Cloud Crunch
Cloud Crunch

Episode · 8 months ago

S2E03: How McDonald's France is Using Data Lakes to Improve Customer Experience

ABOUT THIS EPISODE

Adrien Sieg, Head of Data at McDonald’s Global Technology France, Christina Moss, Director of AWS Cloud Services at McDonald’s, and Mathieu Rimlinger, Director of Global Technology France at McDonald’s, talk about their latest technological advancements in the cloud and how McDonald’s is using data lakes to set customer expectations and improve satisfaction.

...involve solve, evolve, welcome to cloudCrunch, the podcast for any large enterprise planning on moving to or isin the midst of moving to the cloud hosted by the cloud computing expertsfrom Second Watch, Ian will be chief architect Cloud Solutions and SkipBerry, executive director of Cloud Enablement. And now here are your hostsof Cloud Crunch. Welcome back to the podcast. Everybodyvery excited. Today we have, ah, a group of people from McDonald'sCorporation with us today from around the world, and I've had the opportunityto work with McDonald's quite a bit. And it's fantastic. So once again, myname is Ian Willoughby. I'm with a second watch, and today I am joinedwith Adrian Sig, head of Data and McDonald's Global Technology France.Christina Moss, director of a Ws Cloud Services at McDonald's, and Matthew RimLinger, director of Global Technology France at McDonald's as well. They'rejoining us today, and we're gonna be talking about the latest technologyadvancements in the cloud and how McDonald's is using Data Lakes to setcustomer expectations and improve satisfaction. Now, Little commentaryfor me before we get going into this was we had an opportunity to do? Ah,very, very interesting project with McDonald's. We call it the France DataLake. And really, it's gonna be a fun story to share with our audience andreally talking about how this technology used thio drive some changethere. So welcome, everybody. Christina. Goodness. I had the opportunity to seeeverybody while we're recording. Just so I know the audience doesn't Good tosee you. Good to see you, Matthew and Adrian as well. Welcome to the podcast.Thank you. Yeah. Oh, this is great. Yeah. Yeah. Welcome, Welcome. Now I'vegot some questions. I just kind of want to ask you all and what kind of diveinto it if you're ready. But first of all, and I'll let anybody take thisquestion. Let's just I'll start with it out there. How did this idea of usingthis data, like for this goal, really help improve some customer satisfaction?And And how did it come about? How did you kind of conceptualize this? Maybe Ican start s a new training in McDonough. Written off course margin Eri speakingin a fast food industry is that it's no longer customer satisfaction. That isimportant, but just in our experience, which paves the way to an explosion offtypes and and sources off data. Consequently, McDonald said, the needednew way off storing and manipulating data and this fax involved off coursein shifting from the traditional data. Where are usedto more flexible,scalable data like so what we have to keep in mind in McDonald's. We switchfrom a transaction centric view to a customer. Contribute as that super andmore complete perspective on customer lifetime value and consequently, we weneed. McDonald's need full visibility...

