we are an on-demand delivery platform you know like a doordash and we've noticed that our customer wait time has increased across the us and metropolitan areas what would you recommend we do to address this issue [Music] hey there we're so excited to film another exponent show with you today i'm here with angela to go through a metrics type mock interview um angela do you mind briefly introducing yourself for the audience yeah of course hey everyone my name is angela super happy to be here for the mock interview and exponent helped me greatly as i was
preparing for my interview and help me land an opportunity on product strategy operation at google i'm super excited to be here today and hopefully this mock interview will be helpful for those who are preparing for interviews in the community awesome thanks and we're happy to have you alrighty so i have a question for you today um we are an on-demand delivery platform you know like a doordash and we've noticed that our customer wait time has increased across the us and metropolitan areas um what would you recommend we do to address this issue yeah sounds good
so just to clarify our goal is to identify what has you know drive you know drove the increase of overall customer wait time and we wanted to find a solution to address that problem is that correct yeah that's right awesome cool so i think just come to share with you like overall how i wanted to approach on this problem so i think of course starting off by you know asking some clarification questions just help me to kind of understand you know what is impact or magnitude of this issue and this will help us to you
know identify if it's something serious that we need to address immediately or that's something we can just like monitor so in real life this will help us to like prioritize and thinking about allocation of resources and then of course i want to really clarify with you when we say like overall customer wait time what does that really mean um because that could mean like different things so wanted to really clarify that um and then after that i think will be important for us to kind of talk about you know some potential uh were probable causes
that could have led to this issue and then i will share hypothesis and then we can evaluate and identify what the root cause is and then very lastly and based on that findings and i can share it back with recommendations and does that sound like a good approach to you yeah that sounds perfect awesome so yeah so before kind of diving into um the problem just come for me to get a bit more contact in terms of like the overall impact and the time frame of this problem do you mind just kind of share with
me a little bit in terms of like is this a one-time job or that something has happened progressively um you know which part is of the of the market is impacted and how big is impact so it's like a two percent or like 20 percent this help us kind of evaluate how serious the problem is yeah um let's say that this is you know not really a spike it's been happening since say like may um and in areas like sf new york and say like it's a pretty noticeable change let's say you know maybe from
40 minute weights to 45 to about you know 15ish percent got it so it sounds like it's not something within normal fluctuation on a regular day and then sounds like this is a serious problem we wanted to tackle um are there any like other related metrics that we have noticed and have moved along with the increase of overall customer wait time um yeah let's say you know because of the increased wait time we're seeing a dip in csat scores like customer satisfaction and overall the number of orders placed as well got it okay so that's
pretty helpful and it sounds like this is something that is not an isolated market but more kind of like generally like abroad across many u.s metropolitan cities so this could potentially be a broader problem but i just wanted to kind of like dissect into like a couple different uh components so you kind of mentioned that the overall wait time is the moment when the customer plays their order to the moment they receive on the order so if i just kind of putting myself in the shoes of a customer and thinking about my own door dash
on experience i think that's kind of like have a couple different like parts so number one i think it's uh like the order confirmation time so when i place the order to the time the restaurant merchants confirm the order so that's kind of like the number one and the word of confirmation time and then after that i believe it's going to be the the preparation time from the restaurants so the time it takes them to complete the order have that ready for pick up and then after that is like the time between the word of
completion and the carrier of the driver actually arrived at the restaurant to pick up the order and then very lastly is the delivery time which is the time frame the order being picked up to the time of the order being delivered so i think i wanted to kind of evaluate on each of the component to see like where exactly this problem is happening i know we can identify if it's on the merchant side or the carrier side that it's contributing to um the problem here does it sounds like a good approach to you like evaluating
each one of the components yeah that makes sense sounds good and then kind of before going to the first part and and one thing i just also wanted to ask is like are we observing the change happening during a particular day of the time of the day like a lunch time or dinner time and that will also help us to kind of like sculpt down in terms of where to focus on let's say it's pretty consistent for both times got it okay so that's consistent and across and you know lunch and dinner time um so
maybe just to start off with the very first one are we seeing any you know in terms of order confirmation time that has increased meaning like restaurants are taking longer to accept uh customer orders um let's say there's no meaningful change there okay so let's disregard that one sounds like that is not really contributing to the issue here i'm moving to the next one what about any shifts in terms of like preparation time and you know are the restaurant taking longer to you know prepare the orders um let's say there's no significant increase in preparation
