Every article about software pricing in 2026 says the same thing: the seat is dead, usage and outcome pricing have won. We keep 43 verified records with the billing condition written out in full, so we counted. The seat is not dead. Twenty-one of the forty-three tools here still charge per seat, per member or per user, and that is the largest single group by some distance.
What the trend pieces get right is narrower and more expensive than the headline. Seats have not been replaced. They have grown a credit meter bolted on top, and the two are forecast in completely different ways.
Before you compare two prices, check you are comparing the same unit. Across this database there are six: per seat, flat, usage credits, per channel, per brand and per site. A tool that looks half the price of another is often just counting something you have fewer of today and will have more of next year. The unit, not the number, is what decides your bill in month twelve.
Six units, one price column
Pricing pages all present a single figure with a currency in front of it, which invites you to read two tools side by side as if the numbers were comparable. They frequently are not. Here is what the forty-three records in this database actually meter:
- Per seat, member or user — 21 tools. Scales with headcount, including the people who barely log in.
- Flat per account — 12 tools. The only unit that does not move when your business does.
- Usage credits, tasks or minutes — 7 tools. Scales with how much work you push through, which is the hardest thing to forecast.
- Per channel — Buffer, at 6 USD each. Five social networks is five times the headline price.
- Per brand — Metricool, at 20 EUR a month for up to five. An agency's sixth client changes tier.
- Per site — Framer, at 15 € each. Three client sites is three subscriptions, not one.
The credit meter bolted onto the seat
This is the change worth knowing about, and it does not show up as a new pricing model — it shows up inside an old one. Figma charges 20 € per Full seat per month and includes 3,000 AI credits with it. Framer charges 15 € per site and includes 1,000 AI credits a month. Neither is a usage-priced product. Both now have a second meter running underneath the one you agreed to, and neither headline price tells you what happens when it empties. Seven of our records mention credits somewhere in their pricing terms; only four of those are usage-priced products. The rest are seats with a meter.
Credits are not a currency, and each vendor mints its own
The word makes different products sound comparable when they are not. Make sells 10,000 credits for $9. Zapier sells 750 tasks for $29.99. A Make credit is one module call inside a scenario; a Zapier task is one action inside a Zap. A single automation that fetches a record, transforms it and posts it somewhere spends one task in Zapier and three credits in Make. The ratio between the two is not a constant you can look up — it depends entirely on how your own workflows are shaped, which is why the cheaper credit pack is not automatically the cheaper tool.
How to forecast each one before you commit
The practical test differs by unit, and takes about ten minutes per tool.
- Seats: count the people who will need write access in twelve months, not today, and check whether guests, clients and reviewers consume a paid seat. That answer is on a help page, not the pricing page.
- Credits and tasks: run one real workflow, read the consumption, multiply by a busy month rather than an average one. Then check the price of the next pack up, because that is the number you will actually pay.
- Channels, brands and sites: count what you have now and what one new client adds. These units step rather than scale — the sixth brand or the fourth site is a tier change, not a small increase.
- Flat: check what is excluded. A flat price with paid add-ons for the feature you came for is not a flat price.
What we do not know
Our records cover the published entry price and its stated condition, read on the vendor's own monthly view and dated. They do not cover what a credit costs once the included allowance runs out, because most vendors do not publish that in a form we can verify — and we will not estimate it. Where a tool meters something, our record says so and names the unit. Treating that as the whole answer would be the same mistake as reading the headline price and stopping.
Pricing side by side
| Tool | Verified price | Billing condition | Free | Checked |
|---|---|---|---|---|
| Figma | €20.00/mo · Professional | 20 €/month per Full seat billed monthly; 16 €/month billed annually. Dev seat 15 €, Collab seat 5 € (monthly) | Yes | 12.09.2026 |
| Framer | €15.00/mo · Basic | 15 €/month per site billed monthly; 10 €/month billed annually | Yes | 12.09.2026 |
| Make | $9.00/mo · Core | $9/mo for 10,000 credits; annual billing saves 15% or more | Yes | 06.08.2026 |
| Zapier | $29.99/mo · Professional | $19.99/month for 750 tasks with yearly billing; $29.99/month with monthly billing | Yes | 12.09.2026 |
| Buffer | $6.00/mo · Essentials | 6 USD per channel/month billed monthly; 5 USD per channel/month billed yearly | Yes | 05.09.2026 |
| Metricool | €20.00/mo · Starter | 20 EUR/mo billed monthly for up to 5 brands; 16 EUR/mo billed annually | Yes | 20.08.2026 |
| Otter.ai | $16.99/mo · Pro | $8.33/user/mo billed annually; $16.99/user/mo billed monthly | Yes | 12.09.2026 |
| HeyGen | $29.00/mo · Creator | 29 USD/month billed monthly, metered in credits | Yes | 20.08.2026 |
Frequently asked questions
Is seat-based pricing actually going away?
Not at the small-business end of the market. Twenty-one of the forty-three tools in this database still bill per seat, member or user, which is more than any other unit. What has changed is that several of those seats now include a metered AI allowance, so the seat price is a floor rather than the whole bill.
Which unit is cheapest for a freelancer working alone?
Flat pricing, almost always, because a solo user pays the same as a team of one either way and nothing scales against them. The unit to be careful with alone is per site or per channel: those count your outputs, not your people, and a freelancer can easily have more of those than a small agency.
Why do you not convert credits into a per-task cost?
Because the conversion depends on your workflows, not on the vendor. The same automation costs a different number of credits depending on how many steps it has, and publishing an average would invent a precision the vendor itself does not publish. We record the unit and the allowance, and leave the arithmetic to your own usage.
What is the most expensive unit to get wrong?
Credits, because the failure is invisible until the bill arrives. A seat count that is wrong is wrong on day one and you notice. A credit allowance that is wrong runs fine for three weeks and then either stops your automations mid-month or silently bills an overage.