You shipped. The product works. You sat down to pick a price, opened three competitor pricing pages, and did the thing that felt obviously correct: a cheap tier around $9, a middle tier around $29, a "pro" tier around $99. Charm-priced, three clean options. Done.
I've watched a lot of founders make exactly that move. It's the most expensive safe-looking decision on the page.
Thirty years of building systems taught me one thing that applies far outside code: the dangerous mistakes never look dangerous. They look like the sensible default everyone else already picked.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🧩 First — the default really is "correct"
Let's be fair to the ladder. The $9/$29/$99 structure isn't something founders invent in a panic — it's the textbook. The standard template is an entry tier at $9–29, a growth tier at $49–99 as the "main revenue driver," and enterprise above $199. The 9-endings aren't superstition either: prices ending in 9 reliably outperform round numbers, because your brain reads $29 as "twenty-something," not "thirty." The left-digit effect is one of the most replicated findings in pricing.
So this isn't a story about a dumb mistake. It's a story about a smart-looking default that happens to be wrong for one specific person: the founder who has shipped, but can't yet get users.
That person is probably you. So let me show you the trap from the inside.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🪤 The trap, part one: A low price is a bet on volume you haven't earned
Here's the arithmetic almost nobody runs before they pick the number.
🔴 At $9/month, hitting $5K MRR takes roughly 555 paying customers.
🟢 At $99/month, the same $5K MRR takes about 51.
Same revenue. One path asks you to find and convert eleven times more humans.
Now the honest question: which problem do you actually have? If you're reading this, it almost certainly isn't "people think I'm too expensive." It's "nobody's showing up." A low price is a bet that you'll win on volume — but volume is distribution, and distribution is the exact muscle you haven't built yet. Pricing cheap doesn't solve that. It doubles down on the one thing you can't do.
The way out runs the other direction: narrow your audience until your product is the obvious choice for a specific group. Specificity creates pricing power; generality forces you to compete on price. ConvertKit launched at $29 into a market where MailChimp charged $10 — it didn't compete on price, it competed on being "email for creators," and grew to roughly $30M ARR.
Think of a cheap price as a stall selling bottled water at 50¢. To make rent you need a thousand people to walk past. Your real problem is that the street is empty. A low price doesn't fill the street — it just means the few who do walk by are barely worth serving.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
💸 The trap, part two: cheap recruits the customers who leave
Low prices attract customers who cost more to support than they bring in — and a low number quietly signals low quality. The cleanest proof is counterintuitive: when Bannerbear raised its price from $9 to $49, churn went down. The dabblers left; the serious users stayed and paid more.
Read that twice. The cheap tier wasn't bringing in your best customers. It was bringing in the ones most likely to kick the tyres, file the most support tickets per dollar, and vanish by week three. Teardowns show the same shape: founders who cut tiers and raised the entry price saw revenue per user climb ~40% while churn dropped from over 12% to under 8%.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🤖 The trap, part three (the new one): if your product uses AI, the cheap tier loses money on your best users
This is the part most founders haven't internalised yet, and it's where 2026 rewrites the math.
Traditional software has near-zero marginal cost per user — that's the whole reason SaaS prints money at scale. AI doesn't work that way. Every call runs the model again and burns real cash. Per ICONIQ's 2026 data, about $230,000 of every $1M in AI product revenue walks out the door as inference cost before anyone gets paid — the old 80% gross margin is becoming the outlier, not the rule. And usage isn't even: a power user can cost 100x a light user while paying the identical flat fee.
You don't have to take my word for the danger. Watch the biggest player in the room flinch. On June 1, GitHub moved Copilot off flat subscriptions to usage-based billing, openly citing unsustainable costs from rising AI usage. The fallout was immediate — developers reported bills jumping from ~$29 to nearly $750, and in one case ~$50 to ~$3,000.
