Make vs n8n: Credits, Executions, and Real-World Automation Billing

When engineering and operations teams evaluate integration platforms like Make (formerly Integromat) and n8n, pricing comparison tables often obscure how billing units actually scale. A platform that appears dramatically cheaper for small prototypes can quickly become punitive when deployed to handle multi-step data pipelines.

The fundamental difference between Make and n8n is not their visual interface or library of connectors. It is their atomic billing unit: Make charges per individual modular operation (credit), whereas n8n Cloud charges per workflow execution.

1. The Anatomy of Make Credits

According to Make’s official documentation (Make Credits Guide and Make Pricing), standard chargeable module actions generally consume one credit (operation) each.

Standard operations adhere to specific characteristics:

• Standard Module Actions: Standard third-party app steps (e.g. creating a record, sending a message, or calling an API) generally consume 1 credit each in our simplified calculator model.
• Filters and Routers Are Not Inherently Chargeable: In Make, filters between modules and router splitters do not consume credits simply for evaluating or routing a bundle; only the resulting module actions on an active path consume operations.
• Variable Credits per Operation (Especially AI): Credits per operation vary across module types. High-compute actions (such as Make AI Assistant, complex text transformations, or premium connector endpoints) frequently require multiple credits per single operation as documented in Make's credits table.
• Additional Polling Checks: When a workflow runs successfully, its trigger step is already counted within runs × steps. However, non-webhook polling triggers (e.g., querying an inbox or datastore every 5 minutes) consume operations on empty polling checks where zero records are returned. The calculator's additional polling input captures these empty cycles.

Make’s reference Core plan quote starts at approximately $9.00/month (annual-billed monthly equivalent) for 10,000 credits. For our 1,000-run × 5-step scenario (5,000 credits), the workload comfortably fits within the 10,000-credit ceiling, making Make an economical choice for moderate workloads with concise steps.

2. The Architecture of n8n Executions

In contrast, n8n Cloud measures workflow executions. A workflow execution begins when a trigger node (such as an incoming webhook, a scheduled cron trigger, or an app event) fires and completes all associated nodes within the workflow canvas.

Whether your workflow contains 3 nodes or 35 nodes, one completed workflow trigger counts as exactly one execution on n8n Cloud. In our 1,000-run baseline, n8n registers exactly 1,000 executions against its plan ceiling.

n8n Cloud’s reference Starter tier begins at €20.00/month (billed annually, equivalent to approximately $22.00/month at an illustrative 1.10 EUR/USD exchange rate) and includes 2,500 executions per month. The 1,000-execution baseline consumes 40% of the included quota.

3. Where the Billing Curves Cross

Because Make scales with steps and n8n scales with runs, the cost advantage depends heavily on your pipeline's architectural profile:

Scenario A: Deep, Multi-Step Pipelines

Suppose your workflow expands to 18 steps (e.g., input sanitation, three API lookups, data mapping, branch routing, notifications, and error logging).

• At 1,000 runs, Make consumption surges to 1,000 × 18 = 18,000 credits. This exceeds the 10,000-credit Core tier and forces an upgrade to a larger plan tier.
• On n8n Cloud, the execution count remains 1,000 executions, still well within the 2,500 Starter ceiling.
• Takeaway: Complex, step-heavy enterprise automations become dramatically more cost-predictable under per-execution billing.

Scenario B: High-Volume, Single-Step Webhooks

Conversely, consider a high-throughput event logger that receives 20,000 webhooks per month and performs a single lightweight append to a datastore.

• On n8n Cloud, 20,000 executions far exceeds the Starter tier (2,500 executions) and Pro tier quotas, requiring higher enterprise cloud tiers.
• On Make, 20,000 runs with 1 step consume 20,000 credits, which can be acquired on incremental credit tiers at lower monthly price points.
• Takeaway: High-frequency, single-action event routing often favors per-operation pricing over execution-capped tiers.

4. Retries, Polling, and Failure Economics

Real-world automations must handle transient network faults and third-party API rate limits. When a step fails and automatic retry mechanisms fire:

If a retry reruns the full workflow, both Make credits and n8n executions increase by the retry percentage. Our engine models this as Math.ceil(runs × (1 + retryPercent / 100)). In addition, idle polling in Make consumes baseline credits even during business holidays, whereas webhooks in n8n only count when real traffic lands. Note also that multiplying monthly figures by 12 represents an annualized illustration (monthly × 12), not necessarily an upfront contract requirement for month-to-month quotes.

5. Model Limitations & Reproducibility

This analysis is an independent mathematical simulation based on published vendor documentation verified on October 10, 2026. It is not a live performance benchmark of vendor server response times, latency, or uptime.

All calculations in this guide are directly reproducible using the open client-side engine in public/engine.js. Always obtain customized quotes from vendor sales representatives for high-volume enterprise contracts.