SaaS Pricing Models Explained: What You're Actually Paying For
SaaS pricing isn't just about the number on the website. The pricing model determines how your costs scale, where you'll get hit with overages, and how much negotiating power you have at renewal. Understanding the model matters more than memorizing the price.
After evaluating hundreds of SaaS products, we've seen every pricing structure out there. Here's what each one actually means for your budget — and your negotiating position.
Per-Seat (Per-User) Pricing
How it works: You pay a fixed amount per user per month. Add a user, add a charge. Remove a user, save that charge (eventually — most vendors don't prorate mid-cycle).
Who uses it: Salesforce, Slack, Atlassian, Zoom, Microsoft 365, most B2B SaaS.
Where it works well:
- ▸Predictable teams with stable headcount
- ▸Products where usage correlates closely with user count
- ▸Organizations that want simple budgeting
Where it breaks down:
- ▸Fast-growing companies where headcount changes quarterly
- ▸Products with lots of occasional users (do you really need a $25/month seat for someone who logs in twice a year?)
- ▸When vendors bundle admin accounts, API users, or service accounts as "seats"
What to watch for:
- ▸The "active user" vs. "provisioned user" distinction. Some vendors charge for every provisioned seat, even if the person never logs in. Others only charge for active users — but define "active" loosely.
- ▸Minimum seat commitments. Enterprise contracts often require a minimum number of seats for the entire term, even if your team shrinks.
- ▸The SSO tax. Some vendors charge more for SSO/SAML support, effectively penalizing companies for following security best practices. This has become a known frustration in the industry — check whether SSO is included in your tier or requires an upgrade.
Negotiation tip: Ask for a "ramp" schedule if you're growing. Start with 100 seats at the negotiated rate, with the ability to add up to 200 at the same rate over 12 months, rather than committing to 200 on day one.
Usage-Based Pricing
How it works: You pay based on what you consume. API calls, data processed, messages sent, compute hours, storage used. Your bill fluctuates monthly based on actual usage.
Who uses it: AWS, Twilio, Snowflake, Datadog, SendGrid, Stripe.
Where it works well:
- ▸Workloads with variable demand (seasonal businesses, startups finding product-market fit)
- ▸When you want costs to correlate directly with value delivered
- ▸Technical products where usage varies dramatically between customers
Where it breaks down:
- ▸Finance teams that need predictable monthly costs for budgeting
- ▸When a bug or misconfiguration can generate a surprise five-figure bill
- ▸Products where the "unit" of consumption is hard to predict (what does 10 million "events" actually mean for your use case?)
What to watch for:
- ▸Bill shock. Datadog's pricing model has become a cautionary tale, teams have reported bills jumping 3-5x after enabling a new integration they didn't realize would generate millions of additional metrics. Always understand what counts as a billable unit before turning something on.
- ▸Committed-use discounts vs. flexibility. Snowflake and AWS offer significant discounts for committing to annual spend. The discount is real, but you're trading flexibility for savings. If your usage drops, you still pay the committed amount.
- ▸Unit cost at scale. Some usage-based vendors have declining per-unit costs as you scale; others don't. A vendor that charges $0.01 per API call at 1M calls might still charge $0.01 at 100M calls, or might drop to $0.002. Ask explicitly.
Negotiation tip: Ask for a spend cap or "circuit breaker", an automatic alert and optional pause if your monthly bill exceeds a threshold. This protects against runaway costs without committing to a fixed plan.
Tiered Pricing
How it works: The vendor offers 3-4 tiers (typically Free, Starter/Pro, Business, Enterprise) with increasing features and limits at each level. You pick the tier that matches your needs.
Who uses it: Almost everyone. HubSpot, Notion, Asana, Dropbox, Canva.
Where it works well:
- ▸When you can clearly identify which tier matches your requirements
- ▸Organizations that want a defined set of features for a predictable price
- ▸Products where the natural upgrade path aligns with growing team needs
Where it breaks down:
- ▸When the one feature you need is only in the top tier (the "99% of what I need is in Pro, but I need Enterprise for SAML" problem)
- ▸When tiers are designed to push you upward rather than serve your actual needs
- ▸When tier boundaries are set at awkward points (your team of 55 is just past the "up to 50 users" Pro tier)
What to watch for:
- ▸Feature gating as upsell pressure. Vendors know which features make you upgrade. If audit logs, SSO, or API access are gated to higher tiers, that's a deliberate choice to extract more revenue from security-conscious or integration-heavy customers.
- ▸Tier limit resets. Some vendors reset usage limits (API calls, storage, automations) when you upgrade mid-cycle. Others don't, meaning you pay for the higher tier but don't get the higher limits until next billing cycle.
- ▸The "Contact Sales" tier. Enterprise pricing is almost always negotiable. The price behind "Contact Sales" is whatever the sales rep thinks you'll pay. Always negotiate, and always get competing quotes.
Negotiation tip: If you're stuck between tiers, ask for the lower tier with the one or two features you need from the higher tier added à la carte. Many vendors will do this to close a deal, even if it's not on the website.
Flat-Rate Pricing
How it works: One price, one product, all features included. No tiers, no per-seat scaling, no usage metering.
Who uses it: Basecamp (famously), some smaller SaaS tools, and enterprise products that want to simplify procurement.
