An early-stage startup picks its pricing from 4 models: cost-plus (cost + margin, the accounting baseline), value-based (price anchored to the economic value created for the customer, ideal for B2B SaaS), tier-based (2 or 3 plans for a market with established conventions), and usage-based (billing by event, by seat, or by API call). That last model already accounts for 20% of SaaS in production according to the Maxio 2025 report, and its share climbs every quarter. The right model depends on your segment, the market's maturity, and the value metric your product moves.
To set your first price without data, the simplest protocol is the pricing power test across 10 sales: if nobody says no, you're underselling; if more than 50% say no, you're overreaching. The sweet spot sits between 20 and 40% rejection. To calibrate your starting range, 6 public pricing pages from French SaaS companies (Lemlist, Pennylane, Aircall, Akeneo, Lago, Spendesk) give you concrete reference points for each model. Pricing is a decision you make, not the thing left over after you've built the product.
Why most founders get their first price wrong
September 2023. A team walks out of Demo Day with a product that works. The first customer signs at €99/month. Why 99? "Because that's what the competitor charges." Six months later, they find out that competitor loses money on every account.
Pricing is one of the rare early-stage decisions where most founders wing it. Three traps account for the bulk of the mistakes.
The "I copy the competitor" trap
Copying the competitor means adopting their cost structure, their margins, their rookie mistakes, and their market constraints without actually understanding any of it. The competitor you're watching may be sacrificing pricing to grow fast, raising rounds to cover the monthly deficit. You have no way of knowing.
Their price tells you something about the market's psychological ceiling, not the right price for you. Calibrate against their range, never against their exact number.
The "I mark up my cost + 30%" trap
Cost-plus guarantees your margin. Its flaw: it ignores the value the customer gets from the product. A tool that generates an extra €50k of pipeline per month for a team of 5 sales reps shouldn't cost €99/month just because it costs €68 to produce. The cost + margin formula works for commodities. For a differentiated B2B SaaS, it consistently leaves value uncaptured.
The "I'll update it later" trap
"Provisional" pricing has a shelf life: your first 10 customers become the benchmark for everyone who follows. Early adopters talk to each other. Any price increase turns into a churn event rather than a natural progression, because you anchored your reference point too low from day one. Starting low to "validate" often costs more than a solid pricing decision made up front.
The 4 pricing methods (and when to use which)
1. Cost-plus: cost + margin, the accounting baseline
You calculate the full production cost (infra, support, proportional overhead) and add a target margin between 30 and 70% depending on the sector. It's the default model for agencies, managed services, and low-differentiation products.
When to use it. A mature market with established pricing conventions, or when you don't yet have the data to calculate the value created for the customer. It's a starting point, not a long-term strategy.
Critical limitation. Your production cost will drop (cheaper infrastructure, automation). If you're anchored to cost-plus, you don't capture the additional value created. You underprice constantly as your product improves.
2. Value-based: cost avoided or revenue created for the customer, the B2B SaaS ideal
You start from the economic value your product generates for the customer (time saved, churn reduced, additional revenue) and set the price at a fraction of that value. The rule: 10 to 30% of the measurable economic value. If your product sells self-serve rather than through a sales team, your pricing is part of a product-led growth strategy, where the pricing page itself does the conversion work.
When to use it. When you can measure the impact (ROI, hours saved, leads generated). It's the dominant model in B2B SaaS: your software replaces a manual process or improves a conversion funnel.
Why founders avoid it. It takes deep conversations with customers to quantify the value. That's uncomfortable. The result: most founders default to cost-plus. And that's exactly where the margin hides.
3. Tier-based: 2-3 plans, when the market already has conventions
You offer 2 or 3 distinct plans with different features or usage limits. The middle plan is calibrated to be chosen by the majority (classic anchoring). Past 3 plans, the decision gets more complex for the buyer. Below 2 (flat rate), you lose the anchoring effect.
When to use it. When the market already understands this model (most B2B SMB SaaS), when you have several ICP profiles with distinct budgets and usage patterns.
