Delivery in the Age of the Shared Economy

when I was asked if I wanted to give a presentation in the series this year of course I said yes and I had to give a topic six or seven months ago and I decided that I'm surely would be able to say something about delivery and the sharing economy which is a topic that has interested me for a while when I actually started preparing the talk I decided to do a bit of a combination of my perspective of the developments I see in this area and focus also a little bit but not very much on some of the work that has been done here at Georgia Tech in this context so it's a mixture it's a bit of a perspective view introduction maybe to some of the questions that I think are relevant and important and a tiny bit also reflecting on work that we've done here with some of the companies that are actively involved in these questions so what I want to do is first talk a little bit about last mile delivery I mentioned delivery in the title but I'm really thinking about last mile delivery here then I want to introduce the concept of sharing and geek economy and then I want to focus a bit more on meal delivery specifically where both of these are coming together and it's also a very popular and interesting topic and that's an area that we at Georgia Tech have worked on extensively I have to say but as I said I want to start by just looking a bit at last mile delivery which was also the starting point for us I think here at take in terms of what we wanted to work on or have been working on surprisingly Oh things have changed the bit around I'm not exactly sure why in the presentation I had that one was in the upper left hand corner and it clearly shows that we're talking about Amazon here although most of you may have recognized this a little bit but so why is slight on Amazon I think Amazon certainly has been partly responsible for sort of the attention that last mile delivery has been getting and so I wanted to point out a few numbers so the number that's now in the middle about one hundred and fifty million mobile users accessed the app in September so there are a few things I wanted to point out of course that it's a large number but also that this was purely only at access right not on your home computer or your laptop it was through a mobile phone right so and we'll get back to that later the other thing that I think is very important to realize is the number of products that is listed in the upper right hand corner right I remember very well because I'm among the older ones here in the room that Amazon started selling books right now their portfolio is a hundred and nineteen or about a hundred and twenty million different products that you can buy online right so that's a phenomenal change also at your lower left hand corner Amazon Prime day which is their big marketing employee to get you to buy things and you get discounts they sold a hundred million products right so one of the things that interest academics or people like me that amazon has really changed the scale of the business right it's also listed here we have 4,000 items sold per minute right that leads to huge changes in in how you can think about optimization technology to support operations like this which makes it very interesting to all of us there are a few more statistics one of them is Prime Amazon Prime I expect that at least some of us in the room are Amazon Prime members I'm one what I wanted to point out mostly is this light or the the information on the right if you look at why what do Prime members liked about the service right and then it captures two important pieces of information about last mile delivery right they like that it's free and they like that it's gonna be at their home fairly fast within two days right and this seems to characterize what's happening at least on the customer side in last mile delivery people don't want to pay too much for it preferably free and they want to have things as fast as possible right and and this is where a logistics person starts to get kind of itchy because that makes things challenge that that seems to work right is shown here with a fairly recent clip from something that appeared in The Guardian a newspaper that I tend to read I think this was last week or or at least not long ago profits have gone up tremendously as a result at least that's what M is on things by offering next day service instead of two days service right and they plan to roll this out hopefully to all of us right so again MSO investing heavily in making delivery faster and faster and faster right now anybody in the logistics fields that knows that that also means your chances for consolidation will go down if you have less time of course they make it up by volume right so there's some interesting trade offs happening here so let's look at the bit at growth in last mile delivery right what's happening here and clearly the big the big reason is ecommerce right so business to consumer we can purchase almost anything we want now online and get it delivered to our hope right so that the bullet underneath about product availability is very important here as well right the growth is sort of two parts it's more products you can buy online and people tend to buy more online growth of the middle class that's maybe not necessarily the case in the US but certainly in other parts of the world China specifically more people can afford to buy things and they tend to buy it online the other part I mentioned that already the smartphones right we saw what was it a hundred fifty million people sort of actually purchased something online through their mobile phone right this is the technology has played the big role in this we all know I saw some statistics that we look at least twenty five times a day if it wasn't per hour on our phone right and then the other part and we'll get to that later is of course it's not only Amazon that is sort of heavily involved in the online business we've also have really different business models and in part that was motivated by the sharing economy which will we'll get to do a bit later so there's all this growth I want to highlight a few things about e-commerce per se and I'm sure most people have seen statistics like this what to me is the most interesting is the graph on the right and especially the blue line