If Microsoft can't make a flat low AI subscription survive its heavy users, your $9 AI tier won't either. The cheapest plan attracts the most cost-insensitive usage — and you eat the loss on every power user, the exact users you most wanted to keep.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔧 The fix: price the value, charge more to fewer
Under uncertainty, the instinct is to reach for cost-plus — add up your infrastructure, tack on a margin, divide by users. That's backwards. Real pricing starts from the value the product creates, not the cost to build it.
For a founder staring at a flat user graph, the reframe is simple and a little uncomfortable: you can't out-distribute the market yet, so each customer has to be worth more. Fewer people, higher value, a price that reflects it. The rule of thumb is to anchor your price to 10–20% of the value delivered. One founder priced an automation tool at $9 because it "felt fair for something simple" — a user told them it saved $300/month. That's charging 3% of the value created.
And the fear that keeps founders cheap — "I can't raise it later" — isn't true. There's a simple three-move way to raise prices without losing the customers you already have:
🔒 Let your current customers keep their old price. "Grandfathering" just means everyone already paying you stays on the price they signed up at. Paying $29 today? They keep paying $29. They feel rewarded, not punished — so they don't leave.
📈 Charge the higher price to new signups only. Your existing revenue doesn't move at all. Only people who join after the change pay more.
🎁 Ship one real feature first, then announce it. The story becomes "we got better, so the price went up" — not "we got greedy."
Here's why this is close to free, with real numbers. Say you charge $29 and get 100 new signups a month — that's $2,900/month in new revenue. Now you raise the price 30%, to about $38. Even if the higher price scares some buyers off, you only need 77 of those 100 to still sign up to match the old $2,900 — because $2,900 ÷ $38 ≈ 77. In other words, you can lose up to 23 out of every 100 new customers and still break even. Anything better than that is pure profit. And your existing customers are grandfathered, so they aren't going anywhere. That's what "costs you nothing if churn stays flat" means: the downside is tiny, the upside is real.
✅ Save this: the 5-question pricing gut-check (before you commit to a number)
1. Volume reality. How many paying customers would this price actually need? Take your monthly revenue target and divide it by your price. Want $3,000/month at $9? That's 333 customers ($3,000 ÷ $9). At $49, it's just 61 ($3,000 ÷ $49). Now be honest: have you ever gotten that many people to do anything — sign up, buy, even join a free email list? If 333 feels impossible but 61 feels doable, your price is too low.
2. Problem match. A low price only helps if people are already visiting your page and leaving because it feels too expensive. It does nothing if your real problem is that almost nobody is visiting at all. So which is it — are people showing up and saying "too pricey," or are they just not showing up? Don't cut your price to fix a traffic problem.
3. Customer quality. Would tripling the price drive away buyers I actually want — or just tyre-kickers?
4. AI margin. At this price, what does my heaviest user cost me to serve? If I lose money on power users, the price is structurally broken.
5. Value anchor. What's the single concrete outcome my product creates, in dollars? Am I charging 10–20% of that — or 3%?
If you can't answer #1 and #5 with real numbers, you're not pricing. You're guessing and calling it a strategy.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
💬 One question back to you
I'm building in public, and pricing is one of the decisions I'm chewing on for my own product right now — so I'll ask you the same question I've been asking myself:
When you picked your price, were you solving the problem you actually have — or the one the template assumed you had?
Hit reply and tell me the number you landed on, and why. I read every one.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Sources
• Calmops — SaaS pricing models for solo founders
• Prompts to Product — Solo founder pricing playbook
• Freemius — Micro-SaaS pricing strategies (Bannerbear $9→$49)
• Freemius — Micro-SaaS pricing pages that convert
• SaaS Mag — AI COGS & gross margin compression (ICONIQ 2026)
• Monetizely — Economics of AI-first B2B SaaS in 2026
• GitHub Blog — Copilot moving to usage-based billing
• Dataconomy — Copilot token-pricing backlash
• AuditX — Why indie hackers fail at pricing
• Dev.to — The pricing playbook for indie developers