Lyft Product Manager Interview: Driver Cancels 

Don't leave your product management career to chance. Sign up for Exponent's PM interview course today: bit.ly/3IWTSSK

What I learned from this interview is that the root cause has to be identified by scratching the surface of different aspects involved, I've tried to list them down below - 1. Technical challenges (any major release, app versioning issue, bugs or issues etc.) 2. User behavior analysis (at what point the ride is getting canceled) 3. Sentimental analysis (any PR related to Lyft) 4. Demographics analysis (age of users, locations etc.) 5. Competitor analysis (any new feature launched by Uber, updated in marketing campaigns, policy changes) Feedback/suggestions will be appreciated above. BTW really helpful video. :)

Just a tip to everyone, while these mocks are great and I am a paid subscriber. Asking way too many questions will take away flow of the interview. I will suggest ask basic questions and then make a list of assumptions declare it. Understand the interviewer is here to judge how you structure the problem. And giving away 30 mins for one question may not work each time.

Haha, he grilled the interviewer before he started solving the problem. Loved this. Thanks Exponent, your mock interviews are really helpful.

I enjoy this video better than others. Yes he asked a lot of clarifying questions, in plain sight people will think he doesn’t have a good structure. All those clarifying question were high quality analysis, especially with the role play of data scientist and pm. I learned a lot from his questions, he narrowed down an ambiguous prompt to be a detailed answer

I would have asked about two additional segmentations on the business and rates side to make additional determinations. a. Does this issue happen significantly more at a specific time of day (rush hours for instance)? b. Did our compensation model or rates change for our drivers in the last month? If no, and our drivers also drive for competitors, do we know if the competitors have changed their compensation to drivers?

He was all over the places. Some concerns that he raised do not seem to relate to the problem due to lack of explanation. For example, 1. how is a shift from redshift to hive relevant? 2. How is holidays relevant? 3. How is glut of drivers causing driver cancellation? Shouldn’t it be the other way around? 4. I like that he went into funnel analysis. But he seemed to forget the question was about cancellation rate, not number of cancellation.

During the video, i paused in the middle and I started searching for the root cause myself, and since the change was immediate not gradually, I excluded that drivers are canceling the rides, I thought that this action would take time to build an audience, at the same time, 5% doesn't indicate that the problem could be a software bug, which got me back to the conclusion that it might be a campaign or a movement by the riders and I never thought that it would be so clear, but I like that I was close and probably asked the same questions, thanks for the great content anyway!

I would’ve also checked on Lyfts pricing strategy- did we change anything around how we pay drivers or how competitors pay drivers. I would also look at seasonality or time of day. For the follow up question, could there have been an option to look at company policy? Maybe we want to look at the revenue impact that’s causing and instead of focusing on incentivizing drivers who are using both Uber and Lyft, we want to focus our efforts on the drivers who are using Lyft only and modify our matching algorithms to prioritize the Lyft only drivers?

Another question I would ask is whether cancel increase happening in daytime or night - driver might not want to take long rides in the night, when they get to know the destination.

I'm at about <a href="#" class="seekto" data-time="780">13:00</a> min of the interview and I think I cracked it! Drivers are cancelling once when they see a rider without a mask on?! 😎🙂

I would start by breaking down in 1 - drivers behavior Data to check: - record of cancelation requests by driver (I assume you can check if the cancelation was trigged by driver's actions) - if that was not possible, I would simple check drivers that had a huge increase and call them. 2 - tech issues Data to check: - customer support inquiries by drivers - increase in the % of errors in the system log

Amazing interview. Learnt a lot. Thanks so much for making this available!!

Wonderful analysis guys ! Can you make a video about Product Manager vs Engineering and how engineers can switch to product management ?

Verbose. Should have laid down the possible hypotheses in the beginning: Driver side (relative incentive, experience, etc) Tech side Macro side. He should have broken down each section to find the segment as the first thing. But interview pressure is definitely a thing.

Are the declines being caused by Competitors offering better deals and services. Or are are drivers canceling? And if so why? Is it because they have bad internet connection which delays GPS and makes them frustrated and cancel on the customer?

@Exponent It was interesting to see a q&a format between data scientist and PM. Similar questions are asked in a DS interview specifically in case study rounds. Do you think this is the best way to answer them or just layout intuitive hypotheses and explain what data you might want to pull to test the hypotheses.

Few thoughts Solutions: 1. Incentive for zero cancellation. Gamify - streak and rewards for quick wins as well as month, quarter, year streak. It should have minimum rides criteria also for fare play Check driver calls and customer cancellation to catch foul play Check distance and time for customer cancellation to avoid foul play

Hey! You guys do a lovely job. Can you help us get started with what we need to study for interviews or what resources must we study to get a job?

Love the different versions suggestion lmao. Gotta make sure the app is up to date to prevent hacking or remote accessing.

If an interviewer asks- What is your favourite product and why? How would you improve a feature that you don't like in your favourite product. In this case shall we pick up a product in which we have worked in our past experience or any random product like whatsup, facebook, amazon app which is known to everyone but where we not worked in real work scenario or any blank product that I have prepared well before hand like alarm clock for a blind etc where neither we have worked on the product and not necessary that the interviewer has used it. what kind of examples shall we pick ?

