The familiar closed loop: to get work you need a portfolio, and to build a portfolio you need work. Most people try to break it by working free, which occasionally works and usually costs a lot of time.
There is a cleaner route. Do real work on real data and real companies, without waiting for anyone to give you a chance.
First, what a portfolio actually proves
Many people treat a portfolio as a gallery of good-looking work. It is not. Whoever is considering hiring you wants to know one thing: how do you think?
A well-designed ad image says nothing on its own. But if next to it you have written why you chose that message, what alternative you rejected and why, and what you would have measured, that says everything.
The working rule: every project in the portfolio needs a decision and a reason attached. Without that, they are pictures.
Project 1: an analysis on real data
The biggest problem with beginner projects is invented numbers, and anyone working in the field spots them immediately. The fix is that Google provides real data to anyone, free.
According to Google's own documentation, the Analytics demo account is a fully functional account any Google user can access, containing real data from the Google Merchandise Store and the Flood-It mobile game. That is genuine ecommerce data with real traffic sources and transactions.
What to do with it: pick one specific commercial question, such as which traffic source produces actual revenue rather than just sessions, work it through, and write the result on a single page. The question, the answer, the recommended decision.
To be straight with you about the limits, the documentation is explicit: you get a Viewer role only, you cannot export reports or explorations, and it is not available through the Analytics Data API. So you will take screenshots and write the analysis yourself, which is sufficient for this purpose.
Project 2: an audit of a company that exists
Pick a small company in a sector you understand, ideally a real one in your local market. Write an audit: their pages, their messaging, how clear the offer is, how fast they respond, the gap between what they say and what a customer needs to hear.
One important caveat. Do not send it to them as unsolicited criticism, and do not publish it in a way that embarrasses them. This project demonstrates observation, and handling it professionally is part of what is being assessed.
What makes it strong is that it shows you notice things other people have missed. That is the skill people actually pay for.
Project 3: a small campaign with your own money
This is the strongest of the three, the hardest, and the one most often skipped. Spend a genuinely small amount on a real campaign for something real: a product you sell, a page you are trying to grow, a service for someone you know.
The amount does not matter. What matters is that you have been through the experience of making a decision, spending money, and not controlling the outcome. It changes how you talk about advertising in a way that is obvious to anyone listening.
If you are not sure this is the right moment to be spending at all, read when to start spending on ads first.
How to present them
- One page per project. Not a twenty-slide deck.
- Open with the question. What were you trying to find out or fix.
- State the decision and the rejected alternative. This is the part that differentiates you.
- Report the result honestly. Including when it was bad.
- Close with what you would do differently.
That last point works unexpectedly well. The reader sees someone who learns from their own work, which is rarer than someone with a good result.
How long this actually takes
Worth asking, because an optimistic answer makes people quit after a week. These three projects take roughly four to six weeks if you are working on them alongside a job or studies.
The analysis can be done in a week. The audit, another week. The campaign is what takes time, because it has to run long enough for there to be anything worth saying about it. That is a lesson in itself: anyone promising you an advertising result in three days is either confused or selling something.
If you do not have four weeks, do one project and do it properly. One well-written project beats three half-finished ones, which is the same logic as the order you learn in.
What a reviewer is actually looking for
From having reviewed a lot of this work, three things get judged quickly, before anything else.
First: does this person understand the business or only the platform? When someone talks about margin or cost per customer rather than about likes, it registers immediately.
Second: can they say they do not know? Anyone claiming everything worked is visible from a distance, and anyone saying this worked and that did not earns trust much faster.
Third: is the writing organised? If the project page itself is unclear, the reasonable assumption is that the work will be too.
What to leave out
Course projects built from the same brief every other graduate of that course used. Whoever is reviewing has seen it a hundred times and recognises it instantly.
And numbers without context. "Increased engagement 300%" from 2 to 8 is technically true and misleading, and when it is noticed it takes the rest of your work down with it.
The honest limit
This portfolio gets you the interview. It does not get you through the interview. It is weaker than real experience with a paying client, and anyone telling you otherwise is selling you something.
It is still far better than waiting for an opportunity, and it means you walk into a first conversation with something to discuss rather than a description of your intentions.
If the goal is freelancing rather than employment, the next step is different and I covered it in how to start freelancing in marketing. If you want to know which skills are worth investing in first, there is the skills the market is asking for, and how to measure your own work is in how to measure marketing results.
Once you have a portfolio, finding the first client is a different problem entirely. BDG Labs has a practical treatment of where early clients actually come from in five real sources of first customers, aimed at business-to-business selling.