Dispatch: When the Freemium Model Breaks

San Francisco · March 26, 2026

There's a moment in every freemium company's life when someone pulls up a spreadsheet and says the words nobody wants to hear: "We need to talk about the free users."

Ours happened on a Tuesday. The head of infrastructure had been running the numbers on server costs, and he'd discovered something that, in retrospect, should have been obvious: our 127,000 free users were consuming 41% of our compute resources while generating exactly zero dollars in revenue. The cost to serve them was $34,000 a month. Not catastrophic, but not nothing — especially for a company that had just missed its quarterly revenue target by 8%.

"They're not free users," he says, standing at the front of the room with his laptop connected to the projector. "They're negative-revenue users. Every one of them costs us twenty-seven cents a month."

The VP of Marketing, sitting across from me, shifts in her chair. She built the free tier. It was her idea, her pitch to the board, her baby. Three years ago, she stood in this same room and argued that freemium was the future of B2B SaaS, that PLG would replace the sales-led motion, that free users were the top of a funnel that would fill itself.

She wasn't wrong, exactly. The free tier did fill the funnel. We went from 2,000 signups a month to 14,000. The growth chart looked like a hockey stick, and for eighteen months, we were the darlings of our investors' portfolio reviews.

The problem was the conversion rate.

Here's what the funnel actually looked like:

14,000 free signups per month. Of those, 8,200 completed onboarding. Of those, 3,100 were still active after thirty days. Of those, 340 started a trial of the paid plan. Of those, 98 converted to paying customers.

Ninety-eight. Out of fourteen thousand.

That's a 0.7% free-to-paid conversion rate. Industry benchmark for B2B freemium is somewhere between 2% and 5%, depending on who you ask and how they define "conversion." We were less than half the low end.

"But the brand awareness," the VP of Marketing says. She says it quietly, almost to herself.

She's right that there's a brand effect. People know our name. We show up in G2 reviews and comparison blog posts. When someone searches for our category, we're on the first page. But brand awareness is one of those things that's impossible to put a dollar figure on, which means it gets used to justify anything.

Brand awareness is one of those things that's impossible to put a dollar figure on, which means it gets used to justify anything.

I spent a week doing the math that nobody had done. The fully loaded cost of the free tier — servers, support tickets from free users (22% of total ticket volume), engineering time spent on free-tier-specific features, the portion of the marketing budget devoted to top-of-funnel acquisition — came to $73,000 a month. The revenue from the 98 customers who converted averaged $189 per month each. That's $18,522 in new MRR per month.

Customer acquisition cost for a freemium-converted customer: $745. CAC for a direct-signup customer who came through our content marketing or paid ads: $312. The free tier was more than twice as expensive per acquisition as our other channels.

The debate lasted three weeks. I've never seen the leadership team this divided. The CEO, to her credit, didn't take a side. She let us fight it out.

The VP of Marketing argued for keeping the free tier but restricting it — fewer features, usage caps, a "freemium lite" that would cost less to serve but still capture demand. "You don't throw away 127,000 email addresses," she said. "You monetize them differently."

The VP of Sales wanted to kill it entirely and replace it with a demo-request flow. "Give me the marketing budget you're spending on free acquisition," he said, "and I'll hire two more SDRs and double our enterprise pipeline." He had a spreadsheet showing that each SDR generated an average of $41,000 in new ARR per quarter. The math checked out, on paper.

I was somewhere in the middle, which is where I usually end up. The free tier wasn't working as designed, but killing it felt like treating the symptom instead of the disease. The real question was: why weren't free users converting?

So I did what I always do when I'm stuck. I talked to customers.

I set up interviews with three groups: free users who'd been active for more than six months and never upgraded, free users who had upgraded, and people who'd signed up for the free tier and then churned within the first week.

The long-term free users were the most interesting. They'd built the product into their workflows. They used it every day. They just didn't need more than what the free tier offered. When I asked one of them — a marketing manager at a 30-person company — why she hadn't upgraded, she laughed. "Why would I? Everything I need is right here."

We'd given away too much.

In the end, we didn't kill the free tier. We didn't keep it either. We did something in between, which probably means we made everyone equally unhappy.

We replaced the unlimited free tier with a thirty-day free trial of the mid-tier plan. Full features, real data, the whole experience. After thirty days, users could either pay or drop down to a severely limited "starter" plan — one user, one project, basic reporting only. Enough to keep the door open but not enough to run a real workflow.

We also changed what happened during the trial. Instead of a generic onboarding email sequence, we built what the product team called "value paths" — guided workflows that got users to their first meaningful outcome within three days. If you were a marketing team, you'd set up campaign tracking. If you were a product team, you'd build a feature usage dashboard. If you were an ops team, you'd create an automated workflow.

The hypothesis was that if people experienced the full value of the paid product, more of them would pay when the trial ended. Simple enough to state. Hard to execute.

We launched the new model six weeks ago. The data is early and messy and I'm going to share it anyway because I think the mess is the interesting part.

New signups dropped to 4,200 a month. Down 70% from 14,000. The VP of Marketing calls this "the trough." I call it "reality."

Onboarding completion: 74%, up from 59%. People who sign up for a trial with a deadline are more motivated than people who sign up for something that's free forever. Shocking, I know.

Thirty-day trial-to-paid conversion: 14.7%. That's 617 new paying customers, compared to the 98 we were getting under the freemium model. Net new MRR from conversions: $116,613 per month, compared to $18,522.

Server costs for the trial users: $8,200 a month, down from $34,000.

I want to be careful here. Six weeks is not enough time to declare victory. The thirty-day retention numbers look good — 81% — but we won't have ninety-day data until June. The cohort could fall off a cliff. We've seen it before.

Six weeks is not enough time to declare victory. The cohort could fall off a cliff. We've seen it before.

The VP of Marketing is already working on a new plan. She wants to build a "community tier" — a free account that gives access to our community forums, templates, and educational content, but not the product itself. The idea is to capture demand and build brand without the infrastructure costs. I think it's smart. I also think it's going to take six months to build and by then the market will have shifted again.

That's the thing about freemium. Everyone wants a formula. Free up to X users, or Y features, or Z gigabytes, and the conversions will flow. But the formula depends on your product, your market, your competitors, and a dozen other variables that change constantly. The model that worked in 2023 broke in 2025 because a competitor gave away for free what we were charging for. The model we've built now might break next year for reasons we can't predict.

All you can do is watch the numbers and be willing to change when they tell you something you don't want to hear.

The CEO asked me last week if I'd do freemium again, knowing what I know now.

"Yes," I said. "But I'd put a timer on it from day one. Freemium isn't a business model. It's an experiment. And experiments need end dates."

She wrote that down, which either means she agreed or she's saving it to use against me later. In this business, it's hard to tell.