sanity Data Analyst Interview: Questions, Experience & Prep (2026)

Sanity (sanity.io) is a content platform company best known for its headless CMS and structured content tools, used by product and engineering teams globally. As of July 2026, Sanity has 26 open Data Analyst positions, signalling active hiring across functions.

Candidates report the interview process typically spans three to four rounds: a recruiter screen, a take-home SQL or analytics task, a technical panel, and a final culture or leadership round. The focus is on product intuition, SQL fluency, and the ability to translate data into clear decisions.

Salary bands for Data Analyst roles in India, based on knok jobradar data (July 2026):

Across all openings tracked by knok, 319 Data Analyst jobs are live in India right now, with the highest concentration in Bangalore (41), Delhi (22), and Mumbai (19).

These are questions candidates commonly report from Sanity Data Analyst interviews, based on typical patterns for product-led SaaS companies:

1. Walk us through a dashboard or report you built from scratch. What business question were you trying to answer?
2. Sanity's product generates structured content data via APIs and real-time pipelines. How would you model and query event data from a CMS platform?
3. A content team tells you their page views dropped sharply this week. How do you investigate whether this is a tracking issue or a genuine drop?
4. How do you handle missing or inconsistent values when joining tables from different data sources?
5. Write a SQL query to find the top 5 content types by average session duration over a rolling 30-day window.
6. How would you define and measure 'content health' for a SaaS product where customers publish and manage content at scale?
7. Tell us about a time you disagreed with a stakeholder about what the data was showing. How did you handle it?
8. How do you prioritise analytics requests when multiple teams need your help at the same time?
9. Sanity serves international customers. How would you approach building a metric that works consistently across different time zones?
10. Describe your experience with a BI tool such as Looker, Tableau, or Metabase. How do you decide what to put in a dashboard versus what to leave out?
11. What does a good event data model look like for tracking a user onboarding funnel?
12. How do you communicate a complex or surprising finding to a non-technical product manager or executive?

Three STAR-format sample answers for common Sanity interview questions:

Q: Walk us through a dashboard you built from scratch.

*Situation:* Our growth team had no visibility into which content types were driving free-to-paid conversions in our B2B SaaS product.

*Task:* I was asked to build a self-serve dashboard so PMs could answer conversion questions without coming to the data team every time.

*Action:* I mapped the funnel from first content view to plan upgrade, identified three key events to track, wrote the underlying SQL in our data warehouse, and built the dashboard in Metabase with filters for date range, plan tier, and content type.

*Result:* The team found that users who engaged with 'how-to' content in their first week converted at a rate industry surveys commonly place in the top quartile for B2B SaaS. The PM used this to reprioritise the onboarding email sequence.

---

Q: Tell us about a time you disagreed with a stakeholder about what the data was saying.

*Situation:* A marketing lead was convinced that a new campaign had doubled signups, pointing to a spike in the dashboard the week it launched.

*Task:* I needed to validate the claim before it was presented to leadership as proof the campaign had worked.

*Action:* I pulled the raw event logs and found a tracking script had been duplicating pageview events for three days. I reproduced the discrepancy, documented it clearly, and scheduled a call with the marketing lead before the leadership meeting.

*Result:* The campaign had actually generated a real, if smaller, lift. The marketing lead appreciated the early heads-up. We also introduced a daily data quality check that caught similar issues going forward.

---

Q: How do you handle missing or inconsistent data when joining tables from different sources?

*Situation:* I was building a user-level report that joined CRM data with product event data. The user IDs did not match cleanly because the CRM used email as the primary key while the product used a UUID.

*Task:* I had to produce an accurate report for the customer success team within two days.

*Action:* I performed a fuzzy match on email fields to build a mapping table, flagged rows where the confidence was low, and documented which records were excluded and why. I also looped in the engineering team to fix the root cause in the pipeline.

*Result:* The report covered the large majority of accounts. The CS team could use it immediately, and the pipeline fix prevented the same issue from recurring.

STAR (Situation, Task, Action, Result) is the most reliable structure for behavioural questions at Sanity. Keep Situation and Task brief (two to three sentences), spend most of your time on Action (what you specifically did, not what the team did), and close with a concrete Result.