I felt he was hitting a dead end with every lead. Very difficult for the interviewee. But I wonder why he did not write his thoughts down or what dead ends he hit so that he doesn't revisit them ? 🤔 Because this is a mock ? @Exponent is this video recorded live ?

Very useful. Can you made a clip to guide how to do a good take home assignment because during covid many companies interview onl and they request to finish take home assignment 1st.

Nice one Hadar!

Comming also from a consulting background, although it gets to the root cause the interviewee did not articulate its structure and framework upfront so it seems like he is guessing up to the next question. I find a lack of structure, it is not about asking the right question but also stating why you are asking that question...

This is a great initiative Exponent! Could you do more mock interviews on product marketing? That would be of great help!

I enjoyed this interview. I love how he was really digging into the analytic side because he really made sure he had all his bases covered before moving to the next section. I wish the challenge was a bit harder, after the market research / interview about Uber campaign thing was revealed, that was it😂 I wish he made him guess more, like ok its in this region and its at this time.. then create a hypothesis around it. Anyways thank so much for this

Excellent mock interview...

The question is how Uber is going to capture those moments when the driver comes from Lyft app, is it possible technically?

Great content as always @Exponent! One question though: I suppose it would help for the interviewee to write down some sort of decision tree as reference to explore options, right? I mean just to make sure there's a logical path in the discussion to reach the 'solution'.. Thanks.

But if uber has launched a marketing campaign/s, why would the lyft driver cancels the ride after accepting it? If the lyft drivers knows about the incentive benefits on Uber, he wouldn't be accepting the rides on lyft. How does this be factor for the increase in cancel rate?

Just a question How will Lyft have data about who all drivers are on other platform.

Great content tbh....Could you please make some mock interview videos on Program and Project management...could not find much quality content in youtube..

I got frustrated after a point. Stephen, can't we ask quick questions directly in a real interview rather than giving context for everything that we ask. For example : is it happening on ios or Android. Is is happening in a particular app version. Did we do a app release. Did is happen in a particular geography or demography. And let the interviewer ask me why do I wanna know this rather than just filling up the clarification questions with words?

When did Steph Curry quit GSW to work at Lyft!

structuring the approach at the top would definitely improve the chances

Holidays questions are frequently asked.

Could have asked for ‘time of day’ or ‘day of week’ slice too

Try this: "Let's call the drivers that are experiencing this change and ask them some questions"

Can some one just tell me the solution finally? 😅

Sooooo good

This guy is all over the place.(super unstructured)

11 minutes in and not a single theory. Not sure who is being mock interviewed...

wow first person to view :)

The interviewee was a bit painful to watch. He didn't seem as structured so it felt like he couldn't narrow his path. Thus, a lot of irrelevant questions.

One of the dumbest way to solve the problem statement and looks like the candidate is too casual not serious at all. Too naive questions at random stages with no sync. @Exponent - I like all your videos and learn a lot. Please do not drop the ball.

Very unstructured interview to be honest. Difficult to follow the thought process.

Never have I ever changed the root cause during these questions. Nope. Never >: >

Bro looks like Steph Curry

Not a good approach of problem solving. Interviewee talks a lot but didn't approach the pinpoint until 5 ~ 6 min. If this problem solving approach is happening in our company, he will be low rated.

Very unstructured approach and too much of redundant questions

this guy definitely likes to hear himself talk and comes off as though he’s trying to repeatedly flex his intelligence by asking as many questions as humanly possible in what’s supposed to be a 20 min interview. offputting if i was an actual interviewer.

uh.... uhh.... uhhh ... uhhh ... uhhhh. so many "uhs" .. its very distracting..

Too verbose. Would've been better if Hadar filtered out his thoughts and gave it out in consumable bites rather than talk like non-stop!! Just a thought.

A very interesting mock interview I have watched on Exponent so far, thanks for the great content!! I have a clarification question thou, it was mentioned towards the end that even if we know the answer it's better not to put it right away instead ask more questions - which I totally agree, but if I can find the right answer sooner, can I share that first and then say how about we also look into other aspects of it; is that approach right or not right?? For example, when I am first asked that Lyft rides are cancelled by 5%, what could be the potential cause, I would say it this way: Let's do it this way, imagine I am a Lyft driver, this morning I was on my phone, and for your reference I have both Uber app and Lyft app to increase my chances of having more rides. I first got a notification on Lyft that I have a ride to accept, I just accepted it...but right then I also got a notification from Uber that I have got a ride. Now, which one will I accept?? For that I have to see what's the most important value for me? It can be that I trust one company more than the other, or that I like the UI of one app over the other, or that I can make more money using one app over the other. I think for me as a Lyft or Uber driver, I want to make more money. So, it can be that if I cancel with Lyft and choose Uber, it might means that Uber is paying me more or in other words it provides more financial inventive. Putting ourselves into the choose of the driver in this case, and finding what's the most important thing for us can help us to arrive at the answer sooner. But wait, this is one aspect, there can also be all those other aspects like what @Shikhar mentioned below "behavioural, sentimental, demographic and competitor analysis"...now I am going to ask more questions on these aspects as well, and at the end select the one that makes more sense. So clarify my question again, we arrived at the right answer sooner, we shared it with the interviewer, but then said there could also be those other aspects, then we dive into all those other questions and talk about them...and at the end choose the answer that makes more sense...is that the right approach?? I know it's the longest comment, but any feedback or suggestions will be appreciated! Cheers 😊😊