Key Takeaways
- 3 tiers, always. 2 leaves money on the table; 4+ paralyses buyers. The canonical structure: Starter / Pro / Enterprise.
- Don't price at $97. It signals "consumer info product" to B2B buyers. $149 or $199 converts better in most categories.
- Per-seat for collaborative tools, per-usage for APIs, flat-rate for solo-user products. The wrong choice leaks 20-40% of revenue at scale.
- Raise prices when: churn under 2% monthly · 30%+ of demos skip price questions · customers max-out their tier · competitors have moved.
- Grandfather existing customers at the old price for 12 months when you raise. New signups pay the new rate from day one.
SaaS pricing is the single biggest leverage point in any subscription business, and it's the one most founders get wrong on the first attempt. Underpricing destroys the unit economics permanently; overpricing without proof slows acquisition to a crawl. Get pricing right and you can ship a worse product profitably; get it wrong and a great product still loses money.
Why pricing is the #1 lever
Pricing is the only growth lever that costs zero engineering hours, ships in a day, and compounds forever. A 10% price increase delivered cleanly to your existing book typically retains 95%+ of customers — meaning a 9.5% net ARR lift before you've added one new customer. No other lever comes close.
How many SaaS pricing tiers should you have?
Three. Always three. Two tiers leave money on the table because power users can't self-select into a higher plan. Four+ tiers paralyse buyers and create support overhead.
| Tiers | Outcome | Why it fails |
|---|---|---|
| 1 tier | Flat pricing, no expansion path | You leave 30-50% revenue with power users on the table |
| 2 tiers | Starter + Pro | No high-revenue ceiling for enterprise buyers |
| 3 tiers | Starter + Pro + Enterprise | Canonical structure. Use this. |
| 4+ tiers | Multiple intermediate tiers | Decision paralysis cuts conversion 15-25% |
Designing the 3-tier structure
Each tier should be obviously different. The wrong pattern is "Pro has more of everything Starter has." The right pattern is "Starter is for one person, Pro adds team features that one person never needs."
| Tier | Persona | Price anchor | Defining features |
|---|---|---|---|
| Starter | Individual / solo user | $19-49/mo | Core product · 1 seat · basic integrations |
| Pro | Team of 3-15 | 3-5× Starter ($79-199/mo) | Collaboration · admin · advanced integrations · API access |
| Enterprise | 50+ employees | Custom (typically $20K-100K/yr) | SSO · SCIM · audit logs · custom SLA · CSM |
For each tier, ask: "Could a customer at this tier do their job at the tier below?" If yes for any tier pair, your tiers aren't differentiated enough. Add or remove features until the answer is no.
The $97 anchor mistake
Don't price your middle SaaS tier at $97. The $97 price point became famous in the 2000s online-course world for a reason — it's a psychological anchor that signals "consumer info product." For B2B SaaS, $97 communicates "small, undifferentiated, easy to switch away from."
The same product priced at $149 or $199 gets more trial signups in most categories, not fewer, because the higher price implies higher quality. This is called price-quality inference — buyers use price as a heuristic for quality when they can't directly evaluate the product.
Higher if you're differentiated. Lower only if you're trying to compete on price (a strategy that rarely works for SaaS — there's always someone cheaper).
Per-seat vs per-usage vs flat
| Model | Best for | Expansion path | Example companies |
|---|---|---|---|
| Per-seat | Collaborative tools | Each new team member adds revenue | Slack, Notion, Figma, Linear |
| Per-usage | APIs, infrastructure | Volume scaling with customer growth | Twilio, Stripe, AWS, OpenAI |
| Flat-rate | Solo-user tools | Limited; mainly tier upgrades | Basecamp, Carrd, Tally |
| Hybrid | Mid-market multi-product | Both seat + usage | HubSpot, Intercom, Mixpanel |
Default for B2B SaaS: per-seat for Starter and Pro, custom annual contract for Enterprise. Default for API-first products: per-usage with a free tier and tiered usage-based pricing.
Per-seat for an API product, or flat-rate for a product where one customer creates 100× the load of another, leaks 20-40% of potential revenue at scale. Pick the model that scales with the customer's value extraction, not with what's easiest to bill.
Pricing benchmarks across SaaS verticals
When to raise SaaS prices
Most SaaS founders raise prices too late. The signals are clear when you look:
- Churn below 2% monthly. Price isn't the friction; customers stay despite the cost.
- 30%+ of demos don't ask about price. You're competing on value, not cost.
- Customers max-out their tier. Power users routinely hit usage limits and your tier ladder isn't steep enough.
- Competitor prices have moved up. The market has shifted; staying anchored on a 2-year-old number is a slow leak.
- You haven't raised prices in 12+ months. Inflation alone justifies a 5-8% raise on the existing book.
How to grandfather without backlash
Grandfather existing customers at their current rate for 12 months when you raise. Most won't notice. The few who do will feel respected.
"Hi <name> — heads-up that we're raising prices for new customers from <old> to <new> on <date>. Because you've been with us since <month>, your rate stays at <old> for the next 12 months. After <date+12mo> you'll move to <new>. We're investing the additional revenue in <2-3 concrete things>."
New customers pay the new rate from day one. The moral hazard of forcing existing customers to a higher price for the same product is real — grandfather them and you'll keep the goodwill while still capturing the upside on new acquisition.
Frequently asked questions
Validate a startup idea · First 10 customers playbook · SaaS glossary · Free pricing & CAC/LTV calculators