Why RevOps Teams Chose n8n Over Make in 2026

Somewhere around Q3 2025, a pattern showed up across three of my mid-market clients: they had all independently started migrating their automation stacks from Make to n8n, and none of them had talked to each other. The trigger was different each time (a surprise billing spike, a data compliance question from legal, a workflow hitting scenario step limits) but the destination was the same. By early 2026, “n8n vs Make self-hosted” had become the most common infrastructure question I fielded from RevOps leads. This post is what I learned watching those migrations happen in real time, and what I tell teams evaluating the switch today.

The billing inflection point that keeps accelerating the switch

Every team I have talked to had a version of the same moment: the Make bill jumped, and someone finally did the math. At Make’s Growth plan (roughly $29/month for 10,000 operations), costs scale linearly. A mid-market RevOps team running enrichment loops through Clay, syncing contacts into HubSpot, triggering Smartlead sequences, and logging engagement back to Salesforce can burn 500,000 to 2 million operations per month without blinking. At that volume, Make’s enterprise pricing enters the conversation fast.

n8n self-hosted on a $40/month DigitalOcean droplet or a $60/month AWS t3.medium instance has no operation caps. You pay for compute, not executions. One client I worked with at Homegrown Growth Co. was paying $340/month on Make’s Business plan and moved to n8n on a $50/month VPS. Infrastructure cost dropped 85 percent. They added more workflows in the following quarter, because the psychological tax of “every trigger costs money” disappears completely when you own the runtime.

Don’t bother trying to negotiate volume discounts with Make at this scale. The math never pencils. Self-host n8n and spend that budget on better data sources instead.

What operators actually find different in production

The cost story gets attention. It is not the whole reason teams stay with n8n once they have migrated, though. These two tools have genuinely different operating models, and the differences compound over time.

Make is built around a scenario-centric mental model where every automation is a linear chain of modules. Approachable. UI is polished. Non-technical operators can build real workflows without touching code. But that model has hard edges: loops within loops get awkward fast, error handling is limited compared to what you can do in code, and the step-limit ceiling on lower plans becomes a real constraint for enrichment-heavy workflows.

n8n treats every workflow as a graph. Nodes fan out, merge, loop, and branch in ways that map more naturally to how complex GTM logic actually works. The n8n docs on expressions expose a JavaScript-based templating language that lets you transform data inline without spinning up a separate code node. And the Code node itself drops you into full JavaScript when you need it. I have rebuilt Make scenarios in n8n that required 18 modules and found that a single Code node with 30 lines of JavaScript was cleaner and faster to maintain than anything Make could produce.

The data residency angle is not theoretical. By 2026, more RevOps teams at Series B and beyond are fielding questions from security and legal about where prospect and customer data lives during processing. With Make, the answer is always “Make’s cloud infrastructure.” With n8n self-hosted, the answer is “our own VPS in the region we choose,” which is usually enough to close the compliance conversation before it becomes a blocker.

How to choose between them honestly

This is not a universal answer. It is a team capability and volume question. I have talked to RevOps teams where Make is still the right call in 2026, usually because they have a solo ops person, nobody who can manage a Linux process, and fewer than 30,000 operations per month. For them, Make’s operational overhead is genuinely lower even if its per-operation cost is higher. That tradeoff is real.

The migration threshold I use with clients is a simple filter. If you have at least one person on the team comfortable with Docker (or willing to spend an afternoon learning it), and your monthly Make bill is above $80, the math almost always favors switching. The n8n self-hosted setup guide using Docker Compose takes under two hours for a basic production instance. Add an nginx reverse proxy and a free SSL cert from Let’s Encrypt and you have a production-grade setup by end of day.

The workflows that migrate hardest are ones relying on Make-specific features like Watch trigger polling behavior or the advanced error handling modules. The ones that migrate easiest are HTTP Request chains, webhook listeners, and CRM sync flows. That last category is the majority of RevOps automation work anyway.

n8n Self-Hosted vs Make: Which fits your team?

Choose n8n Self-Hosted if

• Your team has someone comfortable with Docker or a basic Linux VPS
• You're processing more than 50,000 operations per month and the Make bill is climbing
• Data residency or compliance is a live concern from your security or legal team

From ~$40/month infrastructure (no per-operation cost) Try n8n →

Choose Make if

• Your ops team is non-technical and visual workflow building is a hard requirement
• You're under 30,000 operations per month and the bill stays manageable
• You need fast time-to-first-workflow without any server configuration

From $9/month (10,000 ops), scales with usage Try Make →

What this shift signals for GTM infrastructure in 2026

The n8n migration trend is part of a larger pattern I am watching across mid-market GTM teams: operators reclaiming control over the infrastructure layer. The SaaS-everything era trained teams to outsource every hosting decision to vendors. But as GTM stacks get more complex (Clay feeding Apollo enrichment into HubSpot properties routing into Smartlead sequences with Gong call data looping back to update lifecycle stage), the cost and compliance overhead of cloud-only automation platforms compounds in ways that eventually become someone’s quarterly problem.

Self-hosted tooling is no longer an engineer hobby choice. It is a legitimate RevOps infrastructure decision that ops leads at Series B companies are now making themselves, without engineering involvement.

G2’s workflow automation category shows n8n’s review velocity accelerating through 2025 and into 2026, and those reviews skew heavily toward operations and RevOps roles rather than developers. That is a meaningful signal. Make built a strong foundation for the first wave of no-code automation adoption. n8n is absorbing the operators who graduated from that wave and need more room to work. According to Make’s own pricing documentation, there is no path to uncapped operations on any plan. At scale, that ceiling is the product.