How to choose a learning project for SEO, YouTube or AI

Which task should you learn on?

You will learn more from one finished piece of work than from a dozen tutorials, so the choice of task matters more than the choice of course. A good learning task is real enough to matter, small enough to complete in the time you have, and based on material you can open yourself and share with a reviewer.

Use a real task only when its material can be shared appropriately. Client data, internal reports and unpublished plans often belong to someone else, and putting them into a training exercise or an AI tool may not be allowed. If that is the case, prepare a training example and label it clearly as one, so that nobody later mistakes it for real data or real results.

The right size depends on the subject. For SEO, a small page map for one section of a website is enough: a handful of pages, the questions each one answers and how they link to each other. For YouTube, a channel brief and a plan for the first video will teach you more than a vague strategy for a year of content. For AI, take one draft workflow and evaluate its output against examples you have checked by hand.

Before you plan the exercise, list what you can access yourself and what depends on another person: analytics access, a CMS account, a channel, an approved data set. A task that stalls for two weeks while you wait for permissions teaches very little.

What should the result look like?

Decide in advance what another person should be able to inspect or use once you have finished. “Learn keyword research” is a topic; “a page map for three service pages, with the search questions each page answers” is an output. The second can be reviewed, discussed and improved, and that is exactly what makes it a learning project.

Write down a few acceptance criteria alongside the output. Is the work complete for the scope you chose? Are the facts accurate, and can the reviewer see where they came from? Is it clear what should happen to this work next, and who would do it? Questions like these keep the review on whether the work answers the original task. Knowing your way around a tool helps, but it is not the point of the exercise.

Keep the first exercise small enough to finish, get feedback on and revise. Plan the revision from the start: the second version, written after a review, is usually where the real learning happens. Keep a note of your mistakes and how you corrected them. That record shows what you have understood far better than a list of completed lessons.

How do you turn one exercise into a method?

After feedback, apply the same method to a related but different case: another service page, another video topic, another workflow. Repeating the exercise on new material shows whether you understood the reasoning or simply followed the instructions the first time round.

As you work through the second case, note which decisions stay the same and which depend on context. In SEO, for example, the way you match a page to a search question stays constant, while the questions themselves change from one service to the next. If you can explain that difference in your own words, the method is genuinely yours.

Keep the reusable parts with the output: the checklist you followed, the sources you used and a short note on why you made each key decision. Next time, you or a colleague can start from there instead of from a blank page.

A note on the Academy itself: our courses are coming soon, and their format and availability will be confirmed before anyone enrols.