Microsoft's Large Solution Provider commissions have collapsed on a schedule that reads like a controlled demolition: $2.5B in 2023, $1.67B in 2024, $583M in 2025, and $0 in 2026. Writing in The Register on June 1, Brendan T. O'Connor (Brushton Group), the architect who designed Microsoft's Enterprise Software Advisor channel in 2001, argues that the 2026 transition is repeating the structural mistake that broke the LAR channel in 1998 and is doing so without the bridge mechanism that made the original ESA reset survivable.
The history matters because the pattern is familiar. In 1998 LAR margins compressed from 4 percent to 2.2 percent. Dell began selling Enterprise Agreements at 0 percent to negative 2 percent margin as hardware loss leaders. Microsoft's own field sales force was going direct on the largest accounts. Reseller economics turned, in O'Connor's word, Kafkaesque. Bill Henningsgaard, then running Microsoft's US enterprise business, summarized the brief to O'Connor in one line: "The channel is going out of business. Give me some ideas."
The 2001 bridge that worked
What O'Connor's team built in response is worth describing in detail, because it is the counterfactual against which the 2026 reset has to be judged. The ESA architecture flipped three things at once. Billing moved from indirect to direct, so Microsoft owned the customer paper. Resellers were redefined as advisors rather than transactional middlemen. And compensation moved from product margin to activity-based fees, paid out of a redirection of part of the existing 17.7 percent volume discount Microsoft was already funding inside list price. Nothing new had to be invented on the P&L; the money was already in the system, it just changed hands differently.
The fee schedule was tiered to match where advisors actually added value. Microsoft-Led covered roughly 1,150 strategic accounts representing about $5B of opportunity at a 4 percent advisor fee. Channel-Assisted covered around 14,000 accounts and $3B of opportunity at 9 percent. Channel-Led covered roughly 60,000 accounts and $3.5B of opportunity at 15 percent. The slope was deliberate: the smaller and harder the customer was to reach directly, the more the advisor earned. Steve Ballmer reportedly called the design "a perpetual motion machine" when it was walked through internally.
The results showed up in the deferred revenue line almost immediately. Microsoft's unearned revenue went from $1.92B in June 2001 to $7.74B in June 2002, a $5.82B swing in twelve months. Across the United States, Canada, and twenty-two Western European countries, 2,577 Enterprise Agreements were closed in the first eighteen months under the new model. The channel did not die; it was redirected, with a clear new job and clear new economics for doing it.
Why the 2026 transition has no bridge
The 2026 reset has the demolition phase without the rebuild. LSP transaction commissions are going to zero on a stepped schedule that partners have been able to forecast for two years, but there is no replacement advisory role attached to it, no activity-based fee schedule, no tiered account segmentation that says which partner gets paid for which motion. Microsoft's only public guidance to displaced LSPs has been to "pivot toward managed services and cloud consumption," which is a strategy statement, not a compensation model. In 2001 the bridge was funded out of the volume discount before the old commissions were cut. In 2026 the old commissions are being cut and partners are being told to find a new river to fish in.
The market is already pricing the divergence. In May 2026 Bytes Technology Group, a UK Microsoft reseller, warned on declining operating profit and cited Microsoft incentive changes directly. Ten days later Softcat, addressing a very similar customer base, raised FY26 guidance to the mid-teens. Same vendor, same geography, broadly the same end customers, opposite outcomes. That is the signature of a transition where the rules of who gets paid for what are unsettled enough that execution differences swamp structural ones, and it is exactly the volatility CIOs absorb at renewal.
Regulators are watching the same picture. On May 14, 2026 the UK Competition and Markets Authority opened a Strategic Market Status investigation into Microsoft's business software, with the public comment window closing June 4. SMS designation would give the CMA conduct-requirement powers over how Microsoft prices, bundles, and routes customers through partners, which is precisely the surface area the 2026 channel reset is reshaping.
What CIOs renewing in the next 18 months should do
At bruno.digital we are seeing the practical consequences land inside renewal cycles already in flight. Three things we would put on the table for any CTO or VP Engineering staring at a 2026 or 2027 EA renewal.
First, separate the advisory relationship from the transaction relationship in writing. The partner who helps architect your Azure landing zones, license your Copilot estate, or model your M365 E5 step-up may no longer be the partner who books the paper. Treat advisory scope, deliverables, and fees as a standalone contract rather than something assumed to be funded inside reseller margin, because in 2026 that margin is gone.
Second, price the optionality of going direct. Microsoft direct billing is now the default path for large enterprise EAs. Get a direct quote alongside the partner quote on the same SKUs and the same term, and use the delta to size what you are actually paying for advisory, procurement, and true-up management. In several engagements we have run this quarter the implicit advisory load was between 3 and 7 percent of total contract value, which is a number worth naming explicitly.
Third, build a fallback for partner failure into the renewal itself. The Bytes versus Softcat split is a warning that some incumbent LSPs will not be standing in their current form by the next renewal. Contractual rights to novate, to access your own usage telemetry directly from Microsoft, and to move support cases off partner-mediated channels should be negotiated now, not after the partner restructures.
What to watch next
Two near-term events will tell us which way this lands. The CMA's SMS comment window closes June 4, and the shape of the responses, particularly from UK resellers and large enterprise buyers, will signal whether conduct requirements on Microsoft's channel economics are on the table for 2027. Then Softcat's interim results in the autumn will show whether its May upgrade was a durable share gain from a weakened Bytes or a one-quarter timing benefit from EA renewal pull-forward. A durable gain says the channel is consolidating around survivors. A reversal says even the apparent winners cannot make the new economics work, and the bridge O'Connor describes is not just missing, it is not buildable from where Microsoft is standing today.
