Three platforms, three philosophies, three targets
In 2026, n8n, Make (formerly Integromat) and Zapier account for more than 70 percent of the no-code automation deployments we observe in small and mid-sized companies in France. They are the three platforms most cited in IT scoping requests, the three most searched on Google in France (more than 25,000 searches a month combined), and the three that really shape the market. Yet behind the shared look (visual interface, drag-and-drop, integration catalog), they embody three radically different philosophies.
Zapier is the heritage of mainstream no-code: maximum simplicity, a massive catalog of prebuilt integrations, and an explicit focus on business users without a technical background. Make is the value-for-money option: more powerful than Zapier on complex workflows, significantly cheaper at equivalent volume, with a somewhat steeper learning curve. n8n embodies sovereign low-code: open source, self-hostable, full control, ideal for organizations that demand independence and control over their data, but it requires minimal technical skills.
This structured comparison, based on more than 30 deployments we observed in small businesses in France between 2024 and 2026, analyzes each platform on 10 criteria, projects real costs across 5 volume tiers and offers a final decision matrix. It is part of our complete guide to no-code automation in business, the pillar page of this series.
Overview: positioning and DNA
Before comparing point by point, understanding each platform’s DNA avoids 80 percent of selection mistakes. Here is a quick ID card for all three.
Zapier
Founded in 2011, USA, the historic no-code leader. SaaS only. More than 7,000 prebuilt integrations. Explicit target: business users.
DNA: maximum simplicity, zero friction, the gentlest learning curve on the market.
Make
Launched in 2012 as Integromat, Czech Republic, acquired by Celonis in 2020. SaaS, EU and US data centers. 1,800+ integrations.
DNA: value for money, scenario-based visualization, more powerful than Zapier on complex workflows.
n8n
Launched in 2019 by Jan Oberhauser in Berlin, Germany. Source-available (Sustainable Use License, part of the fair-code movement). Available in the cloud or self-hosted for free. 1,300+ native nodes.
DNA: sovereignty, control, open source, a blend of no-code and low-code (inline JavaScript/Python).
Detailed scoring on 10 criteria
Each platform is scored from 1 to 5 on 10 key criteria. Scores come from hands-on observation across 30+ small business projects in 2024–2026, not from vendor marketing. The maximum possible score is 50.
| Criterion | Zapier | Make | n8n |
|---|---|---|---|
| 1. Ease of getting started | 5/5 | 4/5 | 3/5 |
| 2. Native integration catalog | 5/5 | 4/5 | 3/5 |
| 3. Power on complex workflows | 3/5 | 4/5 | 5/5 |
| 4. Cost at medium volume (10k exec/month) | 2/5 | 4/5 | 5/5 |
| 5. Cost at high volume (100k exec/month) | 1/5 | 3/5 | 5/5 |
| 6. GDPR and data sovereignty | 3/5 | 4/5 | 5/5 |
| 7. Support quality and SLA | 4/5 | 3/5 | 3/5 |
| 8. Community and resources | 5/5 | 4/5 | 4/5 |
| 9. Generative AI integration | 4/5 | 4/5 | 5/5 |
| 10. Independence and portability | 2/5 | 3/5 | 5/5 |
| Total / 50 | 34 | 37 | 43 |
Beware the overall-score trap: n8n totals 43/50 but loses 2 points on ease of getting started, which can be a deal-breaker for a business team without IT. Zapier scores lower but remains the right choice for a small business that prioritizes speed of setup on a modest volume of simple workflows. The total reflects raw power, not fit with your context.
Pricing comparison across 5 volume tiers
Cost is the criterion where the gaps are most spectacular. The table below shows typical annual costs we observed in France (in EUR, excluding VAT) for 5 monthly execution-volume tiers representative of French small and mid-sized companies, based on public price lists in spring 2026: budget estimates, not US list prices. For reference, US list prices as of November 2026: Zapier Pro is $19.99 a month billed annually for 750 tasks, $129 a month for 10,000 tasks and $489 a month for 100,000 tasks; Make Core is $9 a month billed annually for 10,000 credits; n8n publishes its cloud prices in euros only (EU pricing, €): Starter €20 a month billed annually for 2,500 executions, Pro €50 a month for 10,000, and the self-hosted Community Edition is free.