Where it works well:
- ▸Large teams where per-seat pricing would be expensive
- ▸Organizations that value simplicity and predictability above all
- ▸Products where the marginal cost of additional users is near zero
Where it breaks down:
- ▸Small teams who are effectively subsidizing larger ones
- ▸When the vendor needs to introduce limits (storage, API calls) that undermine the "flat rate" promise
- ▸When the flat rate is set high to capture enterprise budgets, pricing out smaller companies
What to watch for:
- ▸Hidden limits. "Unlimited" almost never means unlimited. Read the terms of service, there are usually fair-use policies, API rate limits, or storage caps that technically make it usage-based with a very high ceiling.
- ▸Price increases. Flat-rate vendors can't increase revenue by adding seats or upselling tiers. That means when costs go up, the flat rate goes up, and you have no negotiating lever to offset it.
Negotiation tip: Flat-rate pricing usually isn't negotiable on price, but you can often negotiate on payment terms (annual vs. monthly), implementation support, or contract length.
Freemium
How it works: A free tier with limited features or capacity, designed to get you using the product. Revenue comes when you hit the limits and need to upgrade.
Who uses it: Slack, Zoom, Notion, Figma, Canva, HubSpot, Airtable.
Where it works well:
- ▸Evaluating a product before committing budget
- ▸Small teams or personal use where the free limits are sufficient
- ▸Products where the free tier is genuinely useful, not just a demo
Where it breaks down:
- ▸When the free tier becomes a data hostage situation, your data is in the product, the free limits are constraining you, and now the upgrade price is whatever they say it is
- ▸When free-tier reliability or support is significantly worse than paid tiers
- ▸When the vendor uses free-tier data to justify investor metrics ("we have 10 million users!") while the product experience for free users degrades
What to watch for:
- ▸The free-to-paid cliff. Some freemium products have a smooth upgrade curve. Others jump from $0 to $25/user/month with nothing in between. Understand the upgrade path before you invest time in the free product.
- ▸Feature removal from free tiers. Slack limiting message history, Google limiting free Workspace features, vendors routinely reduce what's available for free as the product matures. Don't build critical workflows on free-tier features.
- ▸Data portability from free tiers. If you outgrow the free tier and the paid version is too expensive, can you actually get your data out? Test this early.
Negotiation tip: If you're converting from free to paid, ask for an extended trial of the paid tier (30-60 days, not 14) and for the free-tier data migration to be handled by the vendor. They want the conversion badly.
Hybrid Models
Increasingly, vendors combine models. Slack charges per seat but also per feature tier. Snowflake is usage-based but offers committed-use contracts. HubSpot is tiered but charges per "marketing contact" within each tier.
What to watch for with hybrid models:
- ▸Compounding costs. If you're paying per seat AND per usage AND per feature tier, small increases in each dimension multiply. A 10% increase in seats, a 15% increase in usage, and a tier upgrade can turn a $50K contract into $85K.
- ▸Unpredictable budgeting. The more variables in your pricing formula, the harder it is to forecast costs. Ask the vendor to model your likely 12-month spend based on your projected usage, and get that estimate in writing.
How to Evaluate Any Pricing Model
Regardless of the model, ask these five questions before signing:
- What does year-three pricing look like? Not year one with the discount. What's the realistic cost in three years with growth, price increases, and feature needs?
- What triggers a cost increase I can't control? Usage spikes, vendor price changes, SKU restructuring, minimum commitments, identify every variable that's not in your hands.
- What's the cost to leave? Data export fees, migration effort, retraining, workflow rebuilding. The cheaper the product is to buy, the more expensive it often is to leave.
- How does the vendor make money from me over time? If they need revenue growth, it has to come from somewhere. Is it new customers, upselling existing customers, or price increases? The answer tells you what's coming.
- Who else at my company can incur charges? In a per-seat model, can any manager add users? In usage-based, can any developer spin up resources? Understand the blast radius of uncontrolled purchasing.
For Home Users and Freelancers: What This Means For You
You are not negotiating a $200K enterprise deal, but pricing models still affect your wallet.
Per-seat pricing penalizes families. If you want 1Password for your household of four, you are buying a Family plan whether everyone uses it daily or not. Same with Notion, Figma, or any per-seat tool. Check if a family or team plan exists before buying individual seats.
Usage-based pricing can surprise you too. Cloud storage overages on iCloud or Google One, API costs if you are building side projects on Vercel or AWS, even bandwidth charges on media hosting. Set billing alerts on every service that charges by usage.
Freemium is designed to hook you. The free tier is genuinely useful to get you invested. Once your data, workflows, and habits are embedded in the tool, the upgrade price is whatever they say it is. Before committing to any free tool, check what happens to your data if you never upgrade and the free tier gets reduced.
Annual billing saves real money but traps you. That 20% savings for paying annually is $50-100 on typical productivity tools. But if you realize in month 3 that the tool is not for you, that money is gone. Pay monthly until you are certain, then switch.
The real question for personal software: Can I get my data out easily? If the answer is no, the tool owns you, not the other way around. Obsidian stores files as plain Markdown you can move anywhere. Notion requires an export that loses formatting. That difference matters more than any feature comparison.
The Bottom Line
No pricing model is inherently better or worse. The right model depends on your team size, usage patterns, growth trajectory, and risk tolerance. What matters is that you understand the model, not just today's price, but how it scales, where the surprises hide, and what happens when you want to leave.
The most expensive SaaS product isn't the one with the highest list price. It's the one where you didn't understand the pricing model until the first renewal.
Related reads: