n8n vs Make: how I actually choose for a client project

The question I actually ask first

Before comparing nodes or pricing tiers, the real first question is: who is going to look at this workflow six months from now? If it's a non-technical operations person who needs to glance at a run history and understand what happened, that pulls toward Make's linear, visual scenario builder. If it's a developer or a technically comfortable founder who will eventually want to extend the logic with custom code, that pulls toward n8n.

This isn't a preference call, it's a maintenance call. A beautifully built n8n workflow that the client's team can't touch without me becomes a support burden for both of us. A Make scenario that hits a genuine technical wall the client can't work around does the same thing in the other direction.

Where I've used n8n

The Shopify AI customer support agent I built runs on n8n. That workflow needed conditional branching around an AI decision step (can the agent answer confidently, or does it need to escalate to a person), and n8n's code node made it straightforward to shape the data exactly how the AI step needed it, without fighting the platform.

The lead generation system I built pairs a Python scraper with an n8n workflow for the same reason: once real code is already part of the pipeline, n8n's willingness to let a workflow call out to custom logic stops being a nice-to-have and starts being the reason the whole thing works. n8n is also part of the ongoing Habfract Ltd. operations automation, where several workflows route and transform data between Notion, Salesforce, Shopify and Google Sheets.

Where I've used Make

The Cliniko healthcare workflow automation runs on Make, connecting Cliniko, Notion and Google Sheets so appointment and billing data stays aligned. That's a case where the logic itself is fairly linear (a change in one system needs to be reflected in the others), and Make's visual scenario view made it easy for the practice to see exactly what the automation does without reading workflow code.

The AI powered social media automation system is also Make: it generates and schedules content on a recurring basis, which is a good fit for Make's scenario-and-schedule model. And Habfract's operations automation genuinely runs on both — different workflows in the same business, picked for what each one actually needed rather than a single platform decision made once and never revisited.

The parts of the usual comparison that matter less than people think

Pricing matters, but not in the way most comparisons frame it. The real cost question isn't the monthly subscription, it's what happens when a workflow's volume grows 10x. n8n's self-hosted option means execution volume stops being the pricing lever; Make's operation-based pricing means it can. That only actually matters once a workflow is running at real volume, which is a smaller number of my projects than the comparison articles suggest.

Integration count is another one. Both platforms cover the vast majority of tools a small or mid-sized business actually runs — the deciding factor in practice is almost never 'does it have a connector for X,' it's whether the connector for X exposes the specific field or webhook event the workflow actually needs.

A short, honest decision guide

Choose Make if the team maintaining the workflow is non-technical, the logic is mostly linear (trigger, transform, send), and the business wants to see and understand the automation without reading code.

Choose n8n if the workflow needs custom code, conditional AI decision logic, self-hosting for data privacy, or is likely to scale to a volume where per-operation pricing becomes expensive.

And sometimes, like at Habfract, the honest answer is both — picked workflow by workflow, not as a single platform-wide bet.