| Monthly volume (France estimates, €) | Zapier | Make | n8n Cloud | n8n Self-hosted |
|---|---|---|---|---|
| 1,000 exec/month | €240/yr | €120/yr | €240/yr | €120–360/yr (VPS) |
| 10,000 exec/month | €900/yr | €240/yr | €600/yr | €120–360/yr |
| 50,000 exec/month | €3,600/yr | €720/yr | €1,200/yr | €240–600/yr |
| 100,000 exec/month | €7,200/yr | €1,440/yr | €2,400/yr | €360–840/yr |
| 500,000 exec/month | €25,000+/yr | €6,000/yr | €9,600/yr | €600–1,800/yr |
Three major lessons stand out. First, below 1,000 executions a month, the gaps remain manageable and the choice comes down to other criteria. Second, from 10,000 executions a month, Make becomes 60 to 75 percent cheaper than Zapier for equivalent functionality in the French estimates above; at US list prices the per-unit gap is even wider ($129 a month for 10,000 Zapier tasks versus $9 for 10,000 Make credits, billed annually), although the two units aren’t identical. Third, between 50,000 and 100,000 executions a month, self-hosted n8n becomes roughly 6 to 20 times cheaper than Zapier and 1.2 to 4 times cheaper than Make according to the same table, savings that easily fund the in-house technical skills hosting requires.
A point of vigilance that’s often underestimated: the billing unit matters as much as the list price. At Zapier, each action within a Zap consumes a separate task (the trigger itself isn’t counted), so a workflow with one trigger and 4 actions consumes 4 tasks per run, which pushes real consumption well above the apparent volume. Make bills in credits (one per module action, formerly called operations; some AI modules use more) and n8n per complete workflow execution, two units that are considerably more predictable for budgeting ahead.
The TCO of self-hosted n8n must also include a hidden cost: system administration time. Without in-house technical skills, plan on 2 to 4 hours of maintenance a month (updates, backups, monitoring), time that can be outsourced to a fractional CIO managing the infrastructure as part of a broader engagement. To estimate the ROI precisely in your context (volume, time saved, team hourly cost), our automation ROI calculator simulates the gains in a few minutes.
Detailed strengths and weaknesses of each platform
Zapier: the standard of mainstream no-code
Strengths
- • The gentlest learning curve on the market, first workflow in under 30 minutes.
- • A catalog of 7,000+ prebuilt integrations, the largest in the industry.
- • Extensive documentation, a massive community, abundant YouTube and blog resources.
- • Responsive support on the Pro plan and above, with a contractual SLA.
- • Excellent for simple workflows (1:1 sync, notifications, basic alerts).
Weaknesses
- • Price that skyrockets beyond 5,000 tasks a month, the most punishing pricing model.
- • Pricing revised several times in recent years (three changes in two years observed in the field), a variability that complicates medium-term budget forecasting.
- • Limitations on complex workflows (nested loops, advanced data transformations).
- • US hosting by default, GDPR coverage through the Data Privacy Framework only.
- • Strong vendor lock-in, no portable export to competitors.
- • No self-hosting or sovereign option, total dependence on a US SaaS.
Make: unbeatable value for money
Strengths
- • Pricing model 60 to 75 percent cheaper than Zapier for equivalent functionality (our French estimates).
- • Powerful visual scenario interface, ideal for complex branching workflows.
- • Native handling of data transformations, loops and complex conditions.
- • EU data centers (Germany, Czech Republic) with data residency, GDPR-native.
- • Good balance between power and accessibility for analytical profiles.
Weaknesses
- • Learning curve 2 to 3 times longer than Zapier’s for non-technical users.
- • Smaller integration catalog (1,800+ vs 7,000 at Zapier).
- • Support and documentation essentially in English (a drawback for non-English-speaking teams).
- • No self-hosted version, dependence on Celonis’s SaaS.
- • The credit model (one workflow consumes several credits) can be surprising.
n8n: open-source sovereignty
Strengths
- • Free self-hosting, massive savings beyond 10,000 executions a month.
- • Full control over data, ideal for regulated sectors or sensitive data.
- • A blend of no-code and low-code through inline JavaScript and Python nodes.
- • Leader in AI agent integration through the AI Agent node (LangChain).
- • Source code available, active community, independence from a single vendor.
- • Longevity ensured by open source: the code remains accessible and usable even if the vendor changed its pricing policy or disappeared, a guarantee no proprietary solution can offer.
Weaknesses
- • The steepest learning curve of the three, requires a technical background.
- • Self-hosting requires DevOps skills (Docker, monitoring, backups).
- • Smaller native integration catalog (1,300+), some through the generic HTTP node.
- • Community support for the self-hosted version, professional support only on the cloud.
- • Sustainable Use License, not open source in the strict OSI sense.
Generative AI and autonomous agents: who leads in 2026?
Since 2024, native generative AI integration has become the differentiating criterion of this new generation of platforms. All three leaders now offer connectors to OpenAI, Anthropic Claude, Google Gemini and Mistral, but with distinct levels of maturity as of May 2026.
n8n is currently the most advanced on AI agents. Its AI Agent, AI Chain and AI Memory nodes (based on LangChain) let you build complete autonomous agents with tools, persistent memory, retrieval-augmented generation and multi-step reasoning logic directly in the visual interface. Another differentiating asset: native Ollama support for running LLMs locally, a valuable option for organizations that want to keep their data end to end on their own infrastructure without ever going through a third-party API. For small businesses that want to use agentic AI beyond a simple LLM call, n8n offers the most mature experience on the market.
Make offers AI Assistant modules and a solid catalog of LLM connectors. It is very good for classic AI use cases (summarizing, classifying, generating text) and lets you orchestrate chains of LLM calls in its visual scenarios. It lacks, however, the advanced autonomous-agent dimension n8n offers.
Zapier has launched Zapier Central and AI Actions, which greatly simplify adding AI to workflows for business users. The platform is ideal for automating single-LLM tasks (one call, one answer) without worrying about technical complexity. For complex AI use cases (multiple LLMs, persistent memory, RAG), Zapier remains behind n8n.
A note on GDPR compliance (when you process EU personal data): for AI workflows handling personal data, favor LLMs with EU residency (Mistral La Plateforme, OpenAI EU data residency, Anthropic Claude via AWS Bedrock in the EU). All three platforms let you configure these endpoints.
3 typical small business profiles and the recommended platform
Rather than a one-size-fits-all recommendation, here are 3 small business profiles we see in consulting, with the platform that maximizes their chances of success.
Profile 1: sales-driven small business, 15–40 employees, no IT
Typical context: marketing agency, consulting firm, B2B distributor, e-commerce growth team, early-stage SaaS. No in-house IT team, automations built by a sales ops person or a marketer. Initial volume of 500 to 5,000 executions a month, projected at 10,000 a month within 18 months.
Recommendation: Zapier to get started quickly and prove the value, with a plan to move to Make after 18–24 months if volume explodes and workflows become more complex.
Rationale: initial simplicity matters more than optimal cost. A Zapier workflow’s ROI remains excellent in year 1. A future migration to Make will be easier if it’s anticipated from the start through careful workflow documentation.
Typical workflow examples: CRM-to-Slack sync, automatic sending of invoices from the billing tool, automated client onboarding, routing of incoming leads. All connections that can be deployed in a few hours thanks to the prebuilt integration catalog.
Profile 2: manufacturing or services company, 40–150 employees, with an IT manager
Typical context: manufacturing SMB, distribution, multi-store retail, accounting firm, IT services company, construction, an organization with a part-time IT manager or CIO and an active business team. Volume of 5,000 to 50,000 executions a month, multiple CRM-ERP-accounting-HR integrations.
Recommendation: Make as the core platform, complemented by self-hosted n8n for very high-volume workflows or those containing sensitive data.
Rationale: Make offers the best price/power/GDPR compromise for this profile. Its visual scenario interface handles business complexity. EU data residency makes GDPR compliance easier without adding governance overhead.
Typical workflow examples: multi-channel lead capture (forms, chatbot, social networks) with automatic enrichment and scoring, routing to the right contacts, integration between the ERP, the marketing automation tool and reporting spreadsheets.
Profile 3: mid-sized company with 150+ employees or a regulated sector
Typical context: industrial mid-sized company, banking/insurance, healthcare, defense, public sector, telecom operator, an organization with a full IT department, strict GDPR requirements, high volumes (100,000+ executions a month), critical data confidentiality.
Recommendation: Self-hosted n8n as the structural target, deployed in-house on sovereign infrastructure (OVHcloud, Scaleway, private data center).
Rationale: full control over data and independence from a US SaaS vendor outweigh simplicity. The zero license cost largely offsets the DevOps investment. Over 3 years, TCO is 60 to 80 percent lower than SaaS alternatives, with a significantly stronger security posture.
Typical workflow examples: automation of the order process (ERP to CRM to invoicing), AI-driven document processing, orchestration of dozens of interconnected workflows with granular permissions per team and SSO/SAML. See our dedicated page on data and ERP integration.
