Automation Platforms
n8n, Zapier and Make.com, built properly
Advanced automation workflows built on the right platform for your business, connecting your tools and eliminating manual work across your entire operation. I choose on fit and cost, not on what is fastest for me to set up.
Picking the wrong platform is expensive in two ways: monthly cost that scales badly, and workflows that hit a ceiling six months in and have to be rebuilt. Choosing correctly at the start is most of the value.
Choosing badly
The two ways this goes wrong
Both are avoidable with an hour of honest thinking before anything is built.
Pricing that scales against you
Per-task pricing looks cheap at low volume and becomes painful at high volume. A workflow that runs twenty steps per lead costs twenty tasks per lead. At a few hundred leads a month that is fine. At ten thousand it is a serious monthly bill, and by then rebuilding it elsewhere is a big project.
Hitting the ceiling
No-code tools are excellent until you need a loop, a real data transformation, error handling with retries, or a call to an API with unusual authentication. At that point people start bolting on workarounds, and the workflow becomes fragile in ways nobody documented.
How I decide: estimate the monthly execution volume before choosing. That single number usually settles the platform question on its own.
The options
Each platform is good at something different
These are genuine trade-offs, not marketing positioning.
n8n
Self-hostable with no per-task pricing, and it runs real code when you need it. My default for complex or high-volume work where per-task costs would spiral. Steeper to learn and someone has to own the hosting if you self-host.
Zapier
The widest app library and the friendliest interface. Fast to build and easy for a non-technical team to understand and maintain. Costs climb with volume, and complex logic gets awkward.
Make.com
A strong visual builder with genuinely good data handling, branching and iteration. Often the best middle ground on price and capability for multi-step scenarios.
Custom code
When no platform fits, Python or JavaScript rather than bending a tool into something fragile. More power and no per-task cost, but it needs somewhere to run and someone to maintain it.
Direct API integration
Where an off-the-shelf connector is too limited or too slow. Full control over authentication, pagination, rate limits and payload shape.
Error handling and monitoring
Retries with backoff, alerting and logging on every workflow, whichever platform is underneath. The most common way automations fail is silently, and that is what I design against.
Straight comparison
n8n vs Zapier vs Make
General guidance. Your volumes and your team decide the answer.
| Consideration | n8n | Zapier / Make |
|---|---|---|
| Pricing model | Per execution, or flat if self-hosted | Per task, so multi-step workflows multiply cost |
| High volume | Strongest option; cost stays predictable | Gets expensive quickly |
| App library | Good and growing, plus generic HTTP for anything | Widest available, especially Zapier |
| Custom code | First-class; run JavaScript or Python inline | Limited code steps, usually on higher tiers |
| Ease for non-technical teams | Steeper learning curve | Easier to hand over and maintain in-house |
| Hosting | Cloud or self-host for full data control | Fully managed, nothing to run |
| Best fit | Complex, high-volume, data-sensitive builds | Straightforward integrations, smaller volumes, non-technical owners |
How it works
From requirement to running in production
Scope
Volume, budget, technical environment and data sensitivity. This is where the platform decision is really made.
Recommend
I tell you which platform fits and what it will cost monthly at your volumes, including when the honest answer is the cheaper one.
Build
Workflows built and tested in a sandbox against realistic data, never against your live systems.
Harden
Error paths, retries, rate-limit handling and alerting added, so failures surface immediately instead of silently.
Hand over
Documentation explaining every workflow, what triggers it and how to change it, plus a walkthrough with your team.
Engagement
What this costs to build and run
Build cost
Fixed price after scoping. A single well-defined workflow is often a couple of days. A connected set spanning several systems runs one to three weeks.
Platform cost
Paid by you directly to the platform, never marked up through me. I size it at your real volume during scoping so the monthly figure is known before you commit.
Ownership
Workflows are built in your account, not mine. If we stop working together you keep everything, fully documented, with no lock-in.
Platforms and tooling I build on
Common questions
Automation platforms, answered honestly
Related
Often built alongside this
Ready when you are
Not sure which platform you need?
Book a free call. Tell me what you are trying to connect and roughly how often it runs, and I will tell you which platform fits and what it will cost you monthly.
Book a free 30-min call →