Are you ready?

The check has 17 questions in four parts, with a 20-minute timer:

• SQL (6 questions). Filtering and sorting, GROUP BY with HAVING, an INNER JOIN, a LEFT JOIN to find missing records, CASE, and grouping dates by month. You write the query, run it as often as you like and see the first rows of the result.
• Spreadsheets (4). Lookups, SUMIFS, absolute and relative references, and pivot tables with messy data.
• Power BI and DAX (4). Measures against calculated columns, star schemas and relationships, CALCULATE and filter context, and choosing a visual.
• Charts and data sense (3). Reading a value and working out a percentage change, spotting a misleading axis, and telling correlation from causation.

How SQL is marked. We run your query and compare its result with the right answer, so any correct way of writing it counts: a LEFT JOIN, NOT EXISTS or NOT IN all score the same. Column names don’t matter, row order only matters when the question asks for an order, and numbers are compared to two decimal places. If your query returns extra columns alongside the right ones, it still counts, though in an interview you’d return just what’s asked.

How it’s scored. Each question is worth the same. You get a percentage for each part and overall: 75% or more is “Strong foundations”, 50% to 74% is “Getting close” and under 50% is “Not yet”. Skipped and unanswered questions count as not correct. SQL is what junior interviews test most, so we also look at your SQL score before suggesting a call with a coach.

Your three fixes come from the questions you missed in your weakest parts. If the time runs out before you reach several questions, the first fix is about speed.

Its limits. Twenty minutes can’t cover everything a job needs: it doesn’t test Python, statistics or how you explain your work, and real tests use bigger, messier data. The questions are general practice questions written for the Toolkit, not taken from any employer’s test. Use your result as a guide to what to practise next.