For SQL and technical questions, candidates report that Sanity interviewers value thought-out answers over rushed code. Talk through your logic first: state what the query needs to return, identify the tables and joins, then write the SQL. Edge cases such as NULLs, duplicates, and time zone handling are almost always discussed.

For ambiguous product questions (such as 'how would you measure content health'), use a three-step structure:
1. Clarify what the business cares about (retention, activation, or revenue impact).
2. Propose one to three measurable metrics tied to that goal.
3. Describe how you would build and validate the metric before sharing it widely.

For disagreement or conflict questions, lead with what you observed in the data, not with the disagreement itself. Sanity, as a product-led company, values people who protect data integrity without creating friction with partners.

For prioritisation questions, a simple framework candidates use is: urgency (is there a decision being made this week?), impact (how many people or how much revenue does this affect?), and effort (how long will it actually take?). Name the framework, then apply it to a real example from your own experience.

Based on the kind of work a Data Analyst does at a content platform like Sanity, interviewers are typically looking for a few specific things.

Strong SQL and data modelling instincts. Sanity's product is built around structured content and APIs, so analysts work with event tables, content schemas, and customer data that require clean, efficient SQL. Expect questions on window functions, CTEs, and handling sparse or nested data.

Product curiosity. Sanity is product-led, meaning analysts are expected to understand why a metric matters, not just how to calculate it. Candidates who ask 'what decision does this number support?' stand out over those who just answer the technical question.

Clear communication. The role involves working with engineers, PMs, and customer success teams. Interviewers often probe for how you simplify a complex finding for a non-technical audience.

Comfort with ambiguity. Many questions at Sanity are open-ended by design. They are checking whether you can scope a problem, make reasonable assumptions, and flag when you need more information, rather than expecting a perfect answer upfront.

Ownership and follow-through. Candidates who describe not just what they built but what happened after, including iterations, stakeholder feedback, and measurable impact, make a stronger impression than those who stop at delivery.

Week 1: SQL and data fundamentals

Practise SQL at the mid-to-senior level: window functions (RANK, LAG, LEAD), CTEs, aggregations with GROUP BY and HAVING, and self-joins. Focus on funnel analysis and cohort queries, which are common at product analytics companies. Review how to handle NULL values and duplicate rows cleanly.

Week 2: Product analytics and case prep

Read Sanity's public documentation to understand how their CMS and content pipeline works. Study common product analytics frameworks: funnel analysis, retention curves, and feature adoption metrics. Practise answering 'how would you measure X?' for content-specific scenarios such as editor engagement or content publishing velocity.

Week 3: Behavioural prep and mock answers

Write out STAR answers for five to six experiences: a dashboard you built, a disagreement you navigated, a data quality issue you found, a time you influenced a decision, and a project you prioritised under pressure. Practise saying each answer aloud in under two minutes.

Before the take-home task (if applicable)

Candidates report Sanity sometimes sends a dataset with open-ended questions. Structure your answer with a short summary of findings, the SQL or code used, and a recommendation. Clean, commented code with a brief narrative goes over better than a long technical report with no conclusion.

Day before the interview

Review Sanity's recent product updates, blog posts, and the job description one more time. Prepare two to three questions to ask the interviewer about the data stack, team structure, and how the analytics function currently supports product decisions.

Jumping into SQL without thinking aloud. Interviewers at product companies typically want to hear your reasoning before you write code. Rushing to type a query without stating your assumptions can make you seem less structured than you are.

Giving vague results in STAR answers. 'The stakeholder was happy' or 'it went well' does not land. Even without a precise number, say something specific: describe the change in behaviour, the decision that got made, or reference a publicly reported benchmark to give context.

Treating product questions as purely technical. 'How would you measure content health?' is not a SQL question. Candidates who immediately jump to metrics without first discussing what the business needs to decide tend to score lower on product thinking.

Not clarifying ambiguous questions. Sanity interviewers often ask open-ended questions on purpose. Asking one or two clarifying questions before answering shows good analytical instinct, not weakness.

Ignoring data quality in take-home tasks. A common trap is producing clean-looking output without mentioning the dirty rows, NULLs, or anomalies you noticed. Always flag data quality issues, even if you worked around them.

Over-preparing for one tool. Candidates who only know Tableau or only know Python can get tripped up when asked about trade-offs. Be ready to discuss your tool choices and why you made them.