A real-world IoT case study for peak success with less risk
Good morning and welcome to our webinar, How the World’s Largest Hotel Chain Unlocked Profits. My name is Sarah-Lynn Brunner and I will be your host for today.
Today’s presenter is Manish Mistry. Manish, our resident mobile IoT guru brings in 20 plus years of technology and management experience to Apexon. His extensive background delivering and managing all phases of mobile and IoT development testing and certification enables him to create solutions for our customers that help them launch products, services, and content into the mobile IoT arena.
Let’s welcome Manish.
Manish: Good morning folks. This is Manish Mistry from Apexon. It’s a very sunny morning from California. It’s live right now. Let’s talk about one use case.
Sarah-Lynn: Great. Thank You, Manish. Before we get started though, we’re going to talk about the housekeeping item.
First, this webcast is being recorded and will be distributed via email to you, allowing you to share it with your internal teams or watch it again later. Second, your line is currently muted. Third, please feel free to submit any questions during the call by utilizing the chat function on the bottom of your screen. We’ll answer all questions towards the end of the presentation. We will do our best to keep this webinar to the 30 minute time allotment.
At this time, we’re going to hand it over to you Manish.
Manish: Let’s talk about one of the large use case which we install for big industry like hotel, where we have extensively used IoT to generate new mechanisms of revenue for that industry, as well as how we can use technologies to give one of the best user experience and customer experience from the hotel industry side so that anybody who has used hotel knows how painful it is and how seamlessly we can make it happen.
I assume everyone has gone through good or bad scenarios checked into hotels, if anyhow ask for some help. Now, this industry is poised to change the whole ecosystem, how we use technology, how we can use technology to make and simplify the use cases which are of very simple in nature, but always complex to implement.
What I would like to talk about is couple of things before we start. Imagine a hotel industry having multiple branches or their buildings are in various locations, each hotel having multiple rooms. Whenever we think of a solution, we have to think of the scalability of it. At the same time, when we talk about user experience, the people could be of various nature, from various regions, so we have to, also, keep that in mind. When we think from that perspective, the whole dynamics of doing a proof of concepts to the enterprise-grade solution or industrial solution you have to think of so that robustness plays a key role in it, at the same time, how one can come out of the lab mode to a production mode also changes.
To keep that in mind, it’s very important that how we deal in designing a solution at the same time how we can see to it that we can scale, maintain support, and monitor on a constant basis. There are two types of scenarios which are happening. One is behind the scene, what we call it, how optimize we can do the logistics of the hotel industry. On the other hand, there is a front-end which is what we call the user experience which is impacting a person who is the customer and trying to do the– I would say from check-in perspective till day-to-day activity which you would like to see as a user, how seamless that experience becomes.
Let’s look at first the front-end or the user experience scenario. There will be separate webinar for the back-end system because that’s also a very complex scenario which we’d like to address. Let’s, today, look from the front-end perspective. The way we try to solve this is… There are problems when you log in, when you check-in. When you check-in in any hotel, I don’t think I’ve seen any place, I’ve seen a seamless experience where user is able to check-in without having some of the other problems.
Traditional check-in systems have not worked. The whole industry is trying to say, “How can I identify the person? How can I give the person a seamless experience, that I know that person is entering my hotel?” As soon as the person enters the hotel, how can he or she would get automatic check-in? They can go straight to their rooms because they have the room in their hand which is their mobile phone which is being the hubber. At the same time, everyone must have experienced that sometimes your room keys first they don’t work, the air condition temperature is not correct the way you wanted, at the same time everything is around, “You take something and you pay for it.” Then those experiences are not a great experience.
What we try to do as a team together is, of course, we have done tremendous amount of research and taken tons of interviews what would work, what would not. From that, we have tried to come up with how we can create a wallet system which we can– Your phone being your hub which will help you create a centralized experience, which is in your hand, to tackle these situations.
Now, of course, there are many different types of technologies available. There are different challenges. Let’s say how can you open a lock with your phone? After that, so another one. How can you use your phone as a payment machine? How can you use your phone as a controller for air condition, or heater for that matter? There are many different solutions available in the market which we are to either build it from scratch or take off the shelf and see whether that solution we can scale it to the large scale we require.
If you look at a solution, multiple things we want to tackle from one simple, your mobile phone, from access control till your smart room control, till how can you do the payment very seamlessly, how can you do check-in into the hotel seamlessly. There are multiple ways we have tried this. We have tried to do you are check-in on your own or check-out on your own from your phone, but we are, also, trying to put smart concept. As soon as you are within the location of the hotel, we know who you are in case you have registered with us at various avenues. We can find out who you are, what are your preferences. At the same time we can give you and do automatic check-in for you based on your past history.
