"How Would You Improve Google Maps?" | Google PM Mock Interview

Key Takeaways

hey guys my name is bijan from product management exercises and i'm here today with my product manager friend gabe who's a who is in san francisco today we're going to do a product interview a product improvement question we're going to do a mock interview together um and hopefully this will be a good learning experience for you all um so the question that we're going to tackle today is how would you improve google maps okay um so i think to start off with uh i just want to clarify my understanding of uh the product to make sure on the same page um so my understanding is that there's prob there's a few features first is that it provides maps pretty much the entire world including things like street view the second thing is it allows people to navigate around um whether you're walking or bicycling or on the car including turn by turn and the third thing is it allows people to search for i'm going to call it sort of points of interest like restaurants or stores and and find kind of rich information about those points of interest does that sound about right yeah okay awesome um great so um it's a really big product um so for the purpose of the exercise i think i'm going to narrow it down so i think um the most interesting i think to improve is the points of interest section and i kind of that's even itself really big so i want to narrow it down further and say let's focus on improving how um google maps deals with people interacting with restaurants and searching for restaurants and i'm really interested in that because for a couple of reasons um i think that um there's a lot of upside in terms of the functionality in that area um and kind of related to that i think that uh google maps is losing out in a lot of potential uh traffic and engagement to its competitors in that specific domain so for example yelp foursquare even you know open table things like that um right now or maybe even things like door dash and whatnot um a lot of people are completely skipping out on google maps and going straight to those competitor um properties so i think there's a lot of potential upside um to work on there does that sound good yeah okay that's good awesome um but just take a step back um i want to specifically focus on yeah on on how people interact with with restaurants and dining out actually so i mentioned door dash and things like that but i don't want to worry about that one about i want to talk about how they die and actually go out to eat um so the next step is um let's think of a metric and how we want to improve the product so i always like to think about um it's kind of arm metrics the acquisition activation retention and monetization so when i think through those i think acquisition like pretty much everyone pretty much everyone has already used google maps at least once and sort of at least somehow interacted with a restaurant there um and so i'd say they you know been acquired and probably activated as well um but then sort of more the retention or you can also say just generally general engagement um that's where i think there's a lot of a lot of upside and monetization um i think that we don't really worry google makes us money from um you know ads primarily um and we'll leave it we'll leave it that and it's about making this a really sticky ecosystem so i really want to focus on uh the engagement side of things got it okay that makes sense yeah and if i had to operationalize this particularly with interacting with restaurants i probably want to operationalize it as the number of searches for restaurants on google maps in a sort of weekly or monthly period and i also want to look at specifically um how many searches per active user so for for an active user how many searches do they make per week or per month and i want to see that trending up as sort of a success metric okay um so i guess okay so the next step then i would want to be looking at uh which type of um user i'd want to to focus on um i like to think of um i think on on the top end so to speak we have the power users that are always using uh google maps um then on the bottom and we have people that almost never use google maps they're like very infrequent of users and then we have the kind of um so we have infrequent i call them infrequent occasional and um and power and i really want to focus on those occasional users people that are using it from time to time but how do we really increase uh the engagement of those users that sounds nice yeah sounds good all right um so the next thing i'd want to do is to think through some some pain points and problems for that user group um so do you mind if i just take a minute or so to think through um think through that yep that's fine go ahead okay i think to start off with i'd want to um think about the user journey um and then from there i can identify some pain points which which with some of those elements of the user journey so the user journey i think is decide you want to go out for dinner find an appropriate restaurant coordinate with others if you're dining out with someone else make a reservation or try to figure out when you should arrive to avoid the line or figure out like what the wait time is going to be um choose uh the best meal once you're there you eat and then maybe you um provide a review or some people are gonna sort of share what they've eaten on instagram or social media or something like that um i think yeah that's that's pretty much the user journey um and then i've actually written out a little table here um which i guess you can't see but um i think for each of those i can i can write down some pain points so deciding wanting to go out for dinner i don't really think there's uh any that's not really a pain point people just know that they want to go out for dinner i'm not gonna deal with that um finding an appropriate restaurant um pain point there is probably um being overwhelmed with choice and just can't choose a restaurant i know that's a problem for me often there's just so many things that look good uh coordinating with others the pain point there could be just difficult to coordinate with others i don't know you probably everyone's probably had that experience where no one can seem to gain consensus on what time is the best or everyone's coming from different locations and everyone has slightly different tastes and food and different preferences so that can be a