time either okay so the merchants or restaurants are just like taking a normal range of time to have the order ready for a pickup okay that's interesting okay so sounds like that's not also contributing to the issue here what about um the time between order completion to order pick up um you know maybe the orders are completed by just sitting there waiting for the driver to come in and pick it up if that has that time uh increased um since may yeah let's say that has increased got it okay so seems like there's something going
on on the carrier side if the order is ready just like waiting there for and the driver to pick it up so i noted that on my uh notes here i will kind of circle back on this one after evaluating the very last piece since that is also related to carrier and carrier time so the very last piece is the delivery time are we noticing any you know change or increase in terms of the delivery time i think there has to be something going on there because we evaluate the first uh couple and components and
are we seeing that also increased yes let's say we have okay so that's very helpful and then just like super quick recap here so we're basically saying that the increase in time is from the uh from order completion to being picked up and also from the actual uh cure delivery time and so i think those two and sounds like is where the problem is but i also wanted to kind of you know dig a bit deeper to see what's actually happening there so if the problem is more on the carrier side i can think about
a couple of things on top of my head that worth evaluating uh number one is you know insufficient carrier supply do we have the hypothesis here is like do we have enough carriers on the road if our supply does not really match the demand that could have led to the overall increased delivery time and or secondly it could also be like non-optimal navigation or like routing and for our drivers so the hypothesis there is like there might be some change in terms of how we dispatch our drivers or any change in the in-app map or
navigation experience that could have you know negative and experience or negative results on the carrier uh delivery time or like the third part i think we also wanted to consider this is more external factor that could also be like traffic so i think you mentioned that this has increased since like may i think like that's when like you know many cities start to reopen and people are starting to return to work so the traffic could have not been an issue in those metropolitan especially in downtown areas and so that's kind of like just like on
top of my head thinking out loud here a couple of different uh potential drivers and does this sound like you know good area to further investigate and or you know i should be focusing on something else yeah i think this is a good approach um you know let's say for one we have we know we have sufficient carrier supply for two there was no major ui or back-end change and for three um let's say we uh we have like you are correct here um and people have been spending more time on the road got it
okay so it really sounds like you know the traffic components and playing a big goal here and that has like negatively impacted uh in terms of driver delivery time meaning two parts like taking longer for the drivers to arrive at the restaurant to pick up the order and also taking longer for them to actually deliver um to the uh to the customer and so if that's not like where we wanted to focus and solve and so i think there are a couple things that we can consider doing in terms of like addressing on that problem
and then i'll kind of share both of the potential um solutions and then also share kind of like the pros and cons of each and then we can evaluate which one makes more sense and and then we can talk about next steps and also like potential risks associated with it and i think the first one is like if we're thinking about that you know uh the cities start to reopen and there's more traffic and i think we should consider here is the alternative delivery methods i know for like you know drawer dash or some other
um carry uh on demand food delivery there is like different options such as different by bikes scooters motorcycles or even by walking um so i think that's something that we can consider particularly for those like natural or downtown areas um the pros for that is one delivery time could be less impacted by traffic i know like from our own experience in those like highly trafficked areas sometimes it's faster just by walking or like riding bikes compared to like driving a vehicle and and and then secondly is i know some of this functionality has already been
built so if we just like recommend or increase the usage or promote these alternative delivery methods i it shouldn't require a ton of like engineering or products and development time since this is already something um existing but we might need some like lightweights um you know in terms of built on the features but i believe the kind of like the key functionality is there and that's on the pro side but i think we should also consider like the downside or the cons of these like alternative different methods i think number one is there could be
potentially like challenges in terms of like managing uh delivery quality i know that's something very important to doordash and i think just like food could be easily spilled and if you're like delivery via like a motorcycles or bikes scooters or like walking which could ended up like negatively impacting our customer experience if your order arrives it's built um so that's part one and then the second piece in terms of like the cons is that we might need to evaluate like how much um care supply we have who are willing to do the alternative deliberate methods
uh we might need to recruit more of those carriers and have to like on board them um so even if we you know identify there's something that works we also need to make sure that we can support that um so that's kind of like the second and parts of this uh potential solution which is like the alternative delivery methods and the second potential solution i have in mind is um promoting more of the pickup um so i know within after the customer uh when a customer is placed in their order they can choose they want