4. Usage-based: by event/seat/API call, the 2026 shift
The customer pays based on what they consume: emails sent, active seats, API calls, contacts enriched. 20% of SaaS companies already used a pure usage-based model in 2025 according to Maxio, up steadily since 2023. The model has gone mainstream with tools like Lago that let you implement it without a dedicated engineering team.
When to use it. When value is directly correlated with usage (infrastructure, data, enrichment). When you're targeting enterprise accounts that want to control their budget and avoid fixed commitments.
The trap. Usage-based makes revenue less predictable. Implementing a monthly minimum (floor) is essential to secure your base MRR.
How to set your first price when you have no data
The "pricing power" test across 10 sales (0% rejection = too low, >50% = too high)
The protocol comes down to 4 steps: set a first price, pitch 10 qualified prospects within your ICP, count the price-related rejections (not product, not timing), and adjust in 20% increments.
- 0 to 20% rejection: you're underselling. Raise it 20 to 30%.
- 20 to 40% rejection: the sweet spot. You maximize volume without sacrificing margin.
- More than 50% rejection: you're overreaching or targeting the wrong segment. Lower the price or reframe the ICP.
A secondary signal: if everyone accepts without negotiating, you don't just have a price problem, you have a positioning problem. A well-positioned product generates objections. A complete absence of price objections often signals a fuzzy perceived value.
The 10x rule: minimum price = 1/10 of the value generated
If your tool saves a sales rep 5 hours a week, and that rep's loaded cost is €6k/month (€37.50/hr), you're generating roughly €2,400/month in time value. Your price floor: €240/month. Below that, you leave margin on the table. Above €720/month (30% of the value), you'll hit ROI resistance at every close.
The 10x rule applies to B2B SaaS where value is measurable and recurring. For B2C or a commodity, the mechanics are different.
Should you display the price or ask for a quote?
Display it if you're targeting SMBs: short cycle, individual decision, no multi-level sign-off. Switch to "request a quote" if you're targeting enterprise: long cycle, customization, multi-stakeholder negotiation.
The hybrid structure, the most common in B2B SaaS: SMB plans displayed + an Enterprise plan on request. It captures the self-serve bottom of the market without sacrificing flexibility at the top. Practical tip: if you're on the fence, display first. The absence of a price on a B2B SMB site drives bounce.
6 public pricing pages from French startups (worth studying)
Rather than theorize, let's look at what SaaS companies do in production. Six examples, six different models.
Lemlist: 3 per-seat tiers
Lemlist (B2B cold outreach) offers 3 plans (Standard, Pro, Outreach Scale) with per-seat pricing. Each tier unlocks additional automation features: multichannel sequences, A/B testing, advanced CRM integrations. The middle plan is calibrated to convert teams of 2 to 5 sales reps.
What's instructive: Lemlist started flat rate, then migrated to tier-based as customer usage diverged. The decision was data-driven, not theoretical. Source: lemlist.com/pricing
Pennylane: per-user + modules
Pennylane (French accounting) combines a per-user price with add-on modules. The hybrid model reaches solopreneurs (1 user, minimal modules) and accounting firms (multi-user, advanced modules) with one coherent pricing page. The per-user + modules approach works when usage varies widely across customer profiles. Source: pennylane.com/fr/tarifs
Aircall: per-seat with an annual commit
Aircall (cloud telephony) lists 3 plans (Essentials, Professional, Custom) with per-seat pricing and a mandatory annual commit. That commit is a deliberate choice: it reduces monthly churn and improves MRR predictability, at the cost of friction at signup. If your sales cycle runs longer than 30 days, an annual commit can make sense even at the early stage. Source: aircall.io/pricing
Akeneo: value-based enterprise (price on request)
Akeneo (enterprise PIM) doesn't display a price: a contact form replaces the pricing page. This isn't an oversight, it's a positioning choice. At this level of the market (five-figure annual enterprise tickets), value varies enough by product catalog, channels, and integrations that a displayed price would be counterproductive. "Price on request" signals that the product sells consultatively, with ROI calculated customer by customer. Source: akeneo.com/compare-packages
Lago: usage-based open-core
Lago (open-source billing infrastructure) offers 3 plans: Self-Hosted Free (free), Cloud Premium (usage-based on Lago Cloud), and Enterprise (custom). The free self-hosted tier is an acquisition strategy: teams that outgrow their DIY capacity migrate to Cloud or Enterprise. It's the open-core model in its purest form. You give away the core value to create functional dependency, then monetize on convenience or complexity. Source: getlago.com/pricing
Spendesk: per-seat B2B finance
Spendesk (B2B corporate spend management) shows how a finance SaaS aligns its pricing with team size: a per-seat model with tiers differentiated by feature (virtual cards, advanced reporting, ERP integrations). What's instructive: per-seat in the finance vertical grows ARR as the customer's team grows, without renegotiating the base contract. Expansion revenue is automatic.