Final decision matrix
This matrix summarizes the recommendations by crossing the 4 most decisive selection dimensions.
| Main criterion | Volume | Recommended platform |
|---|---|---|
| Maximum simplicity, business users | Any volume | Zapier |
| Value for money, mixed team | 5k–100k/month | Make |
| Sovereignty, sensitive data | Any volume | Self-hosted n8n |
| High volumes and optimal cost | 50k+/month | Self-hosted n8n |
| AI agents and advanced automation | Any volume | n8n |
| Maximum integration catalog | Any volume | Zapier |
The pattern emerging in 2026 among mature small and mid-sized companies in France is hybridization. 35 to 40 percent of the organizations we observe now run 2 platforms in parallel: typically Zapier or Make for simple business workflows built by business teams, and self-hosted n8n for critical, high-volume or sensitive-data workflows. This dual approach maximizes business productivity without sacrificing technical sovereignty.
Going further on this topic
This comparison is part of our full series on no-code automation in business. To dig deeper, see the related articles:
- No-code automation in business: the complete 2026 guide (pillar page of the series: platform overview, 12 use cases, 6-phase methodology).
- Cost of a no-code automation project for a small business (companion article): detailed pricing grid, 3-year TCO model, ROI calculation workflow by workflow.
- 6-phase no-code implementation methodology (companion article): typical schedule, deliverables, pitfalls to avoid, key success factors.
- 12 concrete n8n workflows for small businesses (companion article): catalog of ready-to-implement workflows, screenshots, copy-ready code.
Frequently asked questions
Can you start for free on n8n, Make or Zapier?
Yes, all three platforms offer a free plan with different limits. Zapier’s Free plan offers 100 tasks a month with Zaps limited to two steps (one trigger and one action), enough to test but not enough for real use. Make offers 1,000 credits a month and 2 active scenarios, which already covers simple production workflows at very low volumes. n8n is entirely free when self-hosted (you only pay for your server, about €10–40 a month) and offers a 14-day trial on the cloud. To prove the value in a POC, start with the free plans before subscribing to a paid one.
Is Make the same as the old Integromat?
Make is the rebranded new version of Integromat, deeply overhauled in 2022 by its parent company Celonis. The interface was modernized, performance improved, the pricing model revised and many connectors added. Integromat users were migrated gradually to Make between 2022 and 2023. In 2026, Integromat no longer exists as a separate product, but the name still sometimes comes up in documentation and French-speaking communities out of habit. Old Integromat scenarios still run on Make without modification.
Which platform has the best connectors for French tools?
Zapier and Make have historically had the most native connectors for niche French tools: Sellsy, Axonaut, Pennylane, Pipedrive (international but very popular in France), Brevo (formerly Sendinblue), Cegid, Sage France. n8n is gradually catching up on this front but still sometimes requires building the integration through the generic HTTP node by calling the REST APIs directly. For a company in France where 60 percent of the tools are French-made, Make or Zapier will limit the initial integration investment.
What is the real reliability (uptime) observed on these 3 platforms?
All three platforms show very high availability in practice. Zapier publishes an average uptime of 99.98 percent over the last 12 months, Make 99.95 percent, n8n cloud 99.9 percent. For self-hosted n8n, availability depends entirely on the quality of hosting and monitoring: with careful infrastructure (monitoring, alerts, backups), 99.9 percent is easily reached; without monitoring, it can drop to 98 percent in the event of an undetected system incident. For business-critical workflows, always prepare a continuity plan (redundant workflows or manual failover).
Can you easily migrate from Zapier or Make to n8n?
Migration is feasible but requires manually rebuilding the workflows: there’s no automatic import tool between these platforms. The logic remains similar from one tool to another (trigger, actions, conditions), which makes the transposition easier. For a typical switch of 10 to 20 workflows, plan on 2 to 4 weeks with technical support. The savings on recurring costs generally pay back the investment in 3 to 6 months, a calculation all the more favorable as execution volume grows.
Can you get help with the choice and implementation?
Yes, it’s the approach I offer as a fractional CIO: scoping the automation program (mapping candidates, choosing a platform through a parallel POC, calculating 3-year TCO), steering the rollout (pilot launch, super-user training, governance) and support through industrialization. A typical engagement runs 6 to 12 months, with a path to full autonomy for the company at the end. The engagement’s ROI is typically recouped in 6 to 9 months thanks to the savings generated by the first workflows deployed.