Latest / SaaS Business with Fexingo: Software-as-a-Service Companies, ARR, and Recurring Revenue / The SaaS Margin Trap of Free Upgrades
Transcript
- Lucas: Most software companies today sell unlimited plans because they think it removes friction from the sales cycle. But if you actually look at the underlying infrastructure costs for a data-heavy analytics platform, those flat fees often disappear within the first six months of a customer’s life. Luna: Right, so you’re selling peace of mind while your engineering team pays for every extra gigabyte processed. It feels like a race to the bottom on cost controls. Lucas: Exactly. And the worst part is that it hides in plain sight inside your net revenue retention numbers. When you offer an unlimited tier, you stop seeing the signal of heavy usage that usually tells you a customer is ready to expand or is about to churn due to performance issues. Luna: So the unlimited plan acts as a blind spot for both sales and product teams. How do you track health when everyone is just hitting the ceiling? Lucas: You have to invent new signals. Let’s talk about a mid-market company I worked with last year, let’s call them DataFlow Analytics. They launched a pro-tier unlimited plan to compete with two larger rivals. Within three quarters, their gross margin dropped by twelve percentage points. Luna: Twelve points is massive. That’s not just inefficiency, that’s structural bleeding. Did they try to cap the usage immediately? Lucas: They tried to reverse course, but the backlash was immediate. Customers who had scaled up felt betrayed by sudden limits. So instead of hard caps, they moved to what we call soft throttling—slowing down processing speeds during peak hours rather than cutting access entirely. Luna: That’s a clever pivot. You trade speed for stability, which preserves the relationship without giving away infinite compute resources. Lucas: It works, but only if you communicate it clearly. The problem is that most founders treat unlimited as a marketing feature rather than a financial liability. They forget that server costs scale linearly with usage, while their revenue stays flat. Luna: So the real question is whether the average customer on an unlimited plan actually uses enough to justify the infrastructure cost, or if they’re just gaming the system. Lucas: Usually it’s the latter. You get a small group of power users who consume eighty percent of the capacity, driving up costs for the majority of quiet subscribers who pay the same price. Luna: That sounds like a classic subsidy problem where healthy accounts pay for the abusive ones. Is there a better way to structure those tiers? Lucas: Yes, you introduce progressive overage billing. Once a customer hits ninety percent of their baseline, they start paying per unit for the excess. This aligns their incentive to optimize their own workflows with your need to cover costs. Luna: But doesn’t that kill the perceived value of the unlimited label? Sales teams hate losing that simple headline. Lucas: They do, but sophisticated buyers prefer transparency. A well-designed overage model looks like a safety net, not a trap. It protects you from catastrophic loss while still offering predictability for normal operations. Luna: I’ve seen some companies use fair-use policies written in tiny print, but that always backfires when customers find out. Transparency builds trust, even if it’s less sexy. Lucas: Agreed. And here is the counter-intuitive part: implementing fair-use limits often reduces churn. Why? Because your engineers can focus on optimizing performance for the ninety-five percent who stay within bounds, rather than firefighting crashes caused by the top five percent. Luna: So the limited plan actually improves the experience for the mass market. That is a brilliant operational side effect. Lucas: It is. But getting there requires courage. Most CFOs will approve a flat fee because it’s easy to forecast. They struggle to approve complex tiered structures that require constant monitoring and communication. Luna: Forecasting is king in these meetings. If you can prove that the unlimited plan is costing more than it brings in, does the math win? Lucas: Only if you show the long-term impact on lifetime value. A customer who stays happy on a capped plan is worth more than a noisy one on an unlimited plan who eventually leaves due to poor service. Luna: Lifetime value beats short-term acquisition metrics every time. So the strategy is really about protecting the existing base rather than chasing new logos at any cost. Lucas: Precisely. And for those listening who are building or running a software business, this is a reminder to audit your heaviest users right now. Check their actual consumption against their plan price. Luna: If you find a segment that is consistently breaking even or losing money, you have a clear case for restructuring those terms before the next funding round or earnings call. Lucas: Exactly. It is much easier to adjust terms with current customers than to rewrite your entire pricing architecture after you hit a growth wall. Luna: Before we go deeper into the technical implementation of these limits, I want to pause for a moment about how this show exists. Lucas: If these conversations are useful for what you're building or running, know that we don't take ads. We rely entirely on listener support to keep the production quality high and the research deep. A small group of people already chips in monthly through buy me a coffee dot com slash fexingo, and that is literally what funds making this many of these episodes possible. Luna: It keeps us independent and allows us to dig into topics like this without corporate pressure. If you get value from the analysis, consider joining that community. Lucas: Thanks for keeping the lights on. Now, back to the mechanics. When you implement soft throttling, you need to define what slow means. Is it twenty percent slower? Half speed? Luna: That depends on the latency tolerance of your user base. For real-time trading tools, any delay is unacceptable. For monthly reporting dashboards, a slight lag might be invisible. Lucas: Right. So you must calibrate the throttle to the specific use case. You cannot apply a generic slowdown across all features. It has to be granular. Luna: Granularity adds complexity to the engineering roadmap. Is it worth the development time to build such precise controls? Lucas: For high-growth SaaS companies, yes. The alternative is burning cash on servers that nobody needs. Precision engineering in pricing is just as important as precision engineering in code. Luna: I love that framing. Pricing code is just as critical as application code. One delivers the promise, the other delivers the profit. Lucas: Exactly. And finally, always give customers a grace period when you change policies. Let them upgrade voluntarily rather than forcing a downgrade overnight. Luna: Voluntary upgrades preserve dignity. Forced changes feel punitive. The difference in customer sentiment is enormous. Lucas: So, as you review your own pricing pages this week, ask yourself if unlimited is truly serving your business model, or if it is just a habit from early-stage competition. Luna: It is a good question to sit with. Sometimes the simplest constraint creates the most sustainable growth path. Lucas: We will see how this plays out in the next quarterly reports from major cloud providers. Until then, keep your margins tight and your communications clearer. Luna: Thanks for listening. See you next time.