Believe it or not, everyone has their own patterns. This smart check-in process identifies your pattern based on which it actually keeps your room ready. For example, you like to have the temperature at, let’s say, 70 degrees. As soon as you check-in into the hotel, the room temperature becomes or it, at least, tries to maintain to 70 degrees so that as soon as you enter the room it becomes the right a-ha moment you can get into the room. It actually will turn on the favorite music. Those concepts are already in place. We have already done this. Custom experience for user is extremely important. When we have created his wallet ecosystem within from your phone, we try to address most of the scenarios.
Now, on the other hand, many times rental car is also an important aspect when you are trying to use hotel. Many places there are on-site rental places which is already taken care of- at a push of a button on your phone. You can actually rent a car and it actually gets ready as soon as we know from your pattern or you already when it can be ready.
Looking from the wallet ecosystem perspective, we try to merge right user experience, from access control, till check-in, till rental car, till on-premise shopping, to your in-room shopping. On an average, every room has close to 20 to 30 different sensors which actually is making these things happen. Now, integrating all these things together is not easy. Some are off the shelf solutions, at the same time, we had to create on our own. The philosophy we used is try fast, fail fast. You try it out, any new use cases, see to it that you put in the different parameters and see whether that use case is scalable. If it doesn’t, then we try another system or another mechanism and see whether that can be scalable or not.
Next, I would like to see what are the different activities we had to do. What we did is, we tried to match this to the network protocol seven layers we have. The nature of the work is, on an average, around two years life-cycle. Now, we have to put some methodology to this madness. How do we make it happen? What we tried to do is, from first layer, which is the sensors layer, we tried to identify many different types of sensors available in the market. Of course, that is not one area anyone want’s to go if you want to make it happen on your own. Of course, we had to do tons of industrial sensors as well, at the same time, we had to do the normal typical sensors, but they had to be industry grade, at the same time, can we scale it?
I’ll give you an example why the whole system had to be scalable is important. When we are talking of hundreds of hotels, and hundreds of room per hotel, the extrapolation of problems can increase dramatically. As we say, a high-tech world has high-tech problem. There’s always going to be some problem which we’ll have to see to it, can we monitor it, can we support it, at the same time, can we tackle it remotely? When we looked at those perspectives, we had to decide, “Will my solution, which is available in the market, is going to cater to that. If not, do I have to build it?”
In the second layer, that becomes a very, very important question which we are to answer, where some of the things we had to build on our own because there was no solution outside which would cater to all the requirements we had to put up. Now, saying that will my system become scalable by the nature of doing on your own, it means we had to involve additional external vendors who can build it for us based on our design. That brought another complexity to the whole process.
Now, after this, of course, as you know, in most of the IoT world, there are different protocols on which all these systems are working. The number three and number four layers become very important, the interaction layer and the communication layer, where we had to bring and bridge the gap of different protocols, complexity along with their payloads. That becomes another challenge which we had to solve it as an implementation so we went by the whole platform approach. At the same time, we went to a little bit of a ready-made solutions which helped us craft this together.
Now, on top of that, how we can do a security on top of it, how we can do a solutions on top of it, how we can do a smart use case on top of it. From that perspective, number five, number six, and number seven are the layers which we had to address and see how the use cases map from layer one to layer seven in entirety rather than just looking at application layer or just at the platform layer. When you create a methodology like this as a layer, it helps us guiding every sprint, every implementation in the right fashion, and it creates a whole thought process around thinking into it.
There are two different lifecycles we had to go through in parallel. One which was the whole engineering-based embedded product development lifecycle. That had a different stream of work. At the same time, on the other side of the world we had a system development lifecycle. Now, when you have two different streams going on and as a large-scale use case which is transpiring from both the streams, the product management, as well as a project management was a mammoth task having multiple vendors also. Combining them together is one of the challenges which we have in the past have tackled it seamlessly to make the project very successful.
Let’s look at the activity matrix. What are the things we had to do from embedded development lifecycle until software development lifecycle in entirety? Of course, we had to work with business to identify the scenario, not just with business, at the same time we worked with end-users. What would they like to see? What would they like to experience? When we are going from traditional model to the new generation model, it becomes extremely hard to find out what would help people get the best experience especially when they’ve never experienced it.
Some are table stakes, for example, can I have voice-activated temperature control that becomes a table stake. Nowadays, having Alexa and other systems are so powerful. At the same time, the new generation concept which people may like and may not like. For example, we have a smart mirror concept within the hotel where one can go and see their alerts when they are in the bathroom getting ready for their work or whatever the scenario. Some people like those kinds of concepts, some people don’t like those concepts.
When you want to give an experience, you have to give them