...into every step in the customer journeyto understand what customer allies, what they don't and what McDonald's cando to improve the customer experience and, as a by product improve conversionrate and will die. Alte. Yeah, No, that's That's reallyinteresting, Matthew. Do you have anything? You can add her? Yeah, Andmore generally, we knew that we were sitting on a mountain of data becausewe're talking a lot, A lot, a lot off that, but we're not using it. And thatwas quite a pity. And and I really wanted 18 months ago to start a projectto the eldest, leverages data and do something about the data because we'rejust missing an opportunity to do wonderful thing with the data we hadregarding consumer experience and any other field off data, because I believewe have hundreds of fields of data to to experiment on. And then Cristina,you're sitting at the corporate headquarters. Well, not actuallyphysically these days. But a zoo Most of us were working from home. Is this atrend that you're seeing across the globe with all your different businessobjectives? Yeah, I think you know The interesting thing about McDonald's isthat you don't really think about us as a technology organization, but wereally are, like everything we're doing in House is really like to figure outhow to improve things in the customer experience in the stores. And I dothink this is a trend that is growing. They're just and it's kind of thedirection we're heading more, more more now, even as this year has beenprogressing and I think into next year, yeah, and I have a unique opportunityto not only work with McDonald's is a client of second Watch, but also many,many other multinational and large enterprises. I can honestly say thatwhat you all are doing is much further ahead than most of the industry. A Sfar.The way that you look at digital and it's it's a really, really excitingopportunity. I don't wanna, you know, no shameless plugs here or anythinglike that. But it's just it is really, really fascinating. And I see it fromkind of the whole arc of your business is, well, like it's really it's aboutthat customer experience. And I did enjoy a nice Big Mac the other day andusing the mobile app it was fantastic. So I'm gonna kind of move on here alittle bit. But I want to really kind understand, Like, what were some of thefactors that drove you to this? You know, I I think there's there's a lotof things going on. One is you have, ah, large set of data that you're able tolook at, but also the technology has been changing very rapidly that you canuse. Can you kind of tell me what the catalyst waas and I'm gonna go to theFrance team here initially, is what what was the driver? What was thecatalyst to say now is the time to do this, the technology became much moreefficient and much cheaper to so we entered the opportunity to to build asystem with a nisi return on investment,...

...a fast return on investment that was akey for us. We don't have to invest so much and wait many years to get afeedback and a return engagement. That was a major catalysts toe. Help us sellthe project and engage the project. And after that, we had to find a goodresources and the good people, the good partner to walk with. And then whatkind of data are you actually analyzing? And what are some of the outputs thatare associated with us? Basically, we collect any form of data from fromanywhere within McDonald's, numerous data sources and and zeros from roofnew numbers to show social media streams and anything in between and offcourse, we got a vast amount off operational data. For example, a speedof service is a case in point on its data coming from the kitchen managementsystem. Basically that there is an explosion off sources restaurant haveaccess to much more data and many more data sources and than ever, ever beforeinternal external commercial correction or structure and structure. So we havea lot of data and the underlying goal off our Data Lake and take behind thesaying is to reduce the effort needed to analyze or and process the same dataset for different purposes and and by different application. So that's why toanswer your first question, we was looking for a scalable, fault, tolerantdata platform that processing architectures and framework usingsomething very common in a in the tech world but uh, loosely coupled anddistributed system and basically toe to say in keeping with Mature said Way waslooking for a manage solution that scale seamlessly and put less focus oninfrastructure. Thio Hello, McDonalds team to focus on what really matters sodata and the resulting inside from data. So if I be complete three main type ofdata that we're using today for dash boarding or editing, uh, sales,obviously speed of service and consumer behavior. So the three main type ofdata that we're using but still there is one we are walking that I wantedThio to put in place within the Data lake is any kind of data that comesaround whether you know or don't know what you're gonna do about it. We grabit and we store it in the head. Lake storage doesn't cost much old storageand customers, so we're gonna grab any kind of data internal, external. Thatis that we don't know what we're gonna do about but so that when we're gonnahave a use case, relevant use case about data will have it. We have ahistory about the data and will be more efficient and faster. So that's also apoint off importance in the way we did put that are taking place. And we wannawalk tomorrow. So even though there are...