but the blue line indicates the fraction of retail sales that is gonna be purchased online right now this is for China and China is much further ahead of us in that curve but they expect that more than half more than 60% is going to be purchased online right and it's an interesting statistic the growth in China again just a figure about packages the growth is enormous just look at the red bars they keep going up almost exponentially right so this is big business of course that comes at a price right and I do think we especially as a community logistics community and and everybody is aware of this but need to really think about it and also think about what we can and should be doing right there's also negative consequences of this enormous growth right the most obvious one is traffic in congestion right especially in urban areas second lane parking right there's no space for the delivery vehicles to park so they still have to make their deliveries they cannot just drive around waiting for a parking spot and we get situations like we see in the picture obviously there is a big question about emissions right climate change thinks we need to be mindful of this if more and more is going to be delivered at home we need to start thinking about how to mitigate the effect and and I'll show you some ideas of how that's gonna work I want to mention two trains the one I mentioned already is that the trend is that everything needs to be delivered faster faster and faster right the statistics up there are for the US and it mentions instant delivery and instant delivery is especially relevant when we go to meal delivery the idea of instant delivery is that once you place an order the fulfillment essentially starts right away it's instant it doesn't mean that you have it instantly but at least the process starts so anything where let's say your service guarantee is within an hour within two hours after placing the order is now considered under the category instant delivery if you were not familiar with that and I mentioned already that Amazon wants to get things to us faster and faster the other trend that is important also when it comes to mitigation of the negative externalities of last mile delivery is there's new technology and here there's really a mixture of why people are thinking about these technologies one from the company's perspective is cost and service and part of it is also as I said mitigation especially electric vehicles is a combination right it's better for the environment and and that's an important part of why a lot of people are thinking about it self-driving vehicles the situation is should be different there the hope is that it's really cost reduction you don't need to necessarily and we'll see a few examples have delivery people on board or drivers on board so that saves you cost I'll give you a few examples in case you're not that familiar with droids here are some pictures right this is what we're thinking about they can go on sidewalks make small deliveries Amazon is testing them it's one of the new technologies that's out there I haven't seen them in my street yet but they may be coming soon self-driving delivery vans I find this interesting this particular picture it's not necessarily completely clear how things are going to work right is there still a person that is gonna take the packages out of the van and deliver them to me here we see that their tests in terms of grocery deliveries I find this one the most intriguing and the most likely to be effective right this is kind of a parcel box on wheels right as you can see from the design you can walk up to this vehicle type in your code and one of the lockers is gonna open right so if you think about why are companies thinking about this well there are several reasons right one is this is an autonomous vehicle so there's no driver so that saves on cost the other thing that's interesting parcel boxes take up space right and space is not always available right stores take up space here you have a mobile solution where all you need is a place where you can park for a little bit and then move on so this is kind of a very interesting concept and we'll get back later that there are people that believe that this is what we're gonna see more and more of in the dominant mode of delivering parcels right the other thing that it does is it doesn't go to your home right so it still has a consolidation opportunity right so it's kind of mitigating possibly emissions time and also still making it convenient we could have this come to the parking lot of the isye building which currently doesn't exist and we could go and pick up our packages there right which is a good mode of transport electric vehicles I think here the main motivation is climate change right electric is better than diesel or whatever and I think this is going to be big and it's also a very good thing for society to see that companies like Amazon are gonna use these more and more right and like anything if Amazon orders something they're gonna order a few here are some other examples of the electric vehicles the one thing to note here is that the Amazon one is still pretty large that's more for possibly home deliveries in really downtown areas you want to have smaller versions that may be able if it's legal to go on the sidewalk etc so there's a lot of different designs that are happening right now of course in inner city especially in Europe where vehicles may not always be allowed to go into the city center at certain times bike couriers are in many ways ideal right it's efficient there's no emissions you can have a combination of pure bicycle or electric bicycle it's a very nice option the last thing I want to point out is this is I think another example of where things are going right so here the idea what we're looking at is there is kind of a parcel box where you can go to pick up your package but this is gonna be supplied fully in a fully automated way right so a drone can come and then there is technology inside that can just handle and so everything can be done without human interference other than picking up your package right so so one thing that this shows that that there are a lot of creative people out there that are trying to make last mile delivery both easier cleaner and more efficient right there's a lot of things