Your free users are not your customers

TL;DR: Less than 4% of our stores generate nearly 40% of revenue. Most free-tier users are hobby stores with zero traffic, consuming support in exchange for nothing. Freemium still works — but treat the free tier as a marketing budget, not a customer segment.

Theo Browne published a video called AI Has a Subsidization Problem that made me pause. His argument: companies giving away free AI inference are attracting users who will never pay, who cost more to support than paying users, and who filter your user base down to the least valuable segment.

He’s talking about AI. But he could be talking about every freemium app on the Shopify App Store.

I’ve run RevenueHunt for over five years. We have 20,000+ stores installed. We rank #1 in our category. And the data tells a story that every Shopify app founder running a freemium model needs to hear: the vast majority of your installs will never make you a dollar, and they’ll cost you more than the ones that do.

Key takeaways:

• Less than 4% of our stores (Shopify Plus) generate nearly 40% of our revenue
• The largest segment (~33% of stores) generates under 10% of revenue with a CLV of $21
• Free and basic users consume more support resources per user than paying customers
• Most free-tier stores are “coming soon” pages run by people treating entrepreneurship as a hobby (or therapy)
• If you can’t segment your ad spend by customer quality, you’re paying to acquire users who will never convert
• Usage-based pricing is a natural filter — the stores that drive real traffic upgrade themselves

The numbers nobody wants to show

Here’s what our revenue breakdown looks like by Shopify plan tier:

Segment	% of stores	% of revenue	CLV
Shopify Plus	~4%	~39%	~38x higher than Basic
Growth plans	~10%	~24%	~8x higher than Basic
Basic	~33%	~9.5%	Lowest
Free/trial/development	~10%	~0%	Zero

Read that again. A third of our installed base generates less than 10% of our revenue. The top 4% generates almost 40%.

This isn’t unusual. This is the shape of every freemium SaaS. But most founders don’t look at it this way because the vanity metric — total installs — feels good. “20,000 stores use our app” sounds better than “about 1,300 stores pay us.” Both are true. Only one matters.

The free user tax

Theo makes a point that hit home: free users are often more expensive to support than paying users, not less.

The $0/month client asks endlessly for your time and attention. The $299/month client says “looks good, thumbs up.” This isn’t a meme. It’s our daily support reality.

Our free-tier users generate the most tickets per user. Not because the product is complicated — because their store isn’t ready. Products aren’t organized into collections. The store isn’t live yet. Traffic is zero. They’re obsessing over details — button colors, font choices, border radiuses, transition animations — for a quiz that gets zero traffic.

Meanwhile, our Shopify Plus merchants rarely contact support. When they do, it’s a specific technical question with context. They have a dedicated person handling conversions — usually someone sharp, whose entire job is to optimize the funnel. They install the quiz, configure it, drive traffic with paid ads, and the usage grows organically. They don’t need hand-holding because they know what they’re doing.

The support cost per dollar of revenue between these two groups isn’t close.

Entrepreneurship as a hobby (or therapy)

Here’s my theory about why the free tier looks the way it does.

Most of the stores on our Basic and free tiers are not generating revenue. They’re “coming soon” pages. Stores under construction. Stores with 12 products, no traffic, no ads, no social presence. Stores that launched six months ago and have processed zero orders.

The economics are clear: these stores install our app, configure a quiz nobody will ever take, file support tickets about edge cases that will never matter, and generate zero quiz responses. They cost us support time and server resources in exchange for nothing. Not because they’re bad people — because their store isn’t at the stage where a product recommendation quiz adds value.

A store with no traffic doesn’t need a quiz. It needs traffic. A store with no products organized into collections doesn’t need conditional logic. It needs a catalog. We’re not the priority, and we shouldn’t be. But as long as the tool is free, people will install it anyway — the same way you’d grab a free sample at the supermarket even if you’re not buying.

We recently moved custom CSS and custom JavaScript behind the paywall. These features were free for years. They generated zero additional revenue — but they were the single biggest driver of support tickets. Free-tier users with no traffic and no revenue would spend hours tweaking border radiuses, requesting fonts we don’t ship, and asking our support team to debug their custom code. They had the time because they didn’t have real problems to solve. No traffic to optimize. No conversion funnel to fix. Just pixel-perfect styling for a quiz that three people would ever see — all of them the store owner.

So we made the call: if you want to customize at that level, pay for it. The support volume on those features dropped immediately. The users who actually needed custom CSS — the ones with real stores and real traffic — upgraded without complaint. The ones who were using it as a creative outlet moved on.

And here’s the thing: if setting up a Shopify store is a hobby — and for a lot of these stores, the data says it is — that’s fine. But you pay for hobbies. You want to do jiu-jitsu? You get one free class, then you pay. You want therapy? You pay your therapist. Somehow in SaaS, we’ve normalized the idea that serious business tools should be free forever for stores that aren’t running serious businesses. The free tier subsidizes stores that will never generate a dollar — not for them and not for us.

Why we couldn’t fix this sooner

I always knew the return on ad spend for basic-tier users wasn’t there. The problem was I couldn’t do anything about it.

Before we earned the “Built for Shopify” badge, Shopify’s ad platform didn’t let us segment by merchant plan tier. Our ads in the Shopify App Store were a blank slate — we paid per click, and since the majority of Shopify stores are on the Basic plan, the majority of our clicks were Basic merchants. We were paying to acquire the customers least likely to ever pay us.