pain point um making a reservation or getting a table at a walk-in restaurant um i think getting a reservation um is kind of that's open table and things like that but a pain point definitely is knowing when to show up um to easily get a seat um or walking and actually i believe google maps does kind of address that right now i think there is like estimated wait time um there um and then we have choosing the best item on the menu the pain point there is can't decide what food to order which was very common one eating not going to worry about that providing a review maybe you don't know what to write or it's just too too um too busy to do it and sharing what you've eaten on social media again maybe you don't know what to write or it's difficult to take a good photo or something like that um so to prioritize those um so i want to prioritize them based on just just impact to the user to decide to choose a good um uh pain point so i would say um everyone with choice can't choose a restaurant i said that has a very high impact to the user that's very very common difficult to coordinate with others choosing a restaurant that i'd say is moderate because it can be very painful but it's only painful for some of the users you know some of the time [Music] don't know when to show up um moderate because it's only an issue for some restaurants some of the time can't decide what food to order i'd say high impact that's very common issue i'm providing i don't know what to write what a review to write i'd say low it's not that painful if you don't know what to write it's like you just don't leave a review it's not that impactful to the user and same with um not sure what to to to share on social media i'd say that's also it's only an issue for a very small number of people that really care about that kind of stuff i think i mean it could be interesting if you create some kind of solution that made it really easy i'm sure they could but for the purpose of this i'm going to say it's a low impact to the user got it so that just leaves two high impact pain points which is being overwhelmed with choice when finding an appropriate restaurant and having difficulty choosing what food to order so i'm going to choose those two pain points to continue with and move on to creating solutions got it okay great okay um so okay now i'm just going to cut the solution so if you can give me [Music] a minute or so just to brainstorm some ideas as well sounds good okay um so i've got a few few solutions i'm gonna order them by according to that the pain point that i mentioned um so choosing a restaurant um so actually just to back up for a second i think that um one of google's competitive advantages and core competencies is in building algorithms and incorporating all these different types of data in order to rank things in in order of preference that's what they do with search um so i think that um when i was thinking through so many solutions that was sort of what came to mind and something we could we could do here so in choosing a restaurant you could show a selection of the top three or even the top one restaurant you should try next based on um other data you have so maybe google could infer um what they can they have the data on um what you've searched for previously and they have geolocation data so they can even infer where you've been and maybe you have some reviews you've left previously as well and then they also have data on trending restaurants and things like that so combining all that they could say here is the next restaurant you should try um or even the three and even maybe even show um you know a match rate kind of like netflix or something like that um the next solution is uh showing if your friend went to the restaurant and what they rated as kind of like a facebook type thing um you know i might see that bijan went to this restaurant and he rated it five stars so that might really pique my interest and make me want to go there i also was thinking uh showing the most popular restaurants in the area um but i can probably um i'll just i'll just go through first and we can uh extract those off later um so those are those are solutions for choosing a restaurant now for deciding what food to order um i think similarly you could suggest foods based on prior food that you've ordered um so you could if you go to a restaurant look it up it could say like these this is the dish that's most recommended to you or maybe the you know top three dishes the best appetizer the best main and the best dessert we'd recommend for you at this based on food you previously ordered and you know just general ratings and popularity and things like that i think it would the you know i think the interface might require you to to enter in what you what you like though so that could be a little bit challenging um which brings me to the next solution which is maybe people could fill out a questionnaire or some kind of wizard form on google maps where they get they answer like a list of 10 or whatever however many questions they want to and about their taste and their preferences and then they can get restaurant and dish recommendations based on that and you know the more questions you answer kind of like okay cupid or something like that the higher your matches you know so you could start with just answering 10 but maybe there's like hundreds if you really wanted to right so the more power user types might and that could be something that you invest in and that could hook you into the system because you've invested all this and you really want to know you know want to know what it's going to spit back out at you um and then you could even have you know email program push notifications when there's a new restaurant that has a high match for your preferences um similarly uh you could show the dishes uh that your friends have ordered or rated highly and what they're rated as and then again you could show the most popular dishes at the restaurant so i think that's seven solutions but i think i can immediately cross off a few it's like showing the most popular restaurants and the most popular dishes i think actually google kind of does that so just strike those off um and uh having people fill out a questionnaire actually i like that but it's not really a solution in and of itself i feel like it's more of implementation detail for the the top three restaurants and and the top foods i think it's kind of like how would you get that