.

Automation is one of the categories where this matters most, because a running service costs money every year that a pilot did not, and the arithmetic of that is set out in intelligent automation. The broader picture sits under public sector technology.