This document provides an overview of execution interviews for product manager roles. It discusses three common types of execution interview questions: tradeoffs between two options, setting goals for a product, and performing root cause analysis. Frameworks are presented for each type to help structure responses. The document concludes with an example live mock interview.
8. What are execution interviews
This interview tests your ability to prioritize, identify the correct tradeoffs
and set your team up for success. As a PM you need to create logical
frameworks to make decisions against.
9. Things to keep in mind
1. Be flexible: Most interview conversations don't go as planned. Don't force a
framework.
2. Practice! Practice! Practice!: Practice mock interviews with these
frameworks, and mould them into what works best for you. Your ideal is not
the same as someone else's.
3. Collect your thoughts: As a rule of thumb take a 20 second pause right after
the prompt is given to you. It will help you collect your thoughts and decide
on which framework to use and how.
10. Learn how to design good metrics
Defining Product Success: Metrics and Goals by the Sequoia Data Science
team
12. PMs have to make the best possible
decision with the information
available and pave the path for
getting access to that information.
As a PM you have to define the key
success metrics and true north star
for the team to prioritize against
and push for.
For a lot of PMs understanding
metric fluctuations is a weekly
occurrence (myself included). We
have to be on top it, come up with
hypothesis and validate.
1. Trade off
between A and B
2. Set goal for a
product
3. Root cause
analysis
Example prompt: How many recipe
cards should be shown at the top of
Google search results?
Example prompt: Define goals for
Google search
Example prompt: Uber ride
bookings are down by 5%. How
would you go about identifying the
root cause?
14. 1. Trade off between A and B
1. Clarify: understand what is the tradeoff, the status quo experience, which regions are you planning to make the
change in, what population will be exposed to the change and why.
2. Product and company mission: state your understanding of the mission of the product and how that furthers the
company wide mission.
3. Goal: state the goal that the tradeoff is optimizing for. For e.g. For FB Newsfeed we should prioritize content that is
relevant for the user as measured by user engagement on the feed and with the posts.
4. Key users & stakeholders: who are the internal stakeholders (such as cross functional teams) and users (such as
advertisers and consumers) that will be impacted by this decision.
5. Pros and cons: assume you go all in with both the paths (one at a time), what will be the pros and cons of each.
6. Experiment design: layout your strong hypothesis and the population (such as users with more than the mean
number of friends) you plan to test with.
7. Decision framework: create a framework with the top line and counter metrics. Think of metrics in terms of buckets
(e.g. advertisers, users, performance). Learn more about defining good metrics here.
8. Tradeoff table: prioritize the key metrics and state your hypothesis of how those metrics would move for the
population that you picked, and what would be your expected decision based on that
15. Common Gotchas
1. Not identifying the common long term goal between the different options
2. Not thinking about the product wide impact
3. Not thinking about the correct population to experiment with
16. 2. Set goal for a product
1. Clarify: understand what the product is and what product cycle it is in.
2. Product and company mission: state your understanding of the mission of the product and how that furthers the
company wide mission.
3. Goal: pick a goal based on what cycle the product is in to focus the discussion
a. Growth: Gaining new users
b. Engagement/Transactions: Increasing usage of the app
c. Retention: Ensuring that existing users come back
d. Monetization: Converting app usage to money
4. Key users & stakeholders: who are the internal (such as cross functional teams) and users (such as advertisers and
consumers) that will be impacted by this decision.
5. Key actions: that you think should be encouraged and also the ones that need to be discouraged
6. Metrics: top line and counter metrics to measure the key actions that you prioritized in the previous step. Think of
metrics in terms of buckets (e.g. advertisers, users, performance). Learn more about defining good metrics here.
7. Trade offs: what are the ways in which the metrics that you have prioritized can be gamed. Most people struggle with
this. Before prioritizing the metrics in the previous step think about the tradeoffs you are making.
17. Common Gotchas
1. Picking vanity metrics as the topline
2. Picking ratios or complicated metrics as toplines
3. Incorrectly defining the time period of the metric (e.g. daily vs monthly)
18. 3. Root cause analysis
1. Clarify: understand what the product is and what the exact metric is. For e.g. "active riders went down by 5%", clarify
the definition of active and the time period the metric is computed over.
2. Segment: break down the problem to see if it affects a particular region, platform, device, user type (such as new or
returning, users in a loyalty program etc) or category (such as only Uber XL rides).
3. Contextual: gather context around the metric fluctuation. Could it be seasonality, weekend-weekday effect, dig into
the funnel, check the associated metrics (such as for rider cancellations see pick up times)
4. Internal hypothesis: investigate data quality (e.g. missing/duplicate logs), product changes (direct/in-direct), degrading
product quality (e.g. content on medium), changes in marketing and awareness strategy (e.g. cancelling promotion,
decreasing budget in certain channels, in-product awareness pushes).
5. External hypothesis: investigate competitor behavior (new product launches, new competitors), change in user
behavior, market situation and regulatory changes.
19. Common Gotchas
1. Not clarifying the exact metric definition and time frame
2. Asking for information and stating the hypothesis without the why
3. Asking for irrelevant templated information
20. Live mock with Apoorv Narang, Lyft PM
Prompt: You are the PM for Google recipe
vertical. You are trying to determine if more
recipe cards should be shown at the top of the
search results.
21. You can find more interview tips and
frameworks on rohankatyal.com