Zapier doesn't charge by the Zap. It charges by the task, and that distinction is where most people's Zapier bill goes sideways. A "task" is one successful action step, not one automation, which means a single Zap running once can consume five tasks or one depending entirely on how it's built.
What actually counts as a task
Every time a Zap fires and completes an action step, that's one task. A simple one-trigger-one-action Zap (new form submission triggers an email) uses one task per run. A multi-step Zap that pulls a new lead into a CRM, then posts a Slack notification, then adds a row to a spreadsheet burns three tasks for that same single trigger event, because each of those three actions counts separately.
The trigger itself doesn't cost anything. Zapier checking whether a new event happened, whether that's a new email, a new form response, or a new database row, is free regardless of how often it polls. Tasks only get counted when an action actually executes as a result of that trigger firing.
Filters and Paths change the math in a useful way. A Filter step that stops a Zap from continuing (because the incoming data doesn't match a condition) does not count as a task, since the Zap never completes its action steps. Paths, which route a Zap down different branches based on conditional logic, only consume tasks for the branch that actually executes, not every branch defined in the workflow. Built correctly, filters and conditional paths can meaningfully reduce task consumption compared to running every downstream action unconditionally.
The three tiers and what they actually buy
Free costs nothing and includes 100 tasks per month along with 5 single-step Zaps limited to core apps. It's genuinely useful for testing whether an automation idea works before committing to a paid plan, but multi-step Zaps aren't available at all on Free, only single-trigger-single-action workflows.
Starter runs $19.99/month for 750 tasks and unlocks multi-step Zaps along with Filters. This is the tier where Zapier starts being useful for anything beyond the simplest automations, since most real workflows need more than one action per trigger.
Professional costs $49/month for 2,000 tasks and adds Paths for conditional branching, access to premium app integrations, and webhook support for connecting to services outside Zapier's standard app directory. Teams running automations that need to branch based on incoming data, or that pull from apps outside the free integration tier, end up needing Professional regardless of task volume.
Where the task count actually goes
A 3-step Zap (trigger plus 3 actions) running 250 times a month uses 750 tasks, exactly maxing out the Starter tier's monthly allowance. Add a fourth action step and the same 250 monthly runs blow past Starter and require Professional's 2,000-task ceiling instead, even though the trigger volume didn't change at all. This is the part of Zapier's pricing that catches people off guard: task consumption scales with workflow complexity, not just with how many times something happens.
A team running 10 different Zaps, each averaging 2 action steps and firing 50 times a month, uses 1,000 tasks monthly, comfortably within Professional but already over Starter's limit. Scaling that same setup to 20 Zaps at the same complexity doubles task usage to 2,000, landing right at Professional's ceiling with no room for growth without an upgrade or a plan built around usage-based overage.
The real cost of adding steps
Because tasks are counted per action step rather than per automation, the cheapest way to control a Zapier bill isn't necessarily choosing a bigger plan, it's auditing what each Zap actually does. A Zap with an unnecessary formatting step, or one that posts to a Slack channel nobody reads, is burning task allowance for no real value. Consolidating steps where possible, or removing action steps that don't need to run every single time, is a more direct lever on cost than upgrading tiers.
Filters are worth using deliberately rather than as an afterthought. A Zap that runs an expensive multi-step action chain and only later filters out irrelevant records wastes tasks on records that get discarded anyway. Putting the Filter step as early as possible in the Zap, before the action steps that actually cost tasks, keeps consumption tied to records that matter.
Where Zapier's task model gets expensive
High-volume, multi-step automations are where the task-based model bites hardest. A workflow syncing e-commerce orders across five different systems, with each order triggering five separate action steps, hits 5,000 tasks after just 1,000 orders, well beyond what Professional's 2,000-task ceiling covers. At that volume, the real comparison isn't between Zapier tiers, it's between Zapier and alternatives with different pricing models entirely: Make.com prices lower for comparable volume, and n8n's self-hosted option removes per-task billing altogether at the cost of running and maintaining the infrastructure.
Reddit automation communities consistently flag Zapier's task-based pricing as the single biggest reason teams migrate to Make or n8n once volume grows past a few thousand tasks a month. The ease-of-use advantage that makes Zapier the default starting point stops being worth the premium once a workflow's task count scales into the thousands.
What happens when a Zap fails mid-run
Task consumption on a failed Zap depends on where the failure happens. If an action step errors out partway through a multi-step Zap, the steps that completed successfully before the failure still count as tasks, since Zapier already executed them. Only the step that failed and any steps after it are skipped. This means a Zap that fails on its third of four action steps still consumes two tasks for the steps that ran, not zero.
Zapier retries failed steps automatically on paid plans, and each retry attempt that succeeds counts as a task the same way a first-try success would. A flaky downstream API that requires several retries to succeed can quietly inflate task usage beyond what the raw automation volume would suggest, which is worth checking if a task count looks higher than expected relative to actual trigger volume.
Team versus individual billing
The pricing above covers individual and small-team use on Starter and Professional, both billed per account rather than per user, meaning task allowance is shared across everyone with access to that Zapier account rather than multiplied by seat count. That's a real advantage for small teams collaborating on shared automations, since adding a second or third person to manage Zaps doesn't increase the bill the way per-seat pricing would.
Larger organizations that need role-based permissions, shared team folders, or centralized admin controls typically need to move beyond Professional into higher-volume plans built for that scale, where Zapier shifts toward negotiated or usage-tiered pricing rather than the flat published rates on Starter and Professional. Teams anticipating that kind of growth should factor in that the publicly listed pricing mostly describes the entry and mid-tier experience, not what an automation-heavy org with dozens of active Zaps ultimately pays.
Is the task model worth it
For simple automations under a thousand tasks a month, Zapier's task-based pricing is straightforward and the platform's no-code builder and 6,000-plus app library make it the fastest path from idea to working automation. The task-counting mechanics rarely cause surprises at that scale because workflows tend to be simple enough that task usage is predictable.
The model gets harder to manage as automations get more complex or run more often. Anyone building multi-step Zaps that fire frequently should track actual task consumption against the plan limit before assuming a given tier covers real usage, since the gap between "how many times does this trigger" and "how many tasks does this actually cost" widens fast once a workflow has more than two or three action steps.