happening which is why this is an exciting time I wanted to show you something that was published in a recent study by McKinsey about different mechanisms for parcel livery and I'll flip through them quickly so the first one that you see there is traditional one you have a person driving around in a vehicle stops at the front of your door and hand you the package right drones we've all seen the little videos of Amazon the other one and we'll get to that in a minute is crowd shipping right so it's not the Amazon employee that's gonna make the delivery but somehow there is an app and when you say I need to delivery somebody and in an individual like you and me is gonna make the delivery and we'll talk a lot more about that in a minute then we have the droids that we've seen by couriers semi autonomous ground vehicles and the ones that I pointed out right kind of the mobile parcel boxes which also this study believes is going to be the future right autonomous vehicle like a parcel box that can get to strategically chosen locations and you can go there and pick up your package right that's what they believe is gonna be the big future all right so where does the sharing and geeky economy coming right so let's first make sure that we put the concept there so that we know what we're talking about so if we look at the sharing economy right as the name suggests we are sharing asset among individuals most right that's the big idea the geek economy is more that you are no longer an employee but you work on being right so you for a short period of time you perform a task and that is a gig and you get paid for that right so that's so and clearly if we think about a company like uber this is where these two are coming together right many of the drivers that work for use their own vehicle and so in that sense they're sharing an asset and some of them at least are performing gigs right if you only work for an hour it's considered to gig I mean we've seen and this is a downside of this strength that there are lots of people that actually work for eight hours so de facto they have become an employee but they're not treated as an employee and don't get the benefits that employees get and this is of course where we've seen or at least some of us in California they're trying to change them right the a B five legislation says if you work for eight hours for uber uber should consider you to be an employee right and we'll have to see what happens with that because they're gonna fight it right something interesting dick economy seems to be growing rapidly I grab some statistics here right that we might may find that in the near future most of us are working in the gig economy rather than as an employee of a company or the state like I am all right so as I mentioned when we put sharing and geek economy together we're getting kind of the uber like activities and if we sort of put that together again with last mile delivery then the prime example of that is meal delivery right another important one is groceries that share similar characteristics although the pickup is typically in grocery stores or your Walmart rather than a restaurant but it shares similar features right and I want to specifically look at this a little bit because we have spent quite some time with one of the players in this area so this is not a u.s. phenomenon right we see it happening everywhere I had a chain to talk with people from my toin not so long ago and I'll this is the Chinese version of meal delivery and if I have time I'll come back to that particular example a tiny bit later here are some statistics for the u.s. basically it shows that it's growing the other part that is important is and I don't know if you can easily read that platform to consumer delivery which is the blue part and that is kind of the GrubHub the uber eats etc that we're going to focus on right of course Pizza Delivery we've known that for ages what's new in a way are these aggregators that are really taking over these pcs if we think about market sizes not surprisingly China is the biggest market for meal delivery of course they have a huge population there is big adoption of technology it it's big but the u.s. is not that far behind certainly in terms of money alright so here are a few players in the US and I'm gonna focus on GrubHub which is one of the large players in this arena so we probably all have heard of group up certainly the students here in the audience may have even used it one of the things I want to point out here is that day and I take their website yesterday if this number was still accurate they're making 500,000 deliveries per day and if you think about this clearly that process can no longer be manual right when actually when I started working with group up four or five years ago it was still manual right and it's amazing how quickly and of course they realize that that's not maintainable if they continue to grow right so you need technology and that's partly what I'm gonna talk about so of course we know how it works we have an app we look at what we want to eat we press a button and it gets delivered where we are and if you think about it of course this is one of these examples where you want instant delivery right if you are hungry and you want to eat you don't want to wait three hours right so in this is the the prime example of an environment where the service guarantee has to be short and has to be instant right so I'm gonna talk a little bit more about the o.r side of things and they're basically three takeaways one is that what happens in most of these situations now it's kind of repeatedly solving an assignment problem and we'll get into a little bit of detail in a minute the other one which I think personally is the most intriguing is driver capacity management right if you think about what is the big change in the sharing economy and the use of crowd shipping it's that where in the past companies owned the delivery capacity like UPS owns the delivery capacity you had private fleets etc that's all changing right with the sharing economy you now have tons of people that are not your employee and you want to rely on them to make deliveries like group hop doesn't own for own employee any drivers right they're all individuals that use an app like uber to say hey I'm happy to drive and deliver