Classic 2026 question for every company starting automation: no-code or custom? Four main options, each with a sweet spot.
Zapier — ease at the cost of price
Most popular, simplest. 5000+ integrations, every SaaS has a Zapier connector. Visual "step by step" editor.
Pricing 2026: Free (100 tasks/month), Starter $20/mo (750 tasks), Pro $50/mo (2000 tasks), Team $70/mo, Company $99/mo.
Problem: per-task pricing scales linearly. 10k tasks/mo = ~$400+/mo. 100k tasks/mo = unprofitable ($3-5k).
Make — visual + budget-friendly
Czech Integromat, rebranded Make. More advanced editor (branching, error handling, iterators). Cheaper per operation.
Pricing 2026: Free (1000 ops), Core $9/mo (10k ops), Pro $16/mo (10k ops), Teams $29/mo, Enterprise custom.
Sweet spot: 5k-50k ops/mo, branching logic that is pain in Zapier.
n8n — self-hosted control
Open-source workflow tool, can be self-hosted. Visual editor like Make plus the option to add custom JS code in nodes.
Pricing: $0 self-hosted (your infra cost) or n8n Cloud Starter €20/mo (2.5k executions).
Sweet spot: companies with data privacy concerns, integrations with internal systems, mid scale.
Custom code — full power, full responsibility
Node.js / Python / TypeScript written automation in your infrastructure.
Cost structure: $3-50k setup (one-time) + $50-500/mo infra. Zero per-operation cost.
Sweet spot: scale (100k+ ops/day), specialties (scraping, AI integration, complex transforms), long-term optimization.
ROI breakeven (rule of thumb)
Custom < Zapier when >~3-5k ops/month. Custom < Make when >~50k ops/month. n8n self-hosted < custom when >~10k ops/month and you do not need specialty.