When I joined Microsoft in June 1998, the enterprise licensing channel was collapsing from three directions at once.

Dell had built its "bulldozer" strategy: winning Enterprise Agreement deals at zero or negative margins to use software as the entry point for displacing HP, Compaq, and IBM hardware in the account. No Large Account Reseller could compete with a player pricing at a loss. Microsoft's own field teams had begun driving EA transactions directly, capturing what had been LAR margins. And the customer experience of Microsoft licensing had become indefensible. CRN's Barb Darrow called it "Kafkaesque" in print.

LAR margins had fallen from above 4% to an average of 2.2%, with Dell-contested accounts running at a loss. The channel that built Microsoft's enterprise business was going out of business.

My boss, Bill Henningsgaard, gave me the assignment in one sentence: "The channel is going out of business. Give me some ideas."

Rather than benchmarking other software companies, I studied industries that had solved analogous problems. The State Farm agent model was the most instructive. Independent companies were agents of the parent in the legal and commercial sense. They didn't set price, didn't hold inventory, and didn't compete by discounting. They competed by being better advisors, with activity-based economics that rewarded top performers. That research became the foundation of the Enterprise Software Advisor (ESA) model.

At a closed-door meeting at the Hilton Chicago O'Hare, Henningsgaard brought in the CEOs of three of Microsoft's most influential LARs: Howard Diamond of Corporate Software, Keith Coogan of Software Spectrum, and Paul Jarvie of ASAP Software. When one of them asked how Microsoft would define exactly what services partners had to perform to earn advisory fees, I answered with a philosophy rather than a checklist. We'd give partners the categories and focus areas that mattered, then let the market and customers establish the service levels. Over-specification would commoditize the advisory role. A thoughtful framework would elevate it. That answer defined the architecture of the ESA model for the next twenty-four years.

The economic insight that made the model work was structural. Microsoft was already giving LARs a 17.7% discount on software as part of the existing indirect model. Moving to a 4% activity-based advisory fee was well inside that existing budget. The transition wasn't asking Microsoft for new money. It was redirecting money it was already spending, converting an undifferentiated volume discount into a targeted performance-linked fee.

When I presented this in a mid-year review attended by about forty people including Gates and Ballmer, Ballmer's immediate reaction was: "This is a perpetual motion machine!"

It sounded too elegant. Henningsgaard pulled me aside afterward and was direct. I needed to come back with the math made explicit, and to do that I needed data Microsoft didn't have, on what the channel was actually earning under the existing structure. I got it the only way I was going to get it. During my research and collaboration with our key channel partners, I had cultivated strong relationships and built trust. Howard Diamond of Corporate Software was one of those key partners. During a visit to his company's offices in Boston, Diamond cleared the room of his own VPs and the Microsoft account team, opened the company's live SAP system, and walked me through the actual margin on every deal, line by line. Diamond didn't share that data because he was naive. He shared it because he trusted me to use it to protect the channel's economics rather than exploit them.

That session gave me the verified picture I needed. The mechanics held up. Ballmer accepted the framework and told Henningsgaard to take it forward.

The model split the EA channel into three tiers covering 75,000+ addressable accounts and an $11.5B opportunity envelope: 1,150 Microsoft-led global strategic accounts at a 4% ESA fee, 14,000 channel-assisted corporate accounts at 9%, and 60,000 channel-led medium enterprise accounts at 15%. Microsoft billed the customer directly across all three tiers. What changed was who led the sale, what role the partner played, and how the partner got paid. The channel was converting from a margin model, where partners set end price through discounts, to an advisory fee model, where Microsoft set price and partners earned fees for services delivered. An ESA was required on every deal.

I originally designed three branding tiers: Enterprise Software Advisors, Global Software Advisors, and Software Advisors for SMB. Then my counterpart on the maintenance product team told me their research had landed on a name for the new maintenance offering: Software Assurance, abbreviated "SA". Same letters as my SMB partner tier. I dropped GSA and SA and kept ESA as the single brand. That name is still in use in Microsoft's FY2025 10-K, twenty-four years later.

The architectural work happened inside a five-person Worldwide Licensing & Pricing team co-developing the broader Licensing 6.0 program. I led the channel piece end to end. The team co-developed Software Assurance and the rest of L6.0, taking direction from Steve Ballmer through Bill Landefeld.

Resistance was a leading indicator that the work was consequential. Compaq CEO Michael Capellas called Steve Ballmer in an outraged rant about the transition, mentioning me by name. Like many companies in the dot-com frenzy, Compaq's compensation was tied to top-line revenue and EA software was contributing substantially. Ballmer's response to me was direct: hold the line.

Licensing 6.0 launched October 1, 2001 across the United States, Canada, and 22 Western European countries, with Australia following in February 2002. Across an 18-month enrollment window, Microsoft sold 2,577 Enterprise Agreements.

Microsoft's unearned revenue, the balance sheet line that captures committed multi-year EA value, grew from $1.92B in June 2001 to $7.74B in June 2002. That $5.82B 12-month surge is the financial fingerprint of the transition. Microsoft's 10-K filings attributed FY2002, FY2003, and FY2004 revenue performance directly to the pre-transition enrollment.

Average ESA fees under the new model came in at roughly 3.72%, with top performers above 4%. Partners who delivered value earned more than they had under the discount model. Partners who didn't were exposed. The same dollars produced different outcomes because they were attached to different incentives.

The ESA designation I created remains in Microsoft's FY2025 10-K, twenty-four years later.