for you in the next hour right how do you manage that capacity remember that they meal delivery you want to be sure that the meal is delivered let's say within 45 minutes how do you ensure that you have the delivery capacity to make that right this is a completely new kind of question that we didn't see 10 years ago and a very challenging one and we'll talk more about that in the future the other thing that I have observed and I'll spend a few minutes talking about that naturally when group up started the focus was on what they call the diner the customer that orders the meal right all their focus was on how can we make that customer that diner happy right how should we build our optimization or decision technology to make the diners happy right what we see now is that that there is a big change happening that they realize that they also need to make their drivers happy in part because they're not their employees right and unhappy drivers may instead of accepting tasks foregrip up may actually say I prefer to work for you or uber eat and so this is a big change and there are lots of interesting opportunities for people that like to think about logistics transportation and optimization to see how to go about that or pricing its engine but we'll get back to that there's a lot of works here but we all understand the basic principle you place an order and that gets somehow delivered to your home the bottom part is probably the most interesting part that says that there is an awful lot of uncertainty in this whole process right how many orders are placed because it all happens on the app it's not like I place an order and say well can you tomorrow evening at 6:30 deliver a pizza right it's not how it works it's instant delivery that matters there is some prediction that can happen and we'll get to that in the end when we start talking a bit more about machine learning similarly meal preparation time right is uncertain right these aggregators talk with the restaurant have some agreement on sort of what they can expect but many of the restaurants have also seating and people in the restaurant and they often prefer to keep them happy so there's a lot of uncertainty even in when the meal is ready right and then there is traffic and all those other things parking are you in an apartment complex there's a there's a lot of uncertainty in this particular environment so how do these companies the ones that were aware of uber eats and GrubHub actually solves this problem well the core is repeated assignment and we'll get to that in a minute so so what are the main entities or elements in this process right on the one hand you have the orders and on the other hand you have the drivers right and so an order also has a restaurant associated with it so this is where the pickup takes place and of course an order also has a delivery address and a driver even though I didn't put it here in some cases you know a little bit about how long they're working and we'll get to that in a minute if they're purely signing on then you may not know how long they will be there for you but in the GrubHub case many of them have a fixed period of time when they work and then what I didn't list here is you know what their current list of tasks is that they've already been assigned to them and where they are or where they become available again right and then there is okay so we need to put these together right then if you think about optimizing these decisions then you need to have an objective and so it's actually already quite important to think about what the objectives are right and the two that we have been using most are click to door time which means I push the button when is the meal actually arriving at my door and there is also ready to door which is when was the meal ready at the restaurant and did it arrive at home right and clearly the first one is my total weight and the second one is freshness right and so both are important but of course there are also drivers right because that's your cost in the system so you want to utilize them as best as you can so somehow there are all these different objectives that you have that you need to put together in an optimization based approach and I wanted to look a little bit more at these two basic performance metrics at least a way that GrubHub thinks about them right so when you place an order internally they have a target let's say within 45 minutes we want that order to be placed and if you think about this this already is kind of it's not a strange concept but it is already something where you could have questions whether this is the right way to go right because clearly if you live next door to the restaurant it's gonna be easier for me to make the delivery then if you live far away from the restaurant so should the target be the same but even if you think about the target right should it be a function of the placement time right so should the target be different at 3 p.m. versus 6 p.m. we know that to order volumes at 6 p.m. are going to be quite different right should it be different based on the particular restaurant right some restaurants tend to be busier than others should we take that into account when we use our internal target right and time of day day of the week all these things already in a way you need to think about them and group up it's usually just 45 minutes right but there are lots of questions about that and similarly you can think about ready to door time what in the optimization the way they think about it is if you sort of exceed the target time then somehow something that's not good is happening and you want to avoid that right so this is how the optimization deals with these kind of questions so what are the challenge is there's a lot of uncertainty in this system right so basically what you do is you have a mostly reactive approach right at a certain point in time you look at what you know and do the best you can right so here is that process right here is we sketch how it works so at some point your day start and at some point your day ends meaning the time that you accept your last orders and then during that day every few minutes you make decisions right and at GrubHub this happens every minute right but they're already interesting questions is a minute really important could it be two minutes right