Data Science Product interview questions are known to be difficult questions in data science interviews, since these questions do not have a fixed solution. These product interview questions can become easier by practicing, and understanding the framework! Data Science product interview questions can be split into 3 categories: Analyzing a metric related problem Metric related problems usually involve a certain feature/product that has decreased in a certain metric. When providing your solution, you should explain how you derived your solution and explain the root cause of the problem. Examples of these types of questions: Feature A has been decreased in use by X% Uber asks “Uber Black rides have dropped 10%. How would you investigate this reduction?” The company’s product has decreased in a certain metric by X%, but increased in another metric by Y%, i.e. Google asks “Google noted that there was an increase by 8% in users who mentioned their privacy was respected. On the other hand, engagement in Google Play Services has decreased by 14%. How would you determine the source of this irregularity?” Companies look for: Definition and explanation of relevant metrics and how these metrics are linked to the solution and reason behind the issue Understand what could affect the success of certain features/products Identify potential metric-tradeoffs and provide reasoning on which metric should be favored 2. Measuring impact These questions are asked to understand how the interviewee can determine the success/failure of a product. Examples of these types of questions: “Quora recently implemented a new search toolbar, what metrics would you measure to determine the impact of the search toolbar?” Quora asked YouTube asks “How would you measure the success of YouTube shorts?” Companies look for: Definition and explanation of how each metrics is relevant to the solution of the question Understand what internal/external factors can have an impact on the success of a product Design an experiment to measure how impactful the feature/product is based on the defined metrics 3. Designing a product Designing product interview questions has no correct solution, but tests how well you provide evidence behind each step of your solution. These questions generally involve the company’s product or another company’s product that would help the company. Examples of these types of questions: “How would you design a product that uses a heatmap to help taxi drivers determine where to go to maximize their probability to receive a rider. How would you define which area will have high demand and which drivers would you send to that area?” Lyft asked , “What would you do to summarize a Twitter feed?” Microsoft asked , “If you had to propose a new Yelp feature, what would it be?” Yelp asked Companies look for: How well you understand the company’s products/market How well you understand the process of developing a product (Target audience in mind) How well you can define and explain the relevance of metrics to the solution How well you understand how to measure the impact on a product The solution to these types of product interview questions should be answered in different ways depending on the questions’ category, product/feature in question, and interviewer’s comments. An in-depth explanation on solving these types of questions is discussed in the “ ” article by StrataScratch. The Ultimate Guide to Product Data Science Interview Questions Each question category has various examples, common pointers, and YouTube interview examples. The YouTube interview examples are important, since this explains how important the interviewer's comments are during the interview! While on paper two questions may seem similar, the assumptions and background to the questions will differ! The interviewer will provide more context to the question, so you can understand how to structure your solution. The YouTube interview videos, included under each section, are a simulated environment of product data science interviews questions between an interviewer and interviewee. Do check out these interviews to get more context on how to go about these questions, especially if you have not done an interview before! Here is a list of practice questions classified into their respective category. Analyzing a Metric Related Problem Reddit asks “Traffic went down 5%. How would you determine the reason for this?” YouTube asks “The Daily Active Users in Indonesia increased by 13% compared to yesterday. How would you determine the cause?” Facebook asks “The Facebook 3rd Party Login declining 2% week-on-week — how would you locate the source of the problem?” Snapchat asks “The number of users has dropped by 20%. How would you investigate this issue?” “During a monthly report of uploaded user content, there was a spike in uploaded pictures in October. How would you investigate why this occurred? Yammer asked , “If 70% of Facebook users on iOS use Instagram, but only 35% of Facebook users on Android use Instagram, how would you investigate the discrepancy?” Facebook asked “Explain how network effects might influence your choice of how to assign experimental/control units and measure your main outcome metrics” Uber asked “Uber Black rides have dropped 10%. How would you investigate this reduction?” Uber asks “Google noted that there was an increase by 8% in users who mentioned their privacy was respected. On the other hand, engagement in Google Play Services has decreased by 14%. How would you determine the source of this irregularity?” Google asks “There is an increase of users buying Google One 100GB storage plan by 15%. How would you find the reasoning behind this?” Google asks ”When the Covid-19 pandemic initially started, Microsoft predicted that there would be an increase in Skype daily active users. Later it was found that the daily active users actually decreased by 4%. How would you determine the reason behind this? Microsoft asks “Daily active users have fallen by 5%. What steps would you take to determine why this issue occurred?” Twitch asked General Practice Question to Keep in Mind for any Company Pick one of the company’s products. You are now responsible for building a dashboard that will show all relevant metrics to the team daily. What do you track? This question will allow you to understand what are the important metrics to a company’s product. You should further research what factors could affect the metrics measured. Common Pointers to Keep in Mind for Metric Related Questions Metric questions are often related to change in the number of users. Snapchat (Question #4 above) asks if the user change means the total number of users or the number of unique users? When a metric varies, the change may only occur in certain geographical locations. Using the Snapchat example, the decrease in users may be due to a certain location. Ask the interviewer if the number of users has decreased in any certain geographical location and if it has you need to go about the solution in a different route. When a specific location/date is mentioned, you need to mention how this location/date is related to the product. For example YouTube’s example (#2) A reason for this could be a significant event that occurred the previous day which is important to Indonesians could have been uploaded to YouTube. This would cause a sudden increase in unique YouTube users. An example for the date mentioned could be Yammer’s example (#5). There could be a business trend related to photos or significant events in October. YouTube Interviews Lyft/Uber Metric Interview Question ​ Facebook Product Manager Metrics Interview (Facebook Marketplace) ​ Netflix Product Manager Interview (Inactive Users) Flipkart Product Manager Mock Interview (Root Cause Analysis)​ ​ Google Product Manager Execution Interview (YouTube Watch Time Root Cause Analysis) Measuring Impact Facebook asks “What metrics would you measure to determine the success of Facebook Newsfeed?” Facebook asks “What metrics would you measure to determine the success of Facebook’s ‘Save Item’ feature?” Apple asks “How would you measure the success of Apple's WWDC event?” Airbnb asks “How would you determine how successful Airbnb’s experience feature is?” Snapchat asks “How would you measure the success of Snap Map?” Yelp asks “How would you investigate how successful Yelp reviews are?” “Quora recently implemented a new search toolbar, what metrics would you measure to determine the impact of the search toolbar?” Quora asked “What metrics would you use to measure the success of an Instagram ad?” Instagram asks “How would you measure the success of YouTube shorts?” YouTube asks “How would you measure the success of recommended news based on a searched topic” Google asks General Practice Question to Keep in Mind for any Company Since measuring impact questions are similar, pick any one of the company’s products and understand how data scientists at the company will measure the success of a product/feature. Find metrics or features that are unique to the company’s product and try using this in your answer to show you have a better understanding of the company and its products. Common Pointers to Keep in Mind for Measuring Impact Related Questions Most of the time when a company mentions to measure the impact of a feature, people usually mention A/B testing and think that is good enough. Most companies do not accept this answer and would mention due to certain factors we can not A/B test the feature. A key pointer that people forget for measuring impact questions is not the metrics measuring success, but the metrics that do not worsen in the release of a new product. Using the Instagram question (#8) as an example, while you can mention clickthrough rate as a measurement of success, you should mention at least one metric that should not decrease, such as organic growth rate. Don’t forget when mentioning metrics, explain how these metrics are directly related to the impact of a product. YouTube Interviews ​ LinkedIn Metric Interview Question (Decode and Conquer) Facebook Product Manager Interview (Picking Metrics)​ Designing a Product Twitter asks “How would you determine the negative value of an abusive/discriminatory tweet?” , “What would you do to summarize a Twitter feed?” Microsoft asked , “How would you quantify the influence of a Twitter user?” Twitter asked “How would you design a product that uses a heatmap to help taxi drivers determine where to go to maximize their probability to receive a rider. How would you define which area will have high demand and which drivers would you send to that area?” Lyft asked “eBay has to distinguish cameras from the other objects such as batteries, tripods, and cables. How would you design a software that could classify these objects? Each product has data relevant to the product, which includes ads title, description of the product, price, images, etc.” eBay asked “How would you improve Google Translate?” Google asks Microsoft asks “How would you use VR to improve education?” “What metrics would you track after a user follows a new page and why?” Linkedin asks “What metrics would you use to determine how “close” two friends on Facebook are?” Facebook asks Amazon asks “What metrics would you track after a user ads a certain product to their cart and why? Facebook asks “What metrics would you track to determine a sense of community within Facebook Groups?” General Practice Question to Keep in Mind for any Company Pick one of the company’s products/features. You have to describe how you would improve this product/feature. , “Pick your favorite product or app and describe how you would improve it or design it.” Uber asked , “If you had to propose a new Yelp feature, what would it be?” Yelp asked Common Pointers to Keep in Mind for Designing Product Related Questions Product design interview questions are extremely diverse and hard to predict. Common concepts are based on how you will design a product with some quantification. For example with the Twitter example (Question #3), you will need to understand how Twitter followers can interact with a certain user. For example Engagement from followers can be quantified through likes, comments, and retweets. You can take this further by conducting sentiment analysis on the retweets. If a lot of retweets are positive, then it can be measured that the user has a higher influence on other users. These product interview questions require you to dissect the specifics of the question and see how well you understand the product. Another key factor is how will this designed product benefit the company. While some products, such as the Lyft example (Question #4) are straightforward on how it would benefit the company, the Microsoft example (Question #2) is not as direct as to why Microsoft wants to summarize a Twitter feed. Understand why the company would want to understand the development process of the company. Possibly Microsoft wants to see what people feel about a newly released feature/product. Think about how your solution to the question would help the company to develop its product. YouTube Interviews ​ Facebook Product Manager Mock Interview (FB Movies) ​ Google Product Manager Mock Interview (Fin-Tech Product) Tips for Preparing the Interview This provides insight into answering metrics and success measuring questions. While this video does contain other explanations to questions such as opportunity sizing, this is not required for Data Science product sense interviews, so you could skip those parts! YouTube video A lot of product designing data science interview questions overlap with product manager questions. There are two things to keep in mind when using product manager questions as practice. Product manager questions often go much more in-depth than required by product design data science interview responses. For example, PMs may require you to explain a go-to-market strategy, but this is not required for . Product manager questions may also involve reporting decisions to executives. This would be the summary of your answer. Data Science interview questions Whenever you are researching a new product or going through a product, try to create your product questions! Understand why a company would choose/not choose to implement a certain feature. Companies have probably already thought about your reasoning on why they did not implement a certain feature/product. Question why they would have done that! Also published on . StrataScratch

AirBnB

Amazon

Apple

eBay

Facebook

Google

Instagram

Lyft

Microsoft

Netflix

Target

Twitter

Uber

YouTube

StrataScratch is a community-driven platform for data scientists.

2022 - HackerNoon Contributor of the Year - Data Analysis

2022 - HackerNoon Contributor of the Year - Product Management

Nominated for 2022 - HackerNoon Contributor of the Year - Data Analysis

Nominated for 2022 - HackerNoon Contributor of the Year - Product Management

Too Long; Didn't Read

ML Practitioners - Ready to Level Up your Skills?

Top 40+ Data Science Product Interview Questions

Top 40+ Data Science Product Interview Questions

Too Long; Didn't Read

Companies Mentioned

@nathanrosidi

RELATED STORIES