...three kinds of data today that aremainly used and and more important than the other ones, if I may say so, sales,speed of service and consumers experience, we're gonna grab everything.And we're already reading everything from I t. Data around monitoringdevices, something that we're going to do next year, or ticketing or whateverand also saves etcetera, etcetera. Now, when you were getting ready to deploythis design it, you know, and then you get into the implementation phase Wasthere any unexpected challenges that you experienced while building thisother than like how slow McDonald processes are, like our internalprocesses of, like, just getting things going sometimes? Well, everybody facesthat problem. I think so. You know, universal universal challenge that thatis definitely, definitely universal challenge. But beyond that, let's sayyou know, is there any like, you know, from the technical aspects, theoperational side, any of the collecting of data because you mentioned lots ofdata sources Now that air coming in Was there anything that that you had toovercome any of those challenges? Uh, but basically, from my perspective,it's something very common. It zits rated Thio data inconsistencies becausein in the world, off additional point of sales, we sells happening online,offline in dozens off a different location. Collecting and aggregatingthat fermented data is a fit off. A huge complexity, very huge complexity.And when you try to put in a single place that are coming from many sources,and from many data vendors from mobile labs from loyalty programs from serumtools and many, many more, other digital interface is very hard to havea single premarital. Tito draws in the entire customer journey into the owndigital ecosystem because sometimes the fact you have explicit customer andengaged are missing the different digital platform and and loyaltyprogram. So from my perspective, the challenge we face is to find the linkfor a given customer to to put the food puzzle together and as the big picture.And as Christina said Andi, it was a beautiful project because we work in inmany locations in the world from the U. S. From India, from France, from UK,with with a lot off James. So it waas incredible Thio to put everybody on thesame table and and thio to discuss with the all time. So it was an amazingproject in terms off technical expect with issue we got and off course interms off off projector itself. That is like one of the best things about, likeworking for a global organization, is that you really get to kind of interactwith tons of different people from different places. And the one thingthat's like pretty universal is a...

...common goal, right? Like everybodywants to get the things done and work towards, you know, whatever thesolution is so I think that's probably the biggest challenge. Is just the thedifferent time zones in the different places and everybody working around theclock kind of. But I think it's been good. It was really good. Yeah, it wasfunny. Sometimes We got some called, uh, very early in the morning with withIndia and very late on the evening with us. So it was very funny to work with,ah, lot off people. And, uh, and Christina teams and based in in Indiain in us. So yeah, it was incredible. Yeah. I think your organization reallyunderstands how to work across those global boundaries as well. Better thanvirtually anybody I've seen. So that's it's really interesting. And obviouslywe're all working remote to some degree right now, These us that are officeworkers. And, uh, we still have people obviously out the front lines, which weappreciate their efforts out there. But have you seen the results that you wereexpecting out of this project? Has it helped you meet your goals as it anyunexpected benefits come to the surface that you weren't anticipating? Maybeyou could kind of share some of those stories. In fact, we've reached a fewgoals. We've produced speed of service, dashboard, uh, something that didn'texist previously. And we're now able Thio, analyze and provide census aboutthose those kind of data. But the most important thing is that we've openedtense off those tens of possibilities and opportunities that we didn't havepreviously. So this brings a lot of challenges and a lot of possibilitiesthat we didn't have previously and and that's the most important thing. Butwe're to me. We are only at the beginning of the journey that we havearound data and Algeria will talk about the specific goals and objectives thatwe did rich until now. But the first goal that would read reach from mypoint of view is that we've been able to to show people and show ourorganization that we were losing a lot of opportunities, and now we have theability to go through on to move on into those opportunities and to to walkon new data, new possibilities and to stop walking in data like handcraftedwalk on data. And that's the main goal that we achieved. We we proved theFrench organization that it was possible and important to invest andcontinue investing in data. Yeah, basically, I can deep dive into oneobjective. So I joined McDonald's on the very beginning off February and andmature. Ask give me one goal as the speed of service because, as you canfigure out, it's the only important for fast food. Shame to know, to monitor,to have visibility and on speed of service. Eso Basically we follow ourown national Operation Team toe, get...

