When Microsoft shifted from enterprise licensing to monthly SaaS, the revenue line on our Microsoft productivity business didn't bend. It dropped. The model that had powered years of predictable, upfront revenue suddenly looked like a bad trade: smaller monthly checks, more churn risk, and less obvious room to hide operational mistakes. On the surfce this looked like a licensing change but it was something uch deeper. It was an operating model change most were not ready for.
We did not sirvive by selling harder or accepting lower margins. We got there by redesigning how value was packaged: attaching services to licensing agreements, owning implementations, and effectively inventing the managed services motion that Microsoft VARs and LSPs later adopted as standard. The result was a rebuilt and growing line of recurring business that moved from $750M to $1.2B in annual revenue in under 2 years against a declining market.
The Real Problem We Were Solving
The core problem wasn't "customers don't like enterprise agreements." They dont by the say but that's another issue. It was that our entire commercial and operating model assumed:
- Large, upfront license deals.
- Minimal accountability for actual usage and outcomes.
- Services as optional, episodic line items.
When customers shifted to monthly SaaS, they didn't just change how they paid. They changed how often they could reconsider the relationship. Every month became a mini-renewal. License revenue was no longer pre-sold insulation; it was an ongoing referendum on whether we were still delivering value.
Under the old model, you could get away with selling licenses and walking away. Under the new model, that behavior turned into churn. Our revenue didn't fall because software was worse. It fell because the business was structured for a world where usage and outcomes barely mattered. The old model was poeple in the company and the new model was butts in seats.
How the Old Model Broke in Practice
The first break was in sales behavior. Reps were used to closing big, upfront licensing deals with quota tied to initial contract value. Monthly SaaS made those numbers look anemic. The instinctive response was to discount licenses to preserve deal size — a move that protected short-term optics while quietly eroding long-term unit economics.
The second break was in services positioning. Implementation, migration, and ongoing management were still treated as afterthoughts — nice add-ons if the customer had budget left. In a world where customers could cancel or downsize with far less friction, that posture was suicidal. You can't defend recurring revenue with one-time project thinking.
The third break was in operational accountability. Nobody owned the full lifecycle: license sale, deployment, adoption, and renewal. Sales, services, and support were organized as separate P&Ls and performance systems. Each did its job locally. Nobody was responsible for making sure the new SaaS model held together as a business.
The Operating Play That Turned It Around
The turnaround didn't come from a single program. It came from a series of operating design decisions that redefined what we sold and how we worked.
Attach services as default, not exception
We stopped treating services as optional. Every major licensing agreement came with an attached services motion:
- Implementation baked into the commercial structure.
- Managed services to run, optimize, and support the environment.
- Clear accountability for business outcomes, not just "go-live."
That shifted the economics from "sell once and hope" to "sell a relationship with recurring work attached." It also turned us from a license distributor into an operator: we were now responsible for making the software work in the customer's environment.
Redesign roles and incentives around lifecycle, not transactions
Quotas, compensation, and org design moved away from pure license volume. We built structures where:
- Sales cared about attach rates and long-term value, not just initial contract size.
- Services had a predictable flow of work linked to recurring revenue, not sporadic projects.
- Leadership reviewed the combined performance of license + services, not siloed metrics.
In practice, that meant some people made less money on big one-off deals, but the business stopped depending on them. Revenue became more resilient because it was anchored to ongoing services and adoption, not to one-time license wins. We were actually in control of our finanical destiny.
Make implementation quality a revenue concern, not a cost center
We treated implementation and managed operations as part of the revenue engine, not just delivery cost. If a deployment was weak, that wasn't a line-item overrun; it was a leading indicator of churn. That simple shift justified investing in better playbooks, tooling, and people — because the upside showed up directly in retention and expansion.
Over time, this structure became the pattern others copied. What the market now calls the MSP model — attaching managed services to software and cloud consumption — was, for us, a survival strategy long before it became a category label.
The Parallels to Today's AI-Driven SaaS Apocalypse
Fast-forward to today. The phrase "SaaS apocalypse" is showing up in investor notes and industry commentary as AI tools crush the perceived value of traditional software and compress multiples across the sector. Traders are watching software stocks get hammered as markets price in AI-native competition, feature bundling, and collapsing willingness to pay for point solutions.
The pattern is familiar:
- Creation costs fall; AI can generate or automate what used to be bespoke features or workflows.
- The volume of "good enough" software explodes while per-unit economics compress.
- Value shifts away from standalone tools and toward integration, distribution, data, and managed outcomes.
We went through a softer version of this when we shifted from licenses to SaaS. AI is accelerating it. The economics of pure software are under pressure again. Consumption is no longer measureed by butts in seats but by active token consumption, and the market is punishing vendors that cannot show durable differentiation or clear, outcome-based value.
The lesson from that $1.2B turnaround is straightforward: you do not survive this by arguing valuation with the market. You survive it by redesigning your operating and commercial model so that:
- Software is the substrate, not the product.
- Services, data, and managed outcomes become the center of the revenue story.
- AI labor changes your cost-to-serve and value delivery, not just your demo.
What to Do Next as AI Reshapes SaaS Economics
If you're running a software or services business today, assume you are already in the early stages of a second model shock:
- Audit where your revenue is still effectively "license-like" — large upfront sales, minimal usage accountability, services as optional.
- Identify the equivalent of your services-attached turnaround: what managed outcome, managed service, or AI-powered operating offering should sit on top of your software.
- Decide who owns the lifecycle from sale to realized value, not just the logo.
In the first SaaS wave, the companies that survived and grew didn't just reprice. They redefined their work. In the AI wave, the same will be true. Some SaaS vendors will be marked down as features. Others will quietly rebuild themselves as critical operators of AI-powered outcomes.
The market will call that survival strategy something catchy again. Underneath, it will be the same lesson: when the model breaks, you don't optimize the spreadsheet. You redesign the business.