Data Analyst Interview Questions: What UK Employers Actually Ask in 2026

UK analyst interviews test three things, and since 2025 a fourth. Preparing only for the SQL is why capable candidates get rejected.

UK data analyst interviews test three things and, since about 2025, a fourth. Preparing for the SQL alone is why capable candidates get rejected.

Stage 1: The technical test

Usually SQL, sometimes Excel, occasionally a take-home task.

What comes up most: - Write a query joining two or three tables to answer a business question - Aggregate with GROUP BY and filter with HAVING - A window function problem: latest record per customer, ranking within a group, month-on-month change - Spot the bug in a query or formula - Explain the difference between INNER and LEFT JOIN, and when the choice changes the answer

The question that separates people: "How would you know if this result is wrong?" Most candidates have no answer. Prepare one: check row counts before and after joins, reconcile totals to a known figure, sense-check the magnitude.

Excel variants: build a pivot from raw data, XLOOKUP across sheets, find the error in a broken formula.

If a live coding test is involved, narrate as you go and clarify the question before writing anything. Interviewers frequently leave the brief deliberately ambiguous to see whether you ask.

Stage 2: The case study

The classic form: "Sales dropped 15% last quarter. How would you investigate?"

There is no correct answer. They are assessing structure. A workable framework:

Clarify. Fifteen percent against what: last quarter, or the same quarter last year? Which products, regions, channels?
Verify. Is the drop real, or a data issue? A changed definition, a broken feed and a missing region all look exactly like a sales drop.
Decompose. Volume or price? Which segments? All regions or one? New or returning customers?
Hypothesise. Two or three plausible causes, each with the data you would use to test it.
Recommend. What you would do first, and what you would need to be confident.

Step 2 is the one candidates skip and interviewers are specifically listening for.

Stage 3: Behavioural questions

Standard, but with analyst-specific angles:

Tell me about a time you found an error in your own work
A time you disagreed with a stakeholder about what the data showed
How you handled a request you thought was the wrong question
A time you had to explain something technical to a non-technical audience
How you prioritise when three people want something the same day

Prepare three stories in situation-action-result form and adapt them. The error question is asked more often than people expect and the wrong answer is claiming you have never made one. The right answer is a real mistake, how you caught it, and what you changed.

Stage 4: The AI questions, new and increasingly common

AI now appears in 39.6% of UK data analyst adverts, and interview questions have followed:

"Do you use AI tools in your work, and how?" Say yes, specifically. Denying it reads as either dishonest or out of touch. Describe your actual workflow.
"How do you check AI-generated output?" The real question. Answer concretely: verify the join grain, check row counts, reconcile totals against a known figure, and never ship a number you cannot derive yourself.
"Where would you not use AI?" Good answers involve anything with sensitive data, anything where you cannot verify the result, and any situation where you are accountable for a number you do not understand.

Candidates who handle this well stand out considerably, because most either avoid the topic or over-claim.

The portfolio walkthrough

If you are a career changer, expect this and expect it to take a while. See the preparation notes in our portfolio guide. The essentials: four minutes without a screen, and honest answers about what you got wrong first.

A two-week preparation plan

Week 1: Daily timed SQL practice, concentrating on joins, aggregation and window functions. Write out your three behavioural stories. Rehearse the portfolio walkthrough aloud, recording yourself once.

Week 2: Work through three case studies out loud using the five-step framework. Research each employer: their sector, their data maturity, what they would plausibly be measuring. Prepare your AI answer. Prepare two questions to ask them that show you have thought about their business, not questions about the role's day-to-day.

The thing most candidates get wrong

Treating the interview as a test of knowledge rather than a demonstration of how you think.

Interviewers hire analysts they will have to trust with numbers. Everything they ask is a proxy for one question: will this person tell me when something is wrong, or will they hand me a confident answer they have not checked?

Answer that question well and the SQL syntax matters much less than you think.

Frequently asked questions

How long are UK data analyst interviews? Typically two to three stages: a screening call, a technical test or take-home, and a final interview combining case study and behavioural questions.

Should I do the take-home assignment? Usually yes if it is scoped to two or three hours. Anything larger is a reasonable thing to push back on politely. We cover the approach in acing the take-home assignment.

What should I ask them? Questions about data maturity are excellent: where does the data live, who owns quality, how are analytics requests prioritised. They signal experience and tell you whether the job is real analysis or report maintenance.

Mock interviews and portfolio presentation are built into Uptrail's programme rather than treated as optional extras, because the format itself is what most career changers find hardest the first time.

Sources: ITJobsWatch, Data Analyst skills co-occurrence data, 6 months to 1 September 2026.