this to be delivered or they can pick up um the order themselves and of course this is like a a relatively easy one in terms that you know customer can get their orders faster which address our problem however i think there are more serious like consequences or implications and with this and then just to share a couple of thoughts um number one could be you know if there are fewer delivery orders uh means you know carriers on the platforms they're making fewer dollars and that could potentially lead to like carrier churn issues if they're you
know carriers do not have like enough uh orders being matched to and then the second part is that the customer actually do not get that it's kind of like your food is at your doorstep kind of experience and that could also be another thing to consider and i think just kind of like between these two like i would actually recommend the first one is like the alternative delivery methods um just giving the the advantage that i mentioned is that you know definitely help us address um the uh reduce the overall customer wait time of course
we'll talk about like we should validate that to make sure it actually works and then um true is that i think this one probably requires less um you know development time or building time um so sounds like a relatively lower cost solution to employment um but i think that's kind of high level um the two and then also the one that i recommend on the uh the alternative deal very methods do does this kind of the one you know sounds good to you before i kind of go into the the next steps and and also
like potential risks that we should consider yeah that makes sense i love to get a sense of the next steps yeah yeah cool cool so i think for us it's like basically um hypothesis is that if we do and um increase the usage of the alternative different methods we'll be able to address this problem so i think i wanted to conduct experiments and based on selected metro markets and you know where we can compare uh within a certain distance like orders and being dealer via like bikes we're walking versus waters being delivered uh via vehicle
and then we can compare if you know the alternative methods actually meaningfully reduce the delivery time um which and what they have also to meet the goal of the delivery quality so a couple like key metrics here i wanted to monitor um for our experiment is of course number one the delivery time which we identify is the root cause of the overall customer wait time and then of course the second piece is that we do not want to sacrifice the csat score which you want you know people to get their order earlier but also making
sure the quality of the order is there so um you know just like super high level those like would be the two i think that we should focus on um so this will kind of help us to evaluate you know if um you know this uh potential solution uh makes sense and how meaningful it is uh that is reducing our overall wait time and then of course like following that if that is a possible or validated and way to help us address the problem i think we definitely wanted to evaluate and do we have sufficient
supply of carriers who can do the delivery on these methods and you know because the potential risk is that if without knowing that um you know if we have insufficient supply um then we do actually need to build up the supply first meaning that we have to recruit those uh carriers and also you know onboarding them so that they can actually do those deliveries um so that's kind of like how i think about you know the my recommended solution and also the next steps cool well thank you so much for taking me through this question
um great i think i i was taking some notes uh during the call and i think you did a really great job sort of walking through the funnel right and like breaking it down at each point where things could go wrong and checking in with the interviewer to kind of see where they were at so that was really great and then i think another really great thing you did was when you got down to the solution aspects you explicitly called out the pros and cons and how each would weigh against each other right like some
cons might be bigger than others um one little thing about polish maybe uh when we got down to the metrics um you hit all the right spots but it might be good to sort of bucket them into broad categories like delivery time and cset score or maybe on like the consumer and customer side and maybe some of the stuff about supply is maybe more of like an ops metric but all in all i think you did a really great job um anything else you would like to add yeah thank you so much for the feedback
i think just like super quick self reflection here and one thing i kind of like forgot and is um you know just to make this more holistic uh or exhaustive evaluation i could have also asked you in terms of you know if there's any like internal issues like on the data or like platform issues it's a problem happening on a particular like ios or like android systems um i think i kind of like walk in like um just treated that as a business problem like without you know kind of like validating with you if there's
something going on like a system glitch or like new release or a book um i think that's number one just like to be like more holistic and like exhaustive that should also be like check the box and then the other piece i'm thinking about is like i think initially you mentioned this about like a 15 like increase compared to uh nay i think the other thing just like to helpful to check with the interviewer would be like how does that you know compare to maybe like prequel be time so maybe this was like a you
know within normal range and pre-kobe time this did just kind of like normal like covet and then went back to normal so that would also kind of help to evaluate you know this is something like complete out of range when i never seen it before or this is like kind of like within the normal range compared to pre-covet i think that bigger context will also help to kind of like thinking through the problem so that's just kind of you know quick two things i like super quick self-reflection but yeah thank you so much for your
feedback as well awesome well thank you so much again for being on the show with us and for those of you watching good luck on your upcoming interview [Music]
