Microsoft dramatically cut their distributors

The channel is consolidating fast. Microsoft moved from around 180 global distributors to 60 this year and authorised five companies to integrate with its marketplace. Broadcom removed VMware's lowest partner tier, and TD Synnex CEO Patrick Zammit expects more vendors to follow.

The message is clear: vendors are concentrating more business through fewer partners, while the bar for participation is rising.

Microsoft's CSP rules are a good example. From October last year, a distributor must bill US$30 million in CSP revenue over twelve months in each authorised region. A direct bill partner needs US$1 million.

That may work as a qualification filter, but a revenue threshold cannot tell you whether the surviving distributor knows who owns the renewal 24 or 36 months later.

The list gets shorter while the route to market gets longer

Consolidation looks like simplification on a slide. In practice a single end customer now sits behind some combination of vendor, distributor, marketplace, MSP, cloud provider and services partner, several of which have changed hands since the original sale.

Attribution on day one is solved. Every party in that chain can tell you what they booked and when, to the dollar, with a deal registration number attached.

Ask the same chain who is accountable for the upgrade, the expansion or the refresh in month 12/24/36, and the answers stop agreeing with each other.

Peter Bryant, GSI practice leader at Omdia, expects 20% of distributors to be acquired, merged or closed by 2028. Every one of those events hands an installed base to a company that has never reconciled it against its own.

The asset changing hands is a claim about the future

MSP acquisitions rose 73% year on year in the first quarter of 2026, to 64 deals globally. Private equity was involved in 80% of MSP and MSSP transactions, up from 68% a year earlier. Jessica Davis, principal analyst at Omdia, gave the reason without decoration:

"ARR is what attracted private equity to the managed services space."

Recurring revenue is not a fact about the past. It is a claim about the future, and the evidence for it is installed base data: verified end dates, current contacts, entitlement that matches what is deployed, a record of which partner touched the account last.

That evidence usually sits across several systems owned by several companies, and in a lot of organisations lifecycle revenue is still badly run on ERP, CRM and spreadsheets. Those systems record a transaction accurately. They were never designed to orchestrate what happens to it across four parties in three countries over five years.

So the multiple gets priced off the ARR, and the first renewal cycle after close is where anyone finds out whether the ARR was real. By then the attrition reads as market conditions.

Marketplaces settle the sale and leave the aftermath open

Omdia had distribution revenue running 574.6% larger than hyperscaler marketplace revenue last year, with marketplaces transacting under 1% of the tech and telco industry, so this is not a displacement story yet.

It is a precedent story. A marketplace transaction is the cleanest sale in the channel and the least resolved afterwards. Who holds the relationship, whose installed base record is authoritative, who works the renewal, and who finds out first when the customer moves site or gets acquired.

The buying motion is being copied faster than the ownership model behind it. Vendors are pushing more volume through routes that settle the transaction beautifully and say nothing at all about month 24.

Ten accounts, three questions

Take the ten largest customers that reached you through a partner or a marketplace in the past two years. For each one, before opening a system, name three things: the person accountable for their next renewal, the date it falls, and what is installed at their site today.

Then check what you said against the record.

Where the two agree, you own that lifecycle. Where the record disagrees with you, there is a data problem, and a data problem is fixable inside a quarter. The accounts where nobody can name a person at all are the ones that will fail to renew, and the loss will be written up afterwards as competitive pressure.

Worth doing before the next planning cycle, because what was sold twelve quarters ago is what the revenue looks like now, and that link only appears when someone goes looking for it.

The channel does not have a demand problem. The kit is deployed, the contracts exist, the revenue is sitting there 24 months out with a date attached. What it has is a lifecycle execution problem, and consolidation is concentrating that problem into fewer, larger organisations holding more installed base than they can currently account for.