The models, and what each selects for
| Model | You charge for | Selects for |
|---|---|---|
| Flat subscription | Access, per period | Simplicity; easy to predict, hard to grow within an account |
| Per seat | Each user | Team products; discourages adding light users |
| Usage-based | What they consume | Aligns with value; makes bills unpredictable |
| Tiered | A bundle of limits and features | Most SaaS; the tier boundaries do the selling |
| Freemium | Upgrades from a free tier | Volume products with a cheap free path |
| One-time | The product itself | Tools; no recurring revenue to fund maintenance |
| Commission | A share of what they earn | Marketplaces; you get paid when they do |
Read the third column carefully — that's the part people skip. Per-seat pricing quietly encourages account sharing and discourages inviting the colleague who'd only look occasionally, which is exactly the person who spreads a product through an organisation.
Choosing the value metric
The value metric is the thing you count. It's the most consequential pricing decision, and the test is simple: as the customer gets more value, does the number go up?
| Metric | Verdict |
|---|---|
| Itineraries created per month | Good — grows with their business |
| Seats | Adequate — grows with the agency, not with success |
| Storage used | Poor — unrelated to what they get out of it |
| Support tickets | Bad — charges them for your product being confusing |
| API calls | Depends — good if calls track outcomes, bad if a retry is billable |
- It should be easy to count and easy to explain in one sentence
- The customer should be able to predict roughly what they'll owe
- It should be hard to game without also reducing the value they get
- It should grow slowly enough that a good month isn't a shocking bill
Free trials and free tiers
These are different things with different costs, and they're often confused. A trial is time-limited full access. A free tier is permanent limited access.
| Free trial | Free tier | |
|---|---|---|
| Costs you | Support during the trial | Support and infrastructure, forever |
| Works when | Value is obvious within days | The product gets better with more users |
| Fails when | Value takes a month to appear | The free tier is good enough to stay on |
| Card up front? | Higher conversion, fewer trials | Not applicable |
For a first product with few users, a time-limited trial with a conversation at the end is usually better than either. You learn why people don't convert, which no funnel report will tell you.
Designing tiers
- Three tiers is the practical default — more than four and the page becomes a decision problem
- Differentiate on the value metric first, features second. Feature-gating alone invites resentment
- Never gate security, backups, or basic support. Charging for the ability to not lose data reads exactly as badly as it sounds
- Name tiers after who they're for — Solo, Agency, Network — not Bronze, Silver, Gold
- Show prices. "Contact us" reads as expensive and loses everyone who won't ask
Common mistakes
- Choosing a value metric unrelated to the value delivered
- A free tier so generous nobody needs to upgrade
- Gating security or backups behind a higher tier
- Hiding prices behind a contact form on a self-serve product
- Treating pricing as a number to set once rather than a product decision to revisit
Key takeaways
- Each model selects for different customers and implies different features to build
- The value metric should rise as the customer's success rises
- Trials and free tiers solve different problems and cost differently
- Three tiers, priced visibly, differentiated on the value metric
Try it yourself
Write down what you'd count if you charged by usage, then ask whether that number rises when your customer succeeds. If it doesn't, you've found why your pricing feels arbitrary.