The most expensive automation decision isn't which tool you pick. It's picking one whose billing model doesn't match how your workflows are shaped.
Zapier bills per task — every individual step. Make bills per operation — every module run. n8n bills per workflow execution, regardless of how many steps that workflow contains.
At low volume the difference is invisible. At real volume it's the difference between a rounding error and a line item you have to justify — and it has almost nothing to do with which tool is better.
Here's how to choose without rebuilding in six months.
The short version
Three tiers:
Connectors — Zapier. Broadest integration library, easiest to learn, most expensive at volume.
Builders — Make, n8n. More capable logic, steeper learning curve, cheaper per unit of work.
Platforms — GoHighLevel, HubSpot. Automation bundled with CRM, email, and pages. Convenient, and the lock-in is real.
The billing models are the durable difference:
| Tool | Charges per | Implication |
|---|---|---|
| Zapier | Task (each step) | A 10-step workflow costs 10× a 1-step one |
| Make | Operation (each module run) | Similar, and polling triggers consume operations even when they find nothing |
| n8n | Workflow execution | Step count doesn't affect cost |
That last row is why the gap widens so sharply as workflows get complex.
The three tiers
Connectors do one thing: move data between apps when something happens. Zapier is the category leader, with the largest integration library by a wide margin and the shortest path from signup to working automation. Most non-technical people are productive within an hour.
Builders give you real logic — branching, iteration, error handling, data transformation, custom code. Make offers a visual canvas that suits people who think spatially. n8n is open-source, self-hostable, and the most flexible, with the steepest curve.
Platforms bundle automation with everything else. You're not buying a connector, you're buying a system — and you're accepting that the automation only works well with the platform's own tools.
The mistake people make: buying a platform because they wanted a connector, or fighting a connector's limitations for a year when they needed a builder.
What tasks and operations actually cost
The mechanics that determine your bill, and they're worth understanding before you build anything.
Zapier counts every action step as a task. A workflow with a trigger and four actions, running 500 times a month, consumes 2,000 tasks — not 500.
Make counts every module run as an operation, which is broadly similar. One important wrinkle: polling triggers consume operations even when they find nothing. A scenario checking a spreadsheet every fifteen minutes burns operations around the clock whether or not a row was added, which is a common and invisible source of consumption.
n8n counts the whole workflow run as one execution, regardless of step count. A 20-node workflow processing a batch is still one execution.
Why this compounds: the workflows worth building tend to be the multi-step ones — the ones with branching, enrichment, and conditional routing. Under per-step billing, sophistication is directly taxed. Under per-execution billing, it isn't.
When native integrations beat any connector
Worth checking before you build anything, because people routinely pay a connector to do something their existing tools do free.
Many tools now connect directly. Your CRM probably has a native integration with your email platform. Your form plugin likely posts to your CRM. Your ecommerce platform almost certainly connects to your email tool.
Native integrations are usually better: free, faster, maintained by the vendors, and they don't count against anyone's task allowance.
Where connectors earn their place: tools with no native link, anything needing logic between the steps, and workflows spanning three or more systems.
The habit: before building a Zap, check whether both tools already talk to each other. Ten seconds, and it saves a recurring cost surprisingly often.
All-in-one platforms: the tradeoff nobody mentions
Platforms like GoHighLevel or HubSpot bundle CRM, email, pages, scheduling, and automation into one system. The appeal is obvious: one bill, one login, everything already connected.
What you genuinely get: deep integration between components, no connector costs between internal tools, and one vendor to call.
The tradeoff people underestimate: you're not just buying software, you're buying a ceiling. Your automation is only as good as the platform's own tools, and if its email deliverability, page builder, or CRM is weaker than the specialist you'd otherwise use, you've traded capability for convenience.
And migration is genuinely hard. Your data, workflows, pages, and sequences all live inside one system in its own format. Leaving means rebuilding everything, which is why platforms bundle aggressively — the bundle is the moat.
When a platform makes sense: you're starting from nothing, you value simplicity over best-in-class, or you're an agency serving many similar clients where a repeatable setup matters more than optimality.
When it doesn't: you already have tools that work, or any single component is critical enough that you need the best available rather than the bundled version. The seven GoHighLevel workflows worth building.
Self-hosting: what it saves and what it costs
n8n's community edition is free and self-hostable, with no execution limits. On a small server, that's a genuinely large saving against per-task billing at volume.
What it saves: the per-task tax entirely. Volume stops being a cost variable.
What it costs — and this is the part that gets skipped:
Setup time. Docker on a small VPS is not hard if you've done it before, and it's a real barrier if you haven't.
Ongoing maintenance. Updates, backups, security patches, and monitoring. You now own uptime.
Debugging without support. Self-hosted users rely on community forums rather than a support desk.
A person who can do it. If that's one individual, you have a bus-factor problem — the automation stack becomes something only they can maintain.
The honest threshold: self-hosting makes sense when your cloud automation spend is large enough that the saving justifies the operational burden, and someone on the team is genuinely comfortable running a server. If either condition fails, the maths doesn't work regardless of the sticker price.
Lock-in, and keeping an exit
Every tool creates some. Reducing it is mostly a matter of habits rather than tool choice.
Document your workflows outside the tool. A plain-text description of what each does, what triggers it, and where data goes. If you had to rebuild elsewhere, that document is the spec — and rebuilding from a spec is hours rather than weeks.
Prefer webhooks and APIs over proprietary connectors where the effort is comparable. A webhook works anywhere; a platform-specific connector doesn't.
Keep data in systems you control. If your CRM is the source of truth and automation only moves data around it, switching tools is annoying. If the automation platform is the source of truth, switching is a migration project.
Export what you can, periodically. Most platforms allow workflow export in some format. Even an imperfect export beats a screenshot.
The realistic assessment: migration is usually measured in days per workflow rather than hours, and total time scales with workflow count more than complexity. Worth knowing before you have thirty of them.
What to buy first, starting from nothing
Start with the connector tier, even if you'll outgrow it.
Why: you learn how your tools behave, what actually needs automating, and how much volume you generate — all on something with the shortest learning curve and the best documentation. Those lessons transfer.
The sequence that works:
- Check for native integrations first. Free is better than cheap.
- Start on a connector's free or entry tier and build two or three simple workflows.
- Watch your consumption for a month. This is the number that determines everything downstream, and you can't estimate it accurately in advance.
- Reassess when consumption becomes a real cost or when you hit a logic limitation.
- Then choose deliberately — a builder if you need logic or volume, a platform if you want consolidation.
What not to do: choose a stack before you know what you're automating. The decision that matters is which workflows are worth building, and that's answered by the scoring method, not by a tool comparison.
Frequently asked questions
Is Zapier or Make better?
They're built for different people, and neither is better in the abstract. Zapier has the largest integration library and the shortest path from signup to a working automation — most non-technical people are productive within an hour. Make offers more capable logic and a visual canvas that suits complex branching, at the cost of a steeper learning curve. Both bill per step, so both get expensive as workflows grow in complexity. Choose Zapier if the priority is getting something working today; choose Make if you need real conditional logic.
Is n8n worth self-hosting?
Only if two things are true: your automation spend is high enough that eliminating per-task billing justifies the effort, and someone on your team is genuinely comfortable running and maintaining a server. Self-hosting removes the per-task tax entirely, which is a large saving at volume. What it adds is setup time, ongoing updates and backups, security patching, and debugging without a support desk. If nobody can comfortably run docker compose up, the sticker-price saving is misleading.
Do I need an all-in-one platform?
Only if consolidation matters more to you than having the best version of each component. Platforms bundle CRM, email, pages, and automation with deep internal integration and one bill — genuinely convenient, especially for agencies running repeatable setups across similar clients. The tradeoff is a ceiling: your results are limited by the platform's weakest bundled tool, and migration means rebuilding everything, which is why the bundling is aggressive in the first place.
How much should automation cost?
Less than the time it saves, which sounds obvious and is the calculation almost nobody does. The full equation is time saved per run × runs, minus build time, minus maintenance, minus the cost of errors — and maintenance is the term people omit. Most small operations should be spending very little, because the highest-value automations are simple data movement. If your bill is growing faster than your workflow count, that usually means step-heavy workflows under per-step billing rather than genuinely more automation.
What to do next
Before comparing tools, find out what you'd actually be paying for.
Count the steps in the workflows you're planning, and multiply by how often they'd run. That number — not the headline price — is what determines your cost, and it's the thing per-step billing punishes.
If your workflows are simple and infrequent, any tool works and you should pick the easiest. If they're multi-step and frequent, the billing model matters more than the feature list.
Free: The automation planning checklist.
Related guides
- Zapier vs Make vs n8n — the detailed comparison
- GoHighLevel: 7 workflows to build first — platform-specific
- One workflow, three tools — the same build compared
- What automation actually costs at scale — the pricing curves
- AI marketing automation — what to automate at all
- How to pick your first automation — before choosing a tool
Join the Newsletter
Get practical marketing tactics delivered straight to your inbox.
Written by
Muhammad Basim
Related Articles
What Marketing Automation Actually Costs at Scale
Nobody's automation bill jumps because they added more automations. It jumps because they added steps to workflows that were already running — a filter here, an enrichment lookup there, a second notification — and the billing model charges for every one of them, on every run. That's the mechanic behind almost every "why did this […]
One Workflow, Three Tools: A Build Comparison
Feature tables don't tell you what a tool is like to use. The only thing that does is building the same thing twice and noticing where you got annoyed. So here's one realistic workflow — lead routing with enrichment and conditional assignment — specified once and mapped across Zapier, Make, and n8n. What changes between […]
GoHighLevel: 7 Automation Workflows to Build First
GoHighLevel gives you a hundred things you could automate, which is exactly why most accounts end up with forty half-built workflows and no measurable result. Seven are worth building first. They're the ones tied directly to revenue — catching leads before they cool, recovering appointments that would otherwise vanish, and asking for reviews at the […]