data well as part of that you could have this questionnaire to fill out but i wouldn't say it's necessarily a solution a solution of itself it's just a it's a um a way to um a way to enable the the top restaurant on top uh food recommendations um so that leaves me with uh one two three four um i believe yes lose me four so now i think i'll um i want to prioritize those based on the impact on engagement and the effort to build them um does that sound good yet yeah okay so the top three restaurants you should try next um based on uh prior selections and geolocation data and trending restaurants um and um so in terms of interface i'm envisioning sort of a card at the top of the restaurants tab maybe when you say maybe there i believe there is a actually i don't think there is a restaurant i think there is a local guide tab at the moment on google maps um but um yeah maybe if you search restaurants maybe uh or maybe they will create a tabs like that if they or if they search for restaurants maybe a card comes up and says you know uh here are the top three restaurants we believe you should try next um and a match rate similar to netflix for the data as i mentioned we have trending restaurants based on geolocation data and searches we also have geolocation data and searches that this individual has made and um we can use ratings made by the user and we can also have the user potentially fill out this additional kind of survey for to get to enrich their match rate the impact on engagement i say is really high because this is directly addressed as a real top problem that almost every person faces when choosing a restaurant which is the right restaurant that i should try tonight um the effort to build um i'd say it's it's hard to know i mean this is google here that kind of has a core expertise in these kinds of things so i'm going to say medium a lot of the data is there already um like i mentioned i'm excluding the kind of survey aspect of it but the rest of the data i'd say is already there but you need to make create a new algorithm and train that algorithm so i'd say it's medium um the next solution is uh showing if if your friend went to uh the restaurant and what they rated it as um i vision this would look something similar to facebook shows when a friend likes a post or a page um i'd say that impact engagement is is a medium um first of all it kind of relies on on friends to have rated restaurants so you're not going to have a whole lot of data initially so it's not going to be that impactful and also it doesn't necessarily tell you what you like personally it tells you what your friends like which is a good you know that's that's influential but i don't think it's as impactful um and so i was going to say the effort is i think it's it's definitely easier than than the other algorithm but uh um so i guess it's it's low but i still say that um if the other one's medium and the student needs to be low or if this one's meeting the other one needs to be higher so since i rated the first one medium and say this one's low um but uh it's going to take it could take a long time to get sufficient data to make this interesting so from that perspective it's kind of hard you've got to like it's maybe not technically hard but it's hard because you've now got to figure out like how to get everyone to start reading things actually and liking things to actually make it you know useful um so suggesting the top one to two dishes to order at a restaurant based on um based on prior food you've ordered and ratings by other people so right now google maps shows you the most popular dishes but we could easily take that uh one step further and highlight the one to two that are most recommended to you rather than just like these are the ones that are most popular for the average person um and we can also use that kind of questionnaire format that i had of tool that i had to enrich the data further i'd say that the impact again is high and the effort i'm going to say is high as well um because it kind of not it requires you to enter in like what you've eaten and that's much harder to passively identify than restaurants you visited like google can see based on geolocation data that i probably went to this restaurant around the corner but they don't know that i had the pasta or the fish right um it's gonna be harder to implement that um and then showing dishes your friends have ordered or rated highly and what they're rated as again i see this being kind of similar to facebook um again it's medium impact because it's not as useful it's not it's not as personalized as if it directly uh matched to your personal preferences um and i'm gonna put that as uh um low as well because it just requires you to kind of map what your friends have have liked so based on the prioritization um i would choose to focus first on showing people um well so so i want to focus on the ones that had the highest impact on engagement i'm not as worried about effort to build here just because we're talking about google have an army of engineers and um maybe that's incorrect assumption but i think that they have an army of exceptional engineers so i want to focus on things that are going to have the highest impact on engagement rather than if i was at a scrappy small startup i might focus on the things that are like faster and easier to implement um so i want to focus on showing the top three restaurants and uh um in fact i would actually i would focus on if i had to choose just one i would choose the the top three restaurants because the highest impact and i believe it's um it has high impact and i believe it will be more viable to build than the selecting the one to two top dishes got it um now okay so now i want to look at um what uh like success metrics i think i kind of mentioned them a little bit before but i'll just reiterate there i really want to look at the total number of restaurant searches per month and the median number of uh restaurant searches per active user per month um of course i'd probably want to look at um also i'm not sure if i envision there to be any direct interaction um with those but um if i was going to do some kind of a b test i would probably want to be looking at the people that that inter like experience that feature whether there's a difference in in those um those searches per month and median number of restaurant searches per user per month um and that would give me an idea about whether that feature is