then we maybe can batch more and there are lots of questions there and each time you solve an assignment model and then you have to commit and I'll be more specific about what that means right so we for those of you that have seen our models right we have the drivers on one side and what they do is they look at all drivers that are active at that time right so for all drivers you can look at at what time in what location do they finish their last assigned tasks right so that's the time that matters and then the other thing is you have orders and you have choices to make already there do you look at individual orders or do you already group let's say orders that are close to each other into a single kind of unit which you want to assign to the driver right and then the biggest biggest challenge is how do you actually put a weight or sort of a measure of how good is this particular assignment right that's where in a way all the knowledge and proprietary information should go right every company uber each uses these kind of methods and they think they do better because of how they choose kind of this evaluation about what are good measures to use that it's not easy I try to illustrate with with a few small examples right so let's look at the particular order which is the blue circle and this is the time when it gets ready at the restaurant note that this is already uncertain but in these optimization models it's assumed that we know this and you have two drivers one that's a little bit further away but is ready with this last task or completes his last task a bit earlier so it can drive a longer distance and get there before the order is ready or you have a driver that is much closer but only finishes a little bit later and therefore he can get to the restaurant only a little bit after the order is ready right now which one should you prefer right it's not completely obvious right in a way maybe the driver that is nearby because when he drives through the restaurant he's driving empty and not doing something useful would be preferable but if you think about the customer maybe picking up the order when it's ready may be preferable right so there are a lot of trade-offs that have to be made I mentioned commitment strategies earlier the reason for talking about commitment is as I said to look at all the drivers and all the orders that are in the system and you make the best possible assignments that doesn't necessarily mean that those assignments have to be implemented right there is sort of a second step where you look at all the assignments and say should i implement this assignment or not and the simplest case in a way is if I made an assignment to a driver that is still actively working on a task and it takes three minutes at least for that driver to finish the task and I'm gonna run my optimisation again in two minutes maybe I don't need to commit right because I have another chance to see if more information is available and I can make a better assignment for that driver right so some assignments we simply ignore then there are also assignments that you say ok you have to start driving now to pick up that order there's just nothing we should do otherwise it's going to be very linked you can also make decisions well start driving towards the restaurant but when you get to the restaurant we may actually either change or augment the assignment that we gave you earlier right so in the commitment part there are also a number of different issues that come up another important aspect is bundling right I already mentioned that maybe sometimes you want to group orders or bundle orders and assign a bundle to a driver right and again if you think about this the trade-offs are not obvious right bundling can be more effective in the sense that going deliver come back to the restaurant and then deliver again of course involves partly empty driving takes up more time so it might be better to pick up two orders at the same time then go deliver one end for the other right it's a more efficient use of your driver capacity on the other hand as this example shows is that if we have three orders that we want to put together in a bundle of course I can only pick up the bundle when the last order is written right so order number three determines when I can pick up this bundle but that means that order 1 which was already ready earlier is losing freshness right and then there's also the sequencing of how I'm going to deliver them right if on top of that I'm first doing 3 then 2 then 1 then that order is also in the vehicle all right so even though bundling is necessary at some times during the day how to do it is not necessarily an obvious question so here are another aspect that we looked at with group up is this is some curves that show what happens during a day right so as you expect there is a little bit of activity during the lunch hour and the big activity happens during the dinner hour right now what's interesting or the most interesting thing is the Green Line the Green Line shows how many drivers become available so finish a task for their last assigned tasks in the next 15 minutes right so you can just do a plot and and create such a graph and then the other is of course you can also plot how many orders sort of were placed at a particular time which is the blue curve at least to think about this right and so if if you look at the green graph and the blue graph you see that there is some challenges happening at dinner time right the number of driver that becomes available in the next 15 minutes is relatively small compared to the number of orders so you have to start bundling if you want to have any chance of success right so one of the kind of things we worked on is trying to design a metric that helps recognize when it's time to start bundling more aggressively here is how bundling works we've seen this graph I mentioned this already in a way if you look at it at lunch time on this particular day they had too many drivers they didn't need that many but at dinner time they had too few right and this already points to this capacity management question that I believe is so important you can look at some other things I mean wind drivers one thing that you see here is that you had a large number of drivers during the lunch period you had an even larger number of