When and how to raise your prices
The 3 signals to raise (margin, demand, added value)
Three situations justify a price increase, often at the same time:
- Margin signal: your gross margin drops below 60-65% (the B2B SaaS benchmark). Pricing has to make up for it.
- Demand signal: you've been closing 80% or more of your deals with no price friction for 3 months. The pricing power test is telling you you're in undervalued territory.
- Value signal: you've shipped a significant feature or proven measurable ROI you weren't charging for. Value went up, price follows.
When all three land at once, the increase is urgent.
How to announce an increase without massive churn
The minimal protocol for a clean increase:
- Notice: 45 to 60 days before it takes effect.
- Justification: one sentence on what changed (a major feature, increased value). No apology, no "sorry for the inconvenience."
- Long-term opt-in: offer to lock in the current price by switching to annual before the deadline.
Churn on a price increase rarely comes from the amount. It comes from the communication: too short, too abrupt, no explanation. Customers understand the value if you explain what they got.
Grandfathering: what to offer early customers
Grandfathering (keeping the historical price for existing customers) creates an "early adopter" status that drives loyalty and referrals. Its limit: if your early customers represent a significant slice of your ARR at a discounted price, grandfathering becomes a CAC/LTV problem as you scale.
Practical rule: grandfather for 6 to 12 months max, then a gradual increase with notice. A common middle-ground option: grandfather the current plan, make new pricing mandatory on upgrades.
FAQ: Startup pricing
Should you do freemium at the early stage?
Freemium has a hidden cost: you serve free users who will never convert, which weighs on support, infra, and product focus. Before solid product-market fit, freemium scatters your energy. If you can't pinpoint exactly what trigger turns a free user into a paying one, and on what timeline, avoid freemium. A time-limited free trial (14 days) is often more effective at the early stage: it creates urgency without generating permanent costs.
How do you price a B2B product vs. a B2C one?
In B2B, price is a rational decision: ROI, budget, sign-off cycle. You can justify a high price if you quantify the value created. In B2C, it's an emotional and comparative decision, where the reference point is the competitor's price on the App Store. A B2B SaaS at €500/month is unremarkable. A B2C SaaS at €50/month is already seen as expensive. The acceptable models, the tolerable margins, and the decision cycles are fundamentally different: don't carry your reference points from one segment to the other.
Public pricing or on request for Enterprise?
If your average ticket exceeds €12k annually, on-request pricing is often the better fit: it lets you customize by scope and manage multi-level negotiations without boxing yourself into public pages that weaken your position in big deals. Below €12k annually, public pricing speeds up self-serve and lowers your acquisition cost. The hybrid structure (SMB displayed + Enterprise on request) addresses both segments without sacrificing one for the other.
How many pricing plans should you offer?
Two or three. Two if your market splits into two clear profiles (startups vs. large accounts). Three if you have an intermediate segment to capture, with the middle plan designed as the "default" option that 60% of your customers will pick. Past three plans, choice paralysis reduces your conversion rate. Below two (flat rate), you lose the anchoring effect and the natural upsell path to the higher plan.
Want to go deeper on your GTM model: find your first customers or see how to validate an idea before you build. If you're torn between two pricing models for your next enterprise customer, that's exactly what we work through 1:1 at swanbase.