The only lever I had was timing. I paused ads on weekends because people working on their stores on Saturday and Sunday are probably people with a real job Monday through Friday. Their store is a side project. They’re not the merchant with a dedicated conversion specialist — they’re the merchant doing everything themselves at 11 PM. Not our target.

We unlocked plan-tier segmentation last summer when we got the “Built for Shopify” badge. Now we can target ads specifically at Shopify Plus and Growth merchants. The ROI difference is night and day.

Usage-based pricing as a natural filter

Our pricing is usage-based. The more quiz responses a store gets, the more they pay. This sounds simple, but it creates a powerful self-selection mechanism.

A store with no traffic generates no quiz responses. They stay on the free tier forever. They cost us support resources and server time, and they contribute nothing.

A store with real traffic — the kind that comes from paid ads, organic search, social, email campaigns — generates quiz responses naturally. Their usage grows. They hit the free tier limit. They upgrade. The pricing model rewards stores that are actually running a business.

Shopify Plus stores tend to have dedicated people — a CRO specialist, a growth marketer, someone whose job is funnel optimization. That person installs our quiz, publishes it, drives traffic to it, and watches the conversion data. When the quiz works (and it usually does at that scale), usage grows and the plan upgrades itself. We don’t need to sell them. The product sells itself to stores that are ready for it.

The stores that aren’t ready — the coming-soon pages, the hobby stores, the side projects — never generate enough usage to hit a paid tier. The pricing model filters them out automatically. The problem is they still consume support and infrastructure in the meantime.

What Theo got right

Theo’s spectrum — from “I can only use it if it’s free” to “I’ll try it if it’s free” to “I’ll pay for the better thing” — maps directly to the Shopify App Store.

The gold mine is the range between “I’ll try it because it’s free” and “I’ll pay because it works.” These are the merchants who install the quiz on the free tier, see it convert, and upgrade because the ROI is obvious. We have hundreds of these. They’re our best customers.

The death zone is everyone left of “I can only use it if it’s free.” These are the merchants who will never pay — not because the product isn’t good enough, but because their store doesn’t generate enough revenue (or any revenue) to justify any paid tool. Making the product better doesn’t convert them. Nothing converts them. They need traffic and revenue first, and that’s not our problem to solve.

The mistake Theo identifies in AI companies — attracting millions of free users who will never pay and then being stuck serving them — is the same mistake every freemium SaaS makes at some point. You celebrate the install numbers. You staff up support to handle the ticket volume. You optimize onboarding for users who don’t understand the basics. And then you look at the revenue and realize you’ve been building for the wrong audience.

Why freemium is still worth it

After everything I just said — freemium is still our best marketing tool. I wouldn’t change it.

The free tier is how we got to #1 in the Shopify App Store. The “Powered by” badge on free quizzes was our early growth engine. And here’s the part that makes the math work despite the support cost: free users who play around with the product get hooked.

They install because it’s free. They build a quiz because it’s easy. They publish it. Some of them start exploring the advanced features — conditional logic, custom result pages, integrations. They hit the paywall on a feature they now want. And because they’ve already invested time building their quiz, customizing the design, mapping products to outcomes — the cost of switching to a competitor is higher than the cost of upgrading.

That’s the real edge of freemium. It’s not just distribution. It’s earned switching cost built during the free trial. Every hour a merchant spends configuring their quiz is an hour they’d have to spend again if they moved to a competitor. The free tier lets them invest that time without risk. The investment is what converts them.

So yes, the free tier is a cost. Most free users will never pay. The support burden is real. But the ones who do convert come in pre-committed — they’ve already built something inside your product that they don’t want to rebuild. Freemium isn’t charity. It’s a trap door with a very comfortable floor.

The key is understanding the free tier as a marketing budget, not a customer segment. Budget for the support cost. Don’t optimize the product for users who will never pay. And be deliberate about which features sit behind the paywall — the features that create switching cost should be accessible enough to try, expensive enough to keep.

FAQ

Should I remove my free tier?

Probably not. Freemium is still the best distribution model for marketplace apps. But understand that your free tier is a marketing cost, not a customer segment. Budget for it accordingly — and don’t optimize your product, support, or ads for users who will never pay.

How do I know which free users will convert?

Usage. A free user who completes 50 quiz responses in their first week has traffic and will likely upgrade. A free user who installs the app and never publishes a quiz will never convert. Early usage velocity is the strongest signal.

Is this just survivorship bias?

No. It’s the data. The CLV difference between our top segment and our bottom segment is 38x. That’s not a slight variation in customer quality — it’s a completely different business happening inside the same product.

What about the “12-year-old who becomes a billionaire” argument?

Theo addresses this too. Yes, some free users today will be valuable later. But you can’t subsidize millions of non-paying users on the hope that a few of them will eventually become Shopify Plus merchants. Run the math on what your free tier actually costs you in support and infrastructure, then decide if that bet makes sense.

---

Theo’s full video: AI Has a Subsidization Problem — start at 18:13 for the spectrum breakdown. Related: Distribution is the only moat left for why freemium works for growth, and How a 5-person team uses AI for how we handle support with a small team.