drivers during the dinner period and fewer in between but the work that they do especially the one so a lot of them came on duty at 5:00 right but they spent a lot of time driving to the first restaurant right this is the green area and it's not very effective you're not using them effective so can you do something about that when you think about driver management all right there's a lot more things that you can do I'm going to skip that I do want to point out a few of the practical complications you may actually lose communication with your drivers right it's one thing to say I optimize every minute and I'm gonna commit and then you say well there is this job that you can do or this task that you can do first of all tell me whether you're gonna accept this right they're not your employee so they may actually say no and they say actually no very often which is frustrating of course for a company like grappa I was surprised when I heard that that's in the order of 35 to 40 percent right imagine how to deal with this right that's challenging and I mean of course uncertainty etc etcetera etcetera all right so as I say it and this is one of the things I want you to take away right from an academic perspective this is I think where most of our opportunities are to do really something innovative right how do you manage this uncertain capacity right this is totally new we didn't have UPS doesn't worry about it Amazon worries about it a little bit but certainly companies like grappa and uber eats this is a big big thing and we're currently also working together with roadie which is sort of a local startup that actually is doing home grocery delivery for Walmart relying again on crap shipping crowdsourcing how do you manage to make sure that you can deliver in an hour or two hours how do you know you have enough drivers right so how to ensure that the sufficient number of drivers is available to deliver the placed orders and to meet the service problems right the later part is very important if you promise delivery within an hour or within 45 minutes how can you make sure that you have the right number of people so I split this up in two questions one is how do you even know how many you need right I mean it's one thing to start thinking about maybe how should i reward them but the first thing is how many do I actually need right and that's not an easy question right because it depends on your dispatching strategy it depends on the order patterns it depends on the geography of the city in which you operate these questions are not easy to answer right and with roadie and a PhD student here we're working on using machine learning techniques to try and estimate how many drivers we need for a particular store at particular times of day or for particular order patterns and then the other question is even if you know how many drivers you should have how can you make sure you get it right and so the latter question we looked at the little bit also with group hop because of course the only way that you can control them somewhat is by means of compensation right and so in the GrubHub case which is a reasonable approach is they say we're going to actually request that you commit to working for us for a certain period of the day let's say between five and seven tomorrow and in return for that commitment we want to we we give you a guaranteed minimum pay right and hopefully that attracts enough people to sign up right otherwise you're completely dependent on what orders I assign to you but here at least you have a minimum pay guarantee on top of that we may also ask that you can only per hour reject one of our tasks that we offer you right so you can structure this compensation in different ways right but what that means now is if you start planning such a system there are two different types of drivers in your system one the ones that we label scheduled right which have accepted this minimum pay guarantee and in return tell me or or are available for me to assign but you also still have the people that just open the app and say I'm here if you have a task for me I'm gonna do it right so we looked a little bit at developing technology to determine how many scheduled drivers you would need right so what these blocks would look like how many scheduled drivers are needed and what should their blocks look like right so we developed some technology for that and if you think about it so that's a tactical planning problem that we solve the night before and of course during the day there is dispatching dynamic routing and you have these unscheduled drivers that also show up right so this is quite a different difficult problem and you want to maintain some service guaranteed right 95% of the orders are delivered at or before the target right that's kind of how we incorporate that we use continuous approximation value function approximation if people are interested in more detail we can talk about that a little bit later we tried it on meal delivery data from Iowa that's publicly available for different strategies and our strategy seems to work in the sense that we can make drivers available at the right time for the right amount of time better than other strategies which saves you on cost for the company so as I mentioned the third thing that I wanted to talk briefly about is the fact that now suddenly companies are starting to worry more and more about the drivers as opposed to only the diners one one situation that we encountered early on at in our work with group up was what we label driver drainage right and what that means is drivers are not stupid people right over time they get the sense of where am I gonna be making the most money right should I be in Buckhead or should I be in Alpharetta should I be downtown should I be in Virginia Highlands they somehow start recognizing that if they were in a cooler area either the tips were higher or they got more orders per hour or the distance between the restaurant and the delivery address was shorter which is good for them so drivers start to converge all on areas where they believe they're gonna make more money of course it doesn't work like that because you have too many drivers there then you can't assign but for the company the biggest worry is now you have areas where people place orders and I don't have any drivers