The part of marketing we’ve normalised for too long.

It’s Tuesday morning, and you’ve just sat down with your coffee. By the time the coffee has gone cold, you’ve answered the same enquiry email three times, exported the same numbers from the same dashboard, and lined up the same post across three platforms. The morning is gone, and the work that needs your thinking is still waiting.

That morning is mine, and chances are you've had your own version of it. The longer I've done this work, the more I've noticed how much of it could be running on its own while we sleep.

And yes, before you say it: I know, I should've automated that email a year ago.

So, what is marketing automation? How does it work? And where do I start?

Where this gets interesting.

If you sit down at the end of any given week and write out where the hours went, a familiar pattern shows up. A surprisingly large share of marketing work, mine included, is repetitive: copying numbers between tools, sending the same email with slight tweaks, lining up posts across platforms, pulling the same report on a Monday morning.

What gets squeezed is the work that a human needs: the thinking, the judgment calls, the taste decisions, the choice of what to say to whom. Automation, at its core, is the act of handing the repetitive work back to a system so the thinking work has room to happen.

The reason it's worth paying attention to now, more than it was a few years ago, is that the tools have caught up. The platforms that wire all of this together have become cheap, friendly, and quick to learn. AI sits inside them and handles the small judgment calls that used to demand a person. The barrier has dropped sharply.

Once you see the pattern, you see it everywhere.

The first time I built one of these end to end, the pattern clicked for me in a way no diagram had managed. It's hard to see until you've made one yourself. After that, it's hard to unsee.

Underneath the platforms, every piece of marketing automation comes down to the same simple loop. Something happens, the system reacts, and an outcome lands somewhere useful. A new lead arrives, gets scored, and shows up in the right inbox before you've had your coffee. A blog post goes live, and a queue of social posts and a newsletter draft turn up in your task list.

Behind the scenes, three pieces are doing the work.

Workflow tools, like n8n, Make, or Zapier, are the wiring. They connect everything together and decide what should happen next.

AI handles the bits that need judgment, such as scoring a lead, classifying an enquiry, summarising a long thread, or drafting a first version of a reply.

Integrations are the connections to the tools you already use: your CRM, your email platform, your ad accounts, your analytics.

Those three pieces cover everything. Even the most complicated marketing systems out there are simply more layers of this same loop, with smarter triggers and richer judgment in the middle.

The work you can hand to a system tomorrow.

There are a handful of jobs that show up again and again as the first thing people automate. They sit quietly in the background of a normal marketing week. The five below are where most people start, and once the first one is up and running, the time it gives back tends to make the next ones obvious.

You’ve probably just read that list and recognised at least two of them.

These five are starting points. There are dozens more, and once you've wired up the first one or two, the next ones tend to suggest themselves. The pattern is the same every time: pick the one task that quietly eats your week, and start there.

Where to put your first hour.

If I were starting from zero today, here's what I'd do. Pick one small task you already do every week, something repetitive, like sending the same enquiry reply, pulling the same weekly report, or copying a blog post into three platforms.

Before opening any tool, write the steps out by hand. Map out the trigger, what happens in the middle, and where the outcome should land. Doing this on paper saves a surprising amount of time later, because the moment you start building, the gaps in your thinking show up.

Then build the simplest version in a tool like n8n or Zapier. Get it working before making it clever. Once it runs, you can add a piece at a time, the lead score, the personalisation, the AI step, whichever feels most useful next.

The aim at the start is to get one automation running end to end. Systems, where multiple automations talk to each other, come later, once you've got a couple of working pieces to connect.

A note before I go.

Most of what I've written here is what I'm working through in my own week. Writing it down is how I make sure the learning sticks, and how I figure out what I'm still missing.

The next entry takes the picture one step further: what changes once AI sits inside this loop, and how the same trigger and response shape gets a lot smarter when you let a model do the judgment.

If you want it to land in your inbox, the newsletter is the easiest way to follow along :)

Hope to see you on the other side.

Speak soon,

Ryan.