...visibility into their business and offcourse logistic operation. And they can make accurate decision about where andhow toe to implement changes that we make a new impact. Operationalefficiency and gustatory on DFO from McDonnell Operation Team. Getting dataabout prep time is the first step in gaining insight that can improve otherfulfillment efficiency for our listener, as the speed of service in McDonald'sworld is the time restaurant text to serve the customer, the clock startwhen the customer placed is air order in the restaurant or pull at the drivethru and the clock stop when the food is delivered to the customer. So, asyou can figure out, there are a lot a lot. A lot off data toe give you some,uh, some some figures. We serve 1.5 millions off customers daily and eachtransaction produced maybe 20 or 30 point of data. Eso we are in the in thebig data world because we collect data from prep area and kitchen andoperational team Now can get to better understanding off how long prep time,text and off course in the future with Second Watch Team and the Christian Ateam machine learning and model prep time to provide a more accurateprediction off Our long prep time will take in the in the future. And maybethis is a supporter of big data and the objectivist off mature to monitor, tounderstand, toe get visibility into speeder service. And in France you cantrust us. It zbig revolution toe can monitor as a different prep speed ofservice area speed of service. And because of of course, you can say Okay,here it's like too much time to take other Here. It's like too much time toto prep food and explaining all other some stuff like that. So here is aconcrete. A result we way we have done together. Yeah, this is something wetalked Thio all our customers about. And we actually don't have to talk toyou guys about this because you are all down the road. But if you're notcollecting your data, doing that analysis and starting to think abouthow to apply machine learning for whether it be forecasting predictions,wherever else it is you need to because your competitors air going to get thereeventually as well. So you might as well be in the forefront of it of thatdigital transformation. So it's very, very exciting to see you all do that.Now that you've done this in France and it's it seems like it's a pretty goodsuccess. You're getting some data out of there. You're you're making somedata driven decisions. Are you planning on doing this in other parts of yourbusiness, either today or in the future? I think to like some of the reportingand the visibility you've been ableto have is is really cool, Like some ofthe reports that I've seen that came...

...out of that were really nice, and youcan actually, like collectively see data that you have across from a bunchof different places now in one central location is pretty cool and veryhelpful. I think you know, if you're trying to analyze what's going on, Yeah,and I also think it's interesting, too is because a lot of us have come fromtraditional I t backgrounds. But now we're starting to play into the dataarea. So in France, do you plan on doing any more analysis of datacollecting additional data, looking at expanding this data like, yeah, we someother teams that already doing some of this and we will continue to push toroll this out. We'll continue to kind of marketed and and hopefully kind ofused France is a good example to say, Look at how what they were able tocollect and how this worked. Okay, so let me tell you facts twice today, abusiness analyst, manager, director coming from, of course, anotherdepartment call me and ask me if I can provide him or air more detailedanalysis, or if I can delve into transaction level at purchase tickledevelop mean to increase their analytical, uh, power. Because,historically speaking, our business analysts have used, let's say, excelbased modeling to gain insight from from data. Eso thanks fully advances inin big Data Analytics and concretely speaking with our data platform enablesMcDonald's to reap their data benefits. And we have now means to integrate,analyze our data, the transaction, uh, level to predict the next best productto execute association and the disease to customize so experience of customerto do a lot of stuff as much. You say that just before we have a lot offopportunities to come. So basically the different business lines aware off thisadvantage and they want to You have more data to make more, let's say, moreanalysis and much detailed analysis. So it zvehr e nice to see what he it willcome in the in the month, maybe years to come. That's fantastic. Yeah, I'mvery excited to see where you all take. This is Well, you know, the A IRevolution is coming, and it's it's interesting to see how how you'vealready applied it and it's working. It's really helping the customerexperience, and I'd like Thio see that continue. Well, I want to thank all ofyou for your time today. Adrian, Christina, Matthew fantastic. This hasbeen a really interesting project to get the watch come from conception todeployment and implementation. And I would like to say success so great workout there in the fields with this technology. Okay. I want you to thinkalso, all the teams that helped us during this major project and helpersmay make it a success. Obviously, uh, Christina, for global technology atMcDonnell's, then but also all the...

...second watch people that worked with us.It's been really, really important to have you alongside with us, and youhelp us make this a real success. So thanks. Thanks to all of you. Thank you.I would also say thank you. I think second watchers have been a really,really good partner, and I'm looking forward to doing more of this with morepeople. Well, thanks for those kind words. We don't often try toe selfpromote too much on this podcast. But hey, we'll take it, uh, at the end of2020. Thank you again, everybody. Thanks. Thanks. Thank you again.Appreciate your time audience. Thanks again. We really enjoyed this episode.I hope you do too. As always, please. Email is that cloud crunch at secondwatch dot com with any questions, comments or suggestions. Thanks again, you've been listening to Cloud Crunchwith Ian Willoughby and Skip Very. For more information, check out the blogged.Second watch dot com slash company slash blogged or reach out to secondwatch on Twitter.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (30)