right that's the worst part so how do you deal with such situation right so one opportunity that you can consider is actually restricting drivers to particular regions right so we only assign you orders from this particular area right on the other hand we can also see that that can be limited right if you don't need that many drivers in that particular area for half an hour or an hour in the day how did you start thinking about sharing right so should there be sharing of these kind of ideas the other thing that you notice is drivers tend to like operating in one in the same area they familiarize themselves with the delivery addresses where they can park etc so and especially if they live in that area they don't want to wander around all of New York City for example right they rather work in only the Bronx or whatever so how do you start incorporating these ideas in your decision technology is a big thing that the companies are now thinking about and then as I mentioned before you've probably seen it especially in the uber lyft situation these drivers work for multiple platforms at the same time right so how do you create loyalty how do you make that happen that's not an easy thing another thing that we have been thinking about not really developing technology that GrubHub decided to use but the other thing that is quite interesting and intriguing again to me is that you control what's shown on the app right if let's take a simple example if a customer wants to order a pizza I decide which pizza restaurants that person sees on the first page on the screen right now clearly probably you pick restaurants that are not too far away from his delivery address because that's one what he expects or she but also probably what works best for you but if you happen to have a driver already at a particular pizza restaurant maybe that one should be at the top of the list right a pizza is a pizza at least to me right so how how do you manage which of the restaurants you display now of course your contracts are with the restaurants and so you don't have complete flexibility but you have some flexibility and they're interesting questions of how do you manage orders to minimize your logistics cost while still satisfying the surface promises that to men they're very interesting questions that students here that's still looking for interesting topics to work on certainly this is one of them so I mentioned research opportunities in this space there are plenty around capacity management order management and there's still lots of more things that we can do even in the driver order assignment how can you look ahead how can you incorporate predictions about where sort of the the next batch of orders is going to come from this also indicates that there are great opportunities potentially for machine learning right all of us in the room I'm sure have heard the buzz around machine learning and that it's gonna solve all our problems once you can solve go once you can win it go you can probably solve any problem in the world right of course it's not that simple but there are certain types of things that could be very helpful here right predicting future orders is not easy and the standard forecasting techniques may not be appropriate right because what you really need to know or maybe want to know meet is a big big statement but want to know is delivery location specific right so there's not that much samples from a particular location that that allow forecasting to work with great accuracy and a lots of things that influence these the time of day the weather of the day I mean I mean traffic conditions so it's possible that machine learning can do better than the traditional forecasting techniques order acceptance is a very interesting one we try to do some basic regression modeling with GrubHub to see what influences that decision right we thought maybe if a driver is at the end of his or her shift he wants to make sure that he ends near his home is that important it's the tip and we had a whole range of features it turned out that the best predictor was if he had actually often rejected orders before still useful information maybe but that was a very basic simple model maybe there's more that we can do so I think there are certainly opportunities for machine learning what is the right time to start trying somehow to get more drivers in the system right as we said we in our decision technology developed at least one metric that we use to be more aggressive in bundling but maybe something similar is needed to try and attract more drivers right search pricing or whatever it is right and again machine learning might be helping you to make decisions like that all right I wanted to end I mentioned that I had a chance to talk with people from mate one when I was in Beijing in December and they're one of the biggest meal delivery platforms that exist and also remember that I mentioned that grub hub now delivers 500,000 orders per day as you can imagine in China things are always a bit larger right so here are the numbers that you have so they actually have 27 million orders per day that they need to deal with thousands hundreds of thousands per minute that come into their system right and typically they have 600,000 drivers working for them on a particular day right now I mentioned when I started talking about Amazon that the numbers get big right and that the time to make decisions get smaller and smaller this is another great opportunity for our community the optimisation logistics community to start really thinking differently about how we might think about even addressing these problems right when you get these numbers you need to work hard so they have what they call themselves the mate one super brain and I'll let you see what kind of techniques they use parallel computing that's not the big surprise machine learning data mining prediction lots and lots of things the biggest thing I think that is helping them is how they select for each restaurant the acceptable delivery locations right one way to scale down your problem is to say well I only deliver to the person that lives right next to the rest of course you don't make money doing that but that makes the delisting logistics problem quite easy right and so this is a big thing thinking about the service area for particular