How many times a month does this run?
This is the one that usually settles it. A workflow firing fifty times a month and one firing fifty thousand times belong on different platforms, and the second one gets expensive fast on the wrong choice.
← Back
Choosing a platform
Zapier, Make or n8n, and how I pick.
People usually arrive having already decided which one it is. Sometimes they're right. I'm certified in all three, which mostly means I have no reason to sell you the one I happen to know.
The short version
What each one is actually good at
Zapier
Fastest to get running and by far the biggest library of apps, so it's the right call when the job is mostly connecting SaaS tools that already have a Zapier app sitting there. Costs climb as volume climbs, and that's the thing to model before you commit.
Make.com
Handles branching better and costs less per operation once you're moving real volume. You can also see the whole shape of a workflow on a canvas, which matters more than people expect when you're the one maintaining it a year later.
n8n
What I reach for when there's proper logic involved, when you want to self-host it on your own infrastructure, or when the workflow needs actual code inside it. It's also what I'd pick for anything with AI steps.
The questions that actually decide it
Four things I ask before picking
Does it branch, or does it go in a straight line?
Straight lines are easy anywhere. The moment there are conditions, loops and exceptions, the platform's ability to express that clearly becomes the difference between maintainable and a mess.
Who maintains it after I'm gone?
If somebody non-technical on your team needs to make small changes, that argues for a different tool than if it's going to sit untouched or be handed to a developer.
Where does the data have to live?
If you have rules about data leaving your infrastructure, that rules out the hosted options and points at self-hosting, which changes the answer immediately.
The honest part
Sometimes none of them fit
The order system I built for a subscription kitchen is custom code, not any of these three. Orders arrived in five different shapes including spreadsheets emailed by a nutritionist, and no visual platform was going to parse that cleanly. Forcing it into one would have produced something fragile that broke the first time a spreadsheet came in slightly differently.
That's a real cost either way, and it's worth being clear about it. Custom code means fewer people can pick it up later. A visual platform means a subscription and a ceiling. Which cost you'd rather carry depends on your team, and that's a conversation rather than a rule.
I'd rather lose the job than build it in the wrong tool, because I hand it over and leave, and you're the one living with that choice every month afterwards.
Not sure which one your job needs?
That's genuinely fine, and it's the first thing we'd work out. Twenty minutes, and I'll tell you which platform fits and roughly what it costs to run.
