Comparing two tools by their starting price assumes they are selling the same thing. Frequently they are not. One charges for each person who logs in, another for each time a workflow runs, a third for each step inside that workflow — and the number on the pricing page means something different in each case.
This guide sets out the five metering models you will meet, what each one does to your bill as you grow, and the arithmetic that turns a headline price into a real one. It is the piece of research most worth doing before you commit, because the pricing model is far stickier than the price: vendors adjust prices often and change how they count almost never.
- The billing unit matters more than the headline price — per task and per execution can differ by an order of magnitude on the same workflow.
- Per-seat pricing is the model that punishes success: every hire is a recurring cost increase you did not separately decide on.
- Most advertised prices assume annual billing, and the monthly premium is sometimes far steeper than the usual two-months-free convention.
- Credit and add-on meters run out independently of the main subscription, so one plan can have two ways to stop working.
- Before comparing two tools, restate both prices as cost per month at your real volume. Only then are the numbers about the same thing.
Per seat: the model that charges you for growing
The most common model, and the easiest to underestimate. You pay for each person with access, so the cost is a straight multiple of headcount. Its virtue is predictability: you always know what the next colleague costs. Its problem is that the decision compounds silently — five tools at per-seat pricing means every hire triggers five simultaneous cost increases nobody explicitly approved.
Watch for the occasional seat in particular: the bookkeeper who needs read access twice a month costs the same as your most active user in most products. When you audit spending, that is usually where the recoverable money is.
Per workspace: predictable until it is not
Some tools price the whole workspace rather than each member, which flips the arithmetic entirely. For a small team this is often dramatically cheaper, and it means adding a colleague costs nothing at all. The catch is the tier boundary: crossing it moves you to a materially higher price in one step rather than gradually.
This model rewards teams that stay just under a threshold and penalises those that step just over one. Before choosing it, check where the boundary sits relative to the size you expect to be in a year.
Per task versus per execution: the difference that costs the most
In automation this is the single most consequential distinction, and the vocabulary actively hides it. A tool that meters tasks counts every step a workflow performs: a five-step workflow firing a hundred times consumes five hundred units. A tool that meters executions counts the workflow run itself, so the same activity consumes one hundred. Same work, five times the consumption, depending only on how the vendor defines the word.
This is why the two products below cannot be compared on their starting prices. The only reliable method is to count the steps in your busiest automations, multiply by expected frequency, and convert both pricing pages into a monthly figure at that volume.
| Tool | Verified price | Billing condition | Free | Checked |
|---|---|---|---|---|
| Zapier | $29.99/mo · Professional | $19.99/month for 750 tasks with yearly billing; $29.99/month with monthly billing | Yes | 12.09.2026 |
| n8n | €24.00/mo · Starter | €24/month billed monthly; €20/month billed annually. The self-hosted community edition stays free | Yes | 13.09.2026 |
| Make | $9.00/mo · Core | $9/mo for 10,000 credits; annual billing saves 15% or more | Yes | 06.08.2026 |
Per credit: the meter that runs out separately
Credits appear wherever the vendor's own costs vary by usage — image generation, video rendering, AI summarising, long-context processing. The model is honest in principle, because heavy users genuinely cost more to serve. It is difficult to budget for in practice, because you cannot easily predict how many credits a month of real work consumes until you have had one.
The specific trap is that a credit meter usually runs alongside the subscription rather than inside it, so a plan can stop being useful in two independent ways: the seats are fine but the credits are gone. When a tool has a separate credit allowance, treat it as a second subscription you have not priced yet.
Annual billing: the discount that is not always the usual discount
Nearly every advertised price assumes annual billing, and the convention is roughly two months free for committing to a year. It is worth checking, because the convention is not universal — some vendors set the monthly rate far above the annual equivalent, which turns flexibility into an expensive choice rather than a mildly costly one.
There is also the opposite disclosure problem, where a vendor publishes annual pricing and no monthly rate at all. When that happens, the cost of not committing is unknown rather than high, and a tool you wanted to try for one busy month becomes a year-long decision.
The arithmetic that makes prices comparable
Restate every candidate as one number: total cost per month, at the volume you actually expect, including every meter. That means headcount times seat price, plus credit top-ups, plus the add-ons sold separately from the base plan, converted from annual to monthly if the page quoted a year.
Do this for a realistic twelve months out rather than today. The tool that is cheapest for three people is frequently not the tool that is cheapest for six, and migration between platforms is manual work in every category covered on this site. Choosing the pricing model is a longer commitment than choosing the product.
Frequently asked questions
Which model is best?
None of them, in the abstract — the right model depends on the shape of your usage. Per workspace suits growing teams with modest usage. Per execution suits complex automations. Per seat suits stable headcount with heavy per-person use. The mistake is not picking the wrong model, it is not noticing which one you picked.
How often do these prices change?
Prices change often; billing models change rarely. That asymmetry is the practical argument for choosing on the model rather than the number. Every price on this site carries the date it was verified against the vendor's official pricing page, shown on each tool page, and we re-check on a review cycle rather than continuously.
Is annual billing always worth it?
It is worth it when you are confident you will still be using the tool in six months, and not before. The discount is real, but so is the cost of committing to a product you abandon in month three. For a tool you are still evaluating, pay monthly and treat the premium as the price of the option to leave.
What is the fastest way to audit what we already pay?
List every subscription with its billing unit next to it, not just its cost. The overspend is almost always in one of two places: seats belonging to people who barely log in, and a per-task automation bill that grew as workflows got longer. Both are invisible on a bank statement and obvious in a table like that.