restaurant right and even that is not a simple problem to answer right and that again depends on your dispatching the delivery capacity that you have there's really phenomenally interesting problems here but scale to me is certainly an interesting one not only in Lille delivery we see that also just problems with UPS right more and more packages are going to be delivered problems get bigger and bigger we have less and less time to make decisions how do we deal with that and that requires some creative new thinking all right and that's it for today I am happy to answer questions if you have them yes this is a I mean a valid comment and a Valentine you're making we worked primarily with group up on the transportation side we are transportation experts and but I think it's a very legitimate question to say hey this ready time right is impacting the time you have to make the delivery so focusing some of your attention and effort into streamlining maybe the meal production would be a very valid and useful thing to do it is something that was never mentioned by group up as something that they wanted our help with but I fully agree with you that that's part of the answer possibly to the big picture question yes you're absolutely right yes so I think this is more I mean general question that somehow society has to think about more right the value of all these companies is really determined by the stock market not by their actual business right uber even Amazon is not making money they're only making money on their computer server business not in this package delivery part they're losing money there and so it's an interesting question right I mean part of it comes back to what I said very early on we as customers are not willing to pay for this service right and in a way if you think about it what we're asking for and again this is partly the fault of the companies themselves what they're offering to us instant delivery is costly the less time you have to make deliveries the more costly it is for you I mean it's a simple logistics concept so I think at the moment so one way of course that you can try to do this is by exploiting crowd shipping where you don't pay for benefits and at least your cost base is smaller but even then it's probably not yet small enough to recover the cost I don't know it's something that people have to think about I mean certainly autonomous delivery might be getting there which is why I think at least the package delivery companies are looking at these kind of mobile parcel boxes as the future and I think we need to again try and think creatively if we can actually make it profitable yes it is more likely because that also in so so I mentioned that the the predictors that we found to be the most helpful our tip amount and I mean just early a previous behavior I think another I do think here and I'm making a broad straight statement here is that the other thing that really is shown here is that we need to integrate our optimization with more human behavior modeling right that is not something that we have done in the past and I don't necessarily think that machine learning can figure it all out by itself so talking with psychologists and and thinking about it is going to be quite important there there have I mean the other thing which of course is hard to understand just from let's say group up data if a person doesn't accept an order doesn't mean that it wasn't somehow an okay order for the person but if uber offered him something better at the same time he may just have chosen the uber order right in I at the moment as far as I know they're not tracking or cannot track that information so this makes it very hard to understand exactly what is happening is it the order that the task that we offer or are there other things that play the one that I do know and and you see this at uber as well this is why they start creating compensation scheme that discourage you from not accepting orders right I think uber now has something where you get kind of a bonus if you do a certain number of tricks during the month or during a week right and the the main reason for doing that is to say hey if we offer you something if you want to reach the bonus you might as well better accept it right and it's very hard for an individual to see let's say at the beginning of the month what should I do right they have this target now you have to start thinking once they reach the target do you have you said it wrong do you need to start putting in extra compensation schemes there are lots of questions around this particular aspect and and some of them are not easy to answer because only now people start trying to really investigate the impact of people working for several of these platforms at the same time yes yes and and this mean yes so this is what you have seen a little bit with uberpool right uber pool didn't work very well at the start and again if you think about it the reasons were fairly obvious right if people just say I want to go from A to B now and they expect the vehicle to show up within three four minutes the chance that you can also reroute it to find somebody exactly on the route to where you go or smooth right so the people at uber pool have started to think about different schemes right where you announced earlier to give them more time to find consolidation options so that's an example of where exactly this is happening you need to somehow incentivize right if you were pool is a good example they said well it's going to be cheaper for you and they didn't get anything in return and then couldn't make it work right now they say it's going to be cheaper for you but you have to tell us a bit earlier and we will send you a message when the person is going to be at your door etcetera so clearly working on trying to get information earlier it's very important and we've done studies also for GrubHub to see what the impact would be if they knew 15 minutes earlier and there is a big impact how to make it happen is less clear yeah all right thank you very much I hope you enjoyed it [Applause] you

Channel: GTSCL

Views: 1,148

Rating: 5 out of 5

Keywords: supply chain, logistics, Georgia Tech, transportation, delivery, crowd-sourced, omni-channel strategies, Uber, Lyft

Id: RIDXTDCtOtU

Length: 70min 41sec (4241 seconds)

Published: Mon Feb 24 2020