driving engagement with the restaurant feature of google maps um so to summarize i focus on improving engagement with the restaurant feature of google maps i identified two top pain points being overwhelmed by choice with choosing a restaurant and deciding on what food to order um i identified uh six solutions and narrowed it down to i think i actually identified seven solutions and narrowed it down to one and that's building out an algorithm to match restaurants to users preferences and then displaying the top three to try and i would measure success by looking at the total number of restaurant searches per month and the median number of restaurants such as per active user per month for the users who um uh are uh given that uh are shown that feature as compared to the users who are given sort of uh the um the base case the control condition which is not having that feature got it and that's it yeah all right well thank you so much um so um this was really good and thank you for the answer i'm just gonna go over the um the answer that you provided quickly and then um give some feedback i think it's valuable for the audience here oh yeah i think the structure of the answer was actually really good uh you started out with um clarification and like really narrowing down the scope of the pro the question as soon as possible um you talked about one particular goal that you had in mind which was driving the engagement um you picked a um particular user group which was very important and that basically allowed you to really focus on them later um and then you went we talked about the pain points um you prioritize the pain points um and then you talked about um the solutions that address those pain points and then prioritize them and you came up with metrics um so structure wise i think it was great i think the user grouping was really good um i think the description of the product was good i think the pain points were very meaningful um i think um you came up with good lists of solutions i'll have some feedback around that in a minute but overall i think a couple areas that i think you could have potentially done a little bit better one would have been that i think it's good to have a goal um at the beginning which is like something like engagement and it's okay to not tie yourself with a very very specific metric until you actually have the solutions in mind because if you don't know what the solution is it might be a little bit hard for you to know how to measure the effectiveness of that solution or to effect to measure the success of that particular solution so um maybe it would have been okay to just say hey you know i want to drive engagement and after you've highlighted what solution you want to build then you talk to the actual metrics that are going to help you measure the success of that feature the second feedback that i'll have is related to the pain points associated with the user group that you pick um i wonder if the pain points that you highlighted are more or less the same across these different user groups and if they are then maybe we need it to be a little bit more specific with our pain points um so that they're more specific to the user group or maybe we we can just like say as a disclaimer that you know these pain points that we came up with um you know i can actually see them apply to other user groups as well just so that the interviewer is aware that um you came up with a bunch of pain points and you are aware that they're not necessarily very specific to that particular user group i think you did a pretty good job of like narrowing down the pain points to like two particular ones um i think your solutions were really good um we came up with a large list of them um i would have probably spent less time on evaluation of the solutions i think you can kind of stay at a high level and say um you know the implementation effort for you know showing where friends are i think it's not too hard we can build it or i think it's hard and here's why and move on to the next one um i think same thing with all the other ones i think you could have kind of um spend less time on the evaluation um i i personally really liked um your idea of um encouraging the user to actually go through a bunch of questions and then um basically help google understand them better and then use um then have recommendation made based on that or also like showing friends recommending um showing what places your friends went to i mean i think the trade-off to that is that there could be some privacy concerns that you'll have to figure out how to address um but i think it's interesting um to know for example if i have a foodie friend who's gone to this you know italian restaurant i'd love to probably go there because i i have a lot of respect for by for my friends basically restaurant choices um but i think that's more about um you know how you explain your answers and it's okay if you and i disagree on that um i think it's fine um i think at the end when you came up with um a bunch of metrics um to kind of say how you're going to measure it it'd be good to also highlight metrics that help you measure the usage or the retention of users who are using this product so for example what percentage of users who tried um google recommendation tried it again right and that could be one way to kind of think about it um i think it would be interesting to do that but overall i think you you followed a pretty good structure on i think i would also highlight about the potential trade-offs of you know this solution of like top three recommendations um what would be the reason that someone doesn't use it or um if that happens like you know what does that indicate does that mean that people didn't like the recommendations or um you know they they like recommendations but every time um their taste is different um so like for example today maybe i feel like sushi and if the recommendation was based on my you know last few months of um restaurant experiences maybe it wouldn't be as relevant because now i'm thinking about something else um i think it's important to kind of highlight that but thank you so much i think i hope everyone here enjoyed listening to this mock interview again my name is bijan from product management exercises um feel free to check out the website for interview questions um product manager interview guides and also scheduling mock interviews with other pm's website is productmanagementexercises.com