Pricing a SaaS is a stage decision, not a model-catalog decision. The right answer at $0 MRR is not the right answer at $25K MRR. Pick your model by where your revenue sits today, anchor it to a value metric, and you will outperform the majority of solo founders who price by gut feel alone.
The fastest way to pick a SaaS pricing model (start here)
You have probably read three “six pricing models” articles and still do not know what to change on your pricing page. That is because those articles are written for funded teams with pricing committees, not for you at $5K MRR trying to ship a fix this afternoon.
Here is the one-line rule per band:
- Pre-revenue ($0): publish one simple paid plan. No free tier, no tiers at all.
- Early ($1-5K MRR): anchor a value metric and add a second tier above your current price.
- Traction ($5-50K MRR): layer usage signals, consider a third tier or hybrid, raise prices on new signups.
All three rules hinge on the same concept: the value metric. That is the unit your customer cares about that expands as they get more value out of your product. Lenny Rachitsky calls it “the thing that grows as the customer succeeds.” Get the metric right and the pricing model follows naturally. Get it wrong and no model saves you. The Stage-Gate Pricing Decision Grid below operationalizes this per band.
What “pricing strategy” actually means (and why model lists mislead you)
A pricing strategy is the principle you use to set a price: value-based (price on what the customer gains), cost-plus (price on what it costs you to deliver), or competitor-based (price on what the market charges). A pricing model is the mechanism by which you charge: flat rate, tiered, per-seat, usage-based, freemium, or hybrid.
Most solo founders confuse the two. They answer “how do I charge” before they have even answered “why this number.” That ordering is backwards and expensive.
Value-based pricing is the right strategy for nearly every solo SaaS. The problem is that most founders implement cost-plus by accident, because they price off what feels safe: typically “what I would personally pay” or “slightly less than the big competitor.” The gap between what something costs you and what a customer gains from it is the margin you are leaving on the table.
Two concrete examples. Slack's value metric is active users: pricing per seat captures the expansion as teams grow. An email tool's value metric is contacts reached: pricing per contact captures list growth. Neither starts from AWS cost per user. Both start from “what grows as the customer wins.”
Lenny's Newsletter on pricing strategy is the best public source on value-metric identification. Once you know the metric, the right model at each MRR band becomes obvious.
The Stage-Gate Pricing Decision Grid
Your pricing model decision changes at each MRR band. Here is the decision per band. (The broader stage-gated approach to growth decisions is covered in the AI SaaS growth playbook for solo founders, where pricing is one of the six levers reviewed per MRR stage.)
Pre-revenue / $0: one flat paid planGenerate signal
Anchor on the core outcome your product delivers. Do NOT add a free tier, multiple tiers, or usage metering yet. This week: publish one price and talk to 5 users about it.
Early / $1-5K MRR: two tiers maxFind the metric
Anchor on the specific unit that expands as customers succeed (seats, contacts, API calls). Do NOT use per-seat if single-player, and do NOT ship three or more tiers. This week: add a second tier 2x above your current price and watch take-rate for 30 days.
Traction / $5-50K MRR: two to three tiers or hybridCapture the tail
Use the same metric, now with a usage floor and overage trigger (base fee plus metered overage). Do NOT grandfather existing customers forever. This week: raise prices on new signups only and test a 20-30% increase.
Pre-revenue ($0).At zero MRR, your only job is to find out whether anyone will pay at all. A free tier delays that learning and costs you real money in support time and infrastructure. One simple paid plan forces the question: is this valuable enough to buy? Talk to five people about the price. If it sticks, you have a starting point. If everyone says “too expensive,” you have pricing data, which is more useful than a free-tier signup who never converts. (For the sequenced playbook on reaching those first paying customers, see how to get your first 100 SaaS customers.)
Early ($1-5K MRR). You have paying customers and real usage data now. Identify the metric that separates your power users from casual ones. That is your value metric. Build a second tier above your current price anchored to that metric. Two tiers give you a take-rate signal without drowning you in plan management. One caution: if your product is single-player, per-seat pricing punishes users for growing. Avoid it until the product is genuinely collaborative.
Traction ($5-50K MRR). You have enough usage variance to think about hybrid pricing. A small base fee plus metered overage aligns your revenue with delivery cost. On pricing for new signups: raise prices on new customers first. Do not grandfather your existing base forever. Every SaaS that never raises prices erodes its margin because costs compound. According to @saastr on X in September 2025, “SaaS inflation just hit 11.4% YoY.” That 11.4% hits your infrastructure bill whether or not you raise prices.
Pricing AI features without killing your margin
Flat pricing on AI features hides a margin cliff that does not show up until your top-decile users arrive. If you have shipped an AI feature inside a $50/month flat plan, run this math before you grow.
AI cost-basis worked example (rates as of early 2026: verify at api.anthropic.com/pricing before relying on them):
Claude Sonnet: approximately $3 per million input tokens, $15 per million output tokens. Typical request: 2,000 input tokens plus 1,000 output tokens.
Cost per request: (2,000 / 1,000,000 x $3) + (1,000 / 1,000,000 x $15) = $0.006 + $0.015 = $0.021 per request.
| Customer type | Requests/month | AI token cost | Plan fee | Margin |
|---|---|---|---|---|
| Median user | 200 | $4.20 | $50 | +$45.80 |
| Top-decile user | 2,000 | $42.00 | $50 | +$8.00 |
| Power user (top 5%) | 5,000 | $105.00 | $50 | -$55.00 |
A single power user at 5,000 requests loses you $55 on a $50 flat plan. At 100 such users that is a $5,500 monthly loss hiding inside what looks like a healthy customer count.
