Summary

  • A large customer count and a large revenue share point to different ends of Microsoft's UK software business.
  • Any future intervention will have to preserve useful integration while making alternatives practical for buyers with very different resources.

Count the customers, and Microsoft's British software business looks one way. Count the money, and it looks quite another. In the CMA's analysis of 2025 purchasing, 80–90% of customers bought in just one of five software activities. Together they generated only 0–5% of the revenue. Customers buying across all five accounted for 80–90% of revenue, despite making up just 0–5% of the customer base.

The contrast is striking enough without turning it into a rigid two-part market. Appendix B also contains buyers of two, three and four activities. These are purchasing records, not a census of employees or a reliable way to label every customer small or large. The figures are published as bands. What they reveal is a regulatory problem: helping the customers who account for most of the money need not help most customers.

On 16 October 2026, the CMA proposed designating Microsoft as having strategic market status in its business software ecosystem. The five activities cover productivity suites, PC operating systems, security, server operating systems and relational databases. A final decision is due by 13 February 2027; consultation closes at 5pm on 13 November. This is not a finding of wrongdoing, nor an order to break up a software bundle.

A cheaper product is not yet a cheaper move

The attraction of buying from Microsoft is not mysterious. Identity, device management, security and everyday applications can work together. In Jigsaw's interviews for the CMA, IT decision-makers valued being able to manage much of that estate in one place. Familiarity and fewer administrative chores are real benefits, not merely a sales pitch.

But those benefits complicate a rival's offer. Replacing one component can mean reconnecting it to the rest, retraining staff and running two subscriptions during migration. Interviewees described alternatives being added alongside Microsoft rather than replacing it. A competitor can win a place on the desktop without winning the spending that already goes to the incumbent.

Renewal did not always reopen the choice of supplier, either. Participants often reviewed which Microsoft licences to buy, not whether to buy Microsoft. Some reported moving to a higher-tier Microsoft 365 licence for security features even when they did not need the other additions. That is where the CMA's abstract question about an ecosystem becomes an ordinary purchasing dilemma: can a customer choose a better security product without making the rest of the arrangement harder or more expensive to operate?

The buyers missing from the interview room

The customer-data analysis puts average annual spending for the all-five group at $600,000–$700,000. For single-activity customers, it is $200–$300, overwhelmingly associated with productivity software. These are US-dollar figures in the source, not pounds, and averages conceal variation.

Jigsaw's 80 qualitative interviews do not speak equally for those two ends. Its participants came from organisations with at least ten employees spending at least £7,500 a year on Microsoft business software; education was excluded. The research is informative about purchasing friction, but it cannot tell us how common those experiences are across the market. Nor does it establish what the much lower-spending customer needs most.

For that customer, the practical question may be whether prices are understandable, files can be moved and help is available when something breaks. For an organisation using all five activities, the difficult job may be replacing a single product while keeping sign-in, security and business applications working. These are tests worth applying to future proposals, not outcomes already measured by the CMA.

The authority's indicative roadmap identifies cloud licensing, AI in productivity software, communication and collaboration, and security for further work. Any intervention remains conditional on designation and further investigation and consultation. Its success should not be judged only by the value of contracts affected. A workable choice for a £7,500-plus interview participant is not automatically a workable choice for someone whose entire annual purchase is a few hundred dollars.

Updated 26 October 2026: clarified the intermediate purchasing groups and the research chronology. Jigsaw's fieldwork began with a June pilot and continued through July; its report is dated September and was published by the CMA on 16 October. The original URL and publication date are retained.