How People with Less SQL Than You Win the Interview
The four step framework they use to land interviews.
Dear friends,
Last week I asked you to build one small document and reply with the most interesting thing you found.
Almost nobody did.
I am not going to pretend otherwise, because I think I know why, and it is not laziness.
You did not know what the document was for.
I told you to build a portfolio piece and I never told you two things. What a portfolio piece actually does in the room where the decision gets made, and the shape it has to be built in to do it.
So today you get the framework.
the question every interview contains
Every data interview on earth contains one question.
Walk me through a project.
That question destroys people who know more SQL than you. It destroys them because they prepared for a quiz and got asked for a story. They say um. They describe a course they finished. The interview is over at that moment, it just takes another twenty minutes to end.
I have sat on the other side of that table. It is not that the person is bad. It is that they have nothing to walk anyone through.
the four steps
So here is the framework, and the important part is that you build it into the project itself. You are not learning an interview trick. You are building the portfolio piece in the shape that answers the question before it gets asked.
Step 1. Frame it as a business problem, not a dataset.
Before you write a single query, put one sentence at the top of the document about who needs this and why.
“An online retailer needed to know which regions were slipping. The data was a legacy export with three date formats and duplicate orders.”
Build that line in and it becomes your opening line later. One breath, and the interviewer knows you have touched real-shaped data.
Step 2. Make one decision, and write down why.
Not everything you did. The one place you had to choose.
“I had to define what churned even meant. I went with regulars, three or more orders, silent for 90 days.”
Most portfolios have no decisions in them, which is exactly why they all look the same. Decisions separate an analyst from someone following a tutorial, and the reason you wrote down is the part you say out loud later.
Step 3. End on a number, not a chart.
“South region, down 49% on the year.”
A finding is a sentence with a number in it. If your document ends in a dashboard and no sentence, you have built a screenshot, not proof.
Step 4. Write one line on what you would do next.
Every real analysis is unfinished. Saying what is next proves you know that.
Now read those four back in order.
You just answered walk me through a project.
That is the whole trick. Build the piece in the four steps and the interview answer comes out of it for free, in about ninety seconds, from someone who never had to memorise anything.
build one tonight
Take the weekend document, or start from the last issue’s queries if you never got to it.
Put the four steps into it. One sentence of business framing at the top. One decision, written down with the reason. One finding with a number in it. One line on what is next.
One page. About an hour.
Then read it out loud, top to bottom.
Out loud, not in your head. It will feel ridiculous the first time. Do it anyway, because the first time you say it out loud should not be the day someone is deciding whether to hire you.
And notice what you cannot do. You cannot build this without a project. There is no certificate version of it. There is no course-completion version of it.
That is the whole reason the small document matters.
if you’re already working as an analyst
The same four steps are how I answer “what have you been working on” from my own manager.
Frame. Decision. Number. Next.
I keep a receipts file, one entry a week, in exactly that shape. Half of each entry is drafted by Claude from my rough notes before I edit it, because the writing was never the hard part and the remembering is.
The analysts I see getting promoted are not the ones doing the most work. They are the ones who can account for it in ninety seconds when someone senior asks.
If you keep a brag doc or a receipts file, hit reply and tell me what is in it and how you structure it. I am collecting these and I share what I learn here.
what happens next
For months the question in my inbox has been some version of the same one.
I know the tools, so why am I still stuck?
The answer is that skills are invisible and proof is not. Nobody hires skills. They hire proof.
So I built the complete version of what you have been tasting in these emails.
Three guided projects, one company, real messy data. Excel first, then SQL, then a Power BI capstone. The kind of data that arrives with three date formats in one column and prices stored as text, because that is what actually lands in your inbox on a Monday.
All three are built in the four steps. The business framing is written for you, the hard decision is the thing the project makes you choose, and every project ends as a resume bullet, a LinkedIn post, and a rehearsed answer. The words, not just the work.
And the AI-native workflow underneath all of it, with the prompts included. Not to cheat. To work the way the people getting hired already work.
It is called The First Job Portfolio.
It is finished and I am doing the last checks on it now.
The next email you get from me is the one where it opens. That will be the only time I properly sell you anything this quarter, and then we go back to building things together.
Until then, build one page in the four steps and read it out loud.
Six months of consistency beats six days of motivation.
Talk soon,
Stanley