The median customer is profitable. The power user at 5,000 requests loses you $55 on a $50 plan. The tail is invisible until the usage distribution widens. By then, the flat fee is locked in and changing it means a pricing migration under fire.
As @aakashgupta observed on X in March 2026: “Anthropic just ran the SaaS pricing playbook so cleanly that most people won't notice.” What Anthropic did was price by consumption with a base tier, exactly the hybrid structure that prevents the margin cliff. When AI cost exceeds roughly 10% of plan revenue at median usage, hybrid pricing is warranted.
Per-token pricing means you charge based on the number of AI tokens consumed (input plus output), billed monthly. Stripe's metered billing supports this natively.
For implementation mechanics, see the usage-based pricing implementation guide. For a broader map of AI-native structures, see AI-native SaaS pricing models for 2026.
What founders actually struggle with (data from 224 discussions)
Across 224 public discussions on SaaS pricing strategy we analyzed on 2026-06-22 (Reddit, Hacker News, Stack Exchange, X), the recurring frustrations were not about pricing theory. They were about the gap between what guides prescribe and what founders are actually dealing with week to week.
The top recurring themes:
| Theme | Discussions mentioning | Share of 224 |
|---|---|---|
| B2B SaaS pricing decisions | 33 | 15% |
| Free trial vs freemium tension | 24 | 11% |
| Early-stage pricing uncertainty | 22 | 10% |
| Which pricing model to pick | 21 | 9% |
Free trial and early-stage uncertainty together account for 21% of the discussion volume. These are founders trying to make a call this week with incomplete data. Not founders debating expansion revenue optimization.
Pricing a niche SaaS is hard, especially while you're still figuring out who loves it.
That is the real condition: not “optimize expansion revenue” but “I do not know yet who my best customer is.” The full post captures the early-stage bind precisely.
The Stage-Gate Grid addresses this directly. At pre-revenue and early stage, your goal is not optimizing a pricing model. It is generating signal. One flat paid plan generates more useful data than a freemium funnel: every interaction is a conversion decision, not a usage metric you must interpret. Once you know who loves it ($1-5K MRR band), anchor the value metric and build tiers around what that customer actually does.
Common pricing mistakes at the solo-founder stage
The most expensive pricing mistake at the solo-founder stage is pricing too low. Low prices signal low value and starve your runway: you need more customers to hit the same MRR, which means more support load and less time on product. I have watched founders agonize over whether to charge $19 or $29 when the real question was whether they should be charging $79.
Four more mistakes that compound quickly:
Too many tiers too early. Three or four tiers before $5K MRR creates support complexity and buyer decision fatigue. Two tiers max until take-rate data tells you what a third tier should actually contain.
Copying a competitor's price without their cost basis. A funded competitor at $29/month may be subsidizing with venture capital. Your cost basis is different. Benchmark competitors to sanity-check your range, not to set your number.
Flat-pricing AI features. If AI token cost exceeds 10% of plan revenue at median usage, move to hybrid before the usage distribution widens into the power-user cliff shown above.
Never raising prices. SaaS inflation is real. @saastr reported 11.4% YoY in September 2025. A founder who sets $49/month at launch and never revisits it is effectively cutting prices every year. Raise prices on new signups first; migrate existing customers with 60 days notice.
Freemium before product-market fit. Freemium works when you know your conversion path, activation metric, and support cost per free user. Before $5K MRR, you probably know none of these clearly. A 14-30 day free trial is the better default. Mispriced plans also drive churn once customers realize the value exchange is off; the SaaS churn rate benchmarks for 2026 show what normal looks like by ACV tier, so you can separate pricing-driven churn from product issues.
FAQ and next step
What is SaaS pricing? SaaS pricing is the set of decisions that determine what you charge, how you structure those charges, and what principle anchors the number. It covers strategy (value-based, cost-plus, or competitor-based), model (flat, tiered, per-seat, usage-based, or hybrid), and value metric (the unit that grows as the customer succeeds).
What is the best SaaS pricing strategy? Value-based pricing is the right strategy for nearly every solo SaaS. It anchors price to customer outcome, not your internal cost, so your revenue scales with the value you deliver. The practical challenge is identifying the value metric for your product. Once you have it, the Stage-Gate Grid tells you which model fits at each MRR stage.
What is the difference between a pricing model and a pricing strategy? Strategy is the principle you use to SET a price (value-based means you price on what the customer gains). Model is the mechanism by which you CHARGE it (tiered means multiple plans at different price points). Most solo founders pick a model before committing to a strategy, which is why they revisit pricing every six months without making durable progress.
How do I price an AI SaaS or AI features?Use flat pricing only if usage variance is narrow (top users consume less than 10x more than your median). If variance is high, which is common for AI features, use hybrid: a base fee plus metered overage above a usage threshold. Stripe's metered billing and Anthropic's API both support this natively. Run the margin-cliff math from the section above with your own usage distribution.
Should I offer a free trial or freemium? Before $5K MRR: free trial, not freemium. A 14-30 day trial forces the conversion decision; freemium defers it and adds support load per free user. Above $5K MRR: test freemium only if trial-period usage data reliably predicts conversion for your product.
What is a value metric and how do I pick mine?A value metric is the specific unit your customer cares about that grows as they get more value. To identify yours: look at what your best-retained customers do in the product that casual users do not. That differential behavior points to the metric. Slack's metric is seats; an email tool's metric is contacts. Name that unit and the Stage-Gate Grid tells you which model follows.
The SaaS pricing page optimization guide covers how to structure your pricing page once the model is set. For the downstream unit-economics impact of your pricing tier decisions, the LTV/CAC ratio guide shows how price points propagate into your payback period and growth capacity.
Subscribe to the SaasFlywheel newsletter for one pricing teardown every Friday.