A Window Into Our SQL InterviewsHow We SELECT Data Analysts At The New York TimesThe Times Open TeamBlockedUnblockFollowFollowingMar 14Illustration by Pete GamlenBy Alice Liang, Max Gendler and Robin LeeThe New York Times is a data-driven company with a large team of data scientists, analysts and engineers who provide insight on everything from site analytics to subscription modeling.
Our analysts work with teams in the newsroom, advertising and marketing to run A/B tests, build models for complex data questions and architect data flows.
While analysts may have different responsibilities depending on the team they’re embedded with, every part of the work requires a strong understanding of SQL.
With such a broad range of teams, we needed an effective way to evaluate our candidate pool.
So, we recently implemented a common technical SQL exercise for all open analyst positions.
This means we have a base from which to evaluate incoming candidates, and we know that everyone hired into analyst roles will have similar skills.
Here, we’re sharing an overview of the technical exercise we give candidates that has brought a more transparent, efficient and consistent hiring process to our team.
An Interactive AssessmentThe interview process for each data analyst position starts out similarly: a candidate first has a brief phone call with a member of our Talent and Inclusion team to discuss their experience and whether the open position fits what the candidate is looking for.
When a candidate passes the phone screen, they’re scheduled for a 30-minute technical interview via Google Hangouts with an SQL assessor, who is either an analyst or a manager within the Data & Insights group.
The SQL assessment is designed to understand both a candidate’s technical abilities and how they use that knowledge to answer business-focused questions.
This exercise is not to test how well a candidate has memorized SQL functions, so we encourage candidates to use documentation or other online sources if they get stuck.
We want to see how well they can ask questions about the data, work with documentation to get effective SQL statements and interpret the results of a query.
The exercise allows a candidate to experience the kind of work we do and to evaluate if they are interested in doing it full time.
We settled on 30 minutes for the interview because it’s a short commitment with great benefit for both sides.
We broke the interview into three parts: five minutes of set-up, 20 minutes for the SQL exercise and five minutes for the candidate to ask any other questions.
Right before the interview, the candidate receives an email with a link to join BigQuery, which is the Google database service our analysts use, as well as some relevant documentation.
The first five minutes are spent signing in to and setting up BigQuery.
For the next 20 minutes, we give the candidate access to a randomized and anonymized subset of actual data from the Times’s website, which allows them to run queries on the platform and see the results, just as any Times analyst can.
The exercise can be quite fun.
Candidates who have gone through the exercise have told us that it’s exciting to see and work with the data.
We prompt candidates to use the data provided to look specific patterns in how our readers interact with our main site.
Our prompts don’t ask direct SQL questions such as, “How would you use a CASE WHEN statement in SQL?” but rather more open-ended questions to ask candidates to make comparisons using the data and combine information from different datasets.
We want to see that the candidate can translate business requirements into things that can be done in SQL, mirroring what our analysts do on a day-to-day basis.
We follow a rubric when evaluating a candidate during the SQL assessment.
In particular, we look for the candidate’s process in writing an effective SQL query, as well as how they interpret the resulting dataset in a way that answers the business-focused prompt.
We also encourage candidates to ask insightful questions about the data and the prompt.
Our SQL assessors take notes along the way, and we can provide some helpful hints if needed.
Over the past couple of months, we’ve gathered information from the SQL assessments we’ve given, and we’re still iterating on the best questions we can ask candidates.
We’re continually making other improvements to the process.
Inclusive Hiring PracticesHiring a diverse workforce and being inclusive in our hiring practices is important to our team and to The Times.
With the SQL technical assessment, we have developed a few best practices such as meeting every few weeks to discuss how to make these assessments more consistent.
We also believe making the process transparent may help us hire better qualified and more diverse candidates.
We take extra care to continually make sure the assessments are fair and fun for everyone involved.
Our assessors will not see a candidate’s resume before the technical interview.
Since the SQL assessment’s objective is to evaluate candidates’ skills in SQL and data analysis, information such as a candidate’s school or company do not add signal to the technical evaluation.
A candidate’s background shouldn’t affect how we assess them; it doesn’t matter where they learned SQL or if they used it in a past job.
We’re most interested in how a candidate uses their SQL knowledge in a Times context.
Assessors are able to peer review each other’s evaluations along with the queries run by candidates.
From time to time, we sit in on each other’s interviews to make sure we’re evaluating the candidates in a uniform manner.
The SQL test is only the first part of the interview process, and candidates who do well on it can move forward to meet with the hiring manager and other team members.
Outside of the SQL test, an assessor is not part of the rest of the candidate’s hiring process.
This helps to ensure that the assessor can give an unbiased score to the candidate’s technical ability.
We know technical tests, and especially live ones, can be intimidating — we hope this helps to clarify what our technical interview is like, and what we’re looking for in data analyst candidates.
Please apply to an opening in Data & Insights if you are interested in JOINing our team.
Alice Liang is a Marketing Data Analyst for the Data & Insights team at The New York Times.
Follow her on LinkedIn.
Max Gendler is a Senior Data Analyst for the Data & Insights team at The New York Times.
Follow him on Twitter.
Robin Lee is a Data Analyst for the Data & Insights team at The New York Times.
Outside of work, he’s a data meetup organizer.
Follow him on Medium.
.. More details