n8n vs Make vs Zapier vs Trigger.dev. Which one should you actually use?

How the picker scores your answers

Each answer adds or subtracts weight for the four platforms. The weights aren't evenly distributed. Some questions (hosting, team technical capability) move the needle more than others (monthly volume) because they rule tools in or out rather than shade a preference. A “must self-host” answer, for example, actively penalises Zapier and Make because they don't offer it at all.

The top score wins. The runner-up is your fallback if the winner turns out to be a bad cultural or commercial fit once you poke at pricing. The full ranking is shown so you can sanity check: if two tools are tied at the top, either one is defensible.

When each platform makes sense

n8n

Pick n8n when you have engineers or ops people who can operate infrastructure, you want to self-host for cost or compliance, or you need complex workflows without Zapier's pricing model. n8n works well for internal ops pipelines, scraping, and research automations: anywhere you'd otherwise write cron jobs but want a UI for non-devs to tweak.

Where it falls down: if nobody on your team is comfortable debugging a failed workflow node, n8n will feel exposed. It assumes a level of engineering maturity that Zapier doesn't ask of you.

Make

Pick Make when you want Zapier-level ease but need real branching, loops, and data aggregation, and you're happy on cloud. It's significantly cheaper than Zapier at volume, and its scenario builder handles logic that would require a Zapier Path and three sub-zaps. The learning curve is steeper than Zapier's but gentler than n8n's.

Where it falls down: smaller integration catalogue than Zapier (you'll hit missing connectors for niche tools), no self-host, and debugging a misbehaving scenario can be genuinely painful.

Zapier

Pick Zapier when you're a non-technical team, you value reliability over cost, and your automations are “when X happens, do Y” patterns rather than multi-branch workflows. Zapier is the right default for marketing ops, founder-led admin, and anywhere the builder isn't technical. It's the most expensive per task at volume but the cheapest in operator time.

Where it falls down: once you're running more than about 50,000 tasks a month, the bill starts to distort your P&L. Past that, Make or n8n usually pays back the migration effort inside a year.

Trigger.dev

Pick Trigger.dev when the automation is part of your product, not part of your ops. It's a code-first platform for TypeScript teams who want durable execution, retries, observability, and scheduling without building that infrastructure themselves. If you're shipping AI agents, long-running data pipelines, or any background job that can't afford to silently fail, this is the layer.

Where it falls down: it's not for ops teams, not for non-technical buyers, and has a smaller integration ecosystem than the no-code tools. You write the glue yourself.

Common mistakes when picking an automation platform

• Picking self-host because it sounds cheaper. Self-hosting n8n is cheap on server cost and expensive on ops time. If you don't have someone who enjoys operating infrastructure, the cloud bill is the cheaper answer.
• Optimising for the lowest task price. Task price matters at scale; it's noise at 5k tasks/month. Pick for team capability and reliability first, then tune cost.
• Using a no-code tool for product features. If the automation is customer-visible and paid for, it should live in your codebase or in a code-first platform, not in a Zap your CEO can accidentally delete.
• Assuming “we'll move later” is free. Migrating dozens of live workflows is expensive. If you think you'll outgrow Zapier in 18 months, evaluate Make or n8n now.