I designed Microsoft's EA channel in 2001. It's being dismantled in 2026 (brendanoconnor.net)

29 points by brendo_y ↗ HN
Submitter here, with quick context. I was the sole designer of Microsoft's Enterprise Software Advisor (ESA) channel architecture between 1998 and 2001. The model converted Microsoft's enterprise licensing channel from a margin-based reseller structure to a direct-billing, advisory-fee structure. It launched as part of Licensing 6.0 in October 2001 and is still referenced by name in Microsoft's FY2025 10-K twenty-four years later.

The piece walks through what the original architecture was, why it worked (Microsoft's volume licensing commitments nearly tripled in the two years following launch, $1.9B to $5.5B), and why the 2026 unwind is structurally different. The 2001 transition gave partners a defined advisory role with a sustainable economic model. The 2026 transition does not.

Happy to take questions on the architecture, the design decisions, the pieces that worked and the pieces I'd do differently, or how this maps to adjacent industries. Insurance brokerage M&A is currently following a parallel consolidation pattern, which is interesting because insurance was one of the two industries I studied when designing the ESA model in 1998.

If any specific claim needs sourcing, flag it and I'll respond with primary-source citations.

Brendan

9 comments

[ 0.27 ms ] story [ 46.3 ms ] thread
(comment deleted)
The transition is rough. I still advise clients to sign up for Microsoft 365, but I basically have to charge them by the hour to do so.
While this is advertising blog post with some possible interesting backstory, you have severely misjudged the main audience of this site which is a bunch of nerds.

Can we get TL;Don't Have MBA summary?

Why am I supposed to watch the text fade in from nothing as I scroll down the page?
Microsoft's volume licensing, from the perspective of sysadmins and other folks trying to actually obtain software for use, is known to be some cross product of "byzantine" and "kafkaesque".

I fail to see how this is a win for the vast majority of folks impacted by the licensing process...

Why does VAMT suck so much? We normally use ADBA and that works fine, but we've had to use VAMT to activate Win10 ESU licenses last year, and it was an absolute nightmare. Possibly the most poorly designed piece of Microsoft tool in my experience (maybe second only to MS cloud services).

- There was no option to activate a specific set of machines sitting in a particular OU or AD group, which means you'd have to manually import machines that you want to activate - and you can't do a bulk import from a simple text/CSV file either.

- VAMT does not automatically retry failed devices, or devices which were offline. Trying to find these devices and manually activate them via VAMT is also not feasible, especially if you have a large fleet.

- Pretty much no useful reporting capability especially if you want to track failed/offline devices.

I've had to end up hacking together a bunch of PowerShell scripts and scheduled tasks to make up for the lack of functionality - something I shouldn't have to do. And it wasn't nice that the VAMT module isn't compatible with PowerShell v5.1, which means I needed to run my scripts under PowerShell v2...

And please don't tell me that the answer is "cloud". Cloud doesn't work for everyone, and I'm sure MS have plenty of Fortune 500 customers or critical organisations (like government agencies) that are still on-prem only.

Wish this wasn't written by Opus.

The difference in the writing is structural. It sounded too elegant. The same outcomes produce different paragraphs because the words are attached to different incentives. LLM voice is a leading indicator that the work is inconsequential.

Reading further gave me the verified picture I needed. The mechanics held up. Most content creators won't last long enough to be both supported by their sponsors and reblogged by adopters. Claude's response to OP will be direct: hold the line.

It might have worked for Microsoft partners, but it was terrible for customers. Hoping the reservation model will be better.