
Typical stages (varies by company)
- A conversation about your background and interest in the company
- A conversation about past analysis work
- A technical or case exercise, often live, sometimes take-home
- A conversation with the team or stakeholders
Smaller companies may combine these. Ask the company which stages to expect.
Prepare your project stories
Choose two or three analyses you have done and be ready to explain:
- the question and who needed the answer
- where the data came from and what you did to check it
- the method you chose and why
- what you found, what you recommended and what happened next
Use the STAR method to keep answers focused.
Practise the technical part
- Revisit the tools named in the advert, such as SQL, spreadsheets, Python or R, and a dashboard tool.
- Practise writing queries and explaining each step out loud.
- Refresh the basics: summary statistics, distributions, comparing groups, correlation versus causation and sampling.
Case and take-home tasks
Ask how long to spend and what is being assessed. In your answer, state the question, your assumptions, what you checked, the finding and a recommendation. Say what you would do with more time or data. A short, clear write-up beats a long one.
Explain results to non-specialists
Startups and scaleups often want analysts who can talk to colleagues outside the data team. Practise a two-minute explanation of one analysis with no jargon.
Questions to ask them
- What decisions will this role inform?
- What data does the team have, and how is it kept up to date?
- Who will I work with most?
- What would the first project be?
Before the day
Follow the basics: read the job description, research the company and prepare examples. See our research guide.
Distinguish the role before practising
Read whether the vacancy is mainly about reporting, product decisions, experimentation, modelling or production machine learning. Prepare the tools and questions that match it. Do not assume a data analyst and a data scientist will have identical assessments.
Data-science hiring processes can include a coding task using SQL or Python, discussion of assumptions and trade-offs, and communicating results to a non-technical audience. This is a published employer example, not a guaranteed process at other companies.
Try these original practice prompts:
- A signup metric fell. How would you check the definition, tracking changes and affected groups before recommending action?
- A query joins customers and purchases. What is the unit of each row, and could the join count a customer more than once?
- Two campaigns have different conversion rates. What information would you need before comparing them?
- A dashboard supports a recommendation. What evidence is missing, and how would you explain that uncertainty?
Write a short answer with the question, checks, finding and next action. Explain what is observed and what is an assumption.
Next steps
Read our general startup interview guide, then browse data roles on the jobs page.
FAQ
What do data analyst interviews at startups cover?
Usually your past work, technical skills such as SQL and spreadsheets, a case or take-home exercise and how you communicate findings. Ask the company for the stages.
Do I need Python for a data analyst role?
It depends on the advert. Check which tools are named and prepare those.
How should I approach a take-home data task?
Ask the time expected, state your question and assumptions, check the data and present a clear finding and recommendation. Keep it concise.
How do I explain data work to non-technical interviewers?
Start with the question and the answer, then explain how you got there without jargon. Practise a two-minute version.
What if I have no professional data experience?
Use projects, coursework or volunteer work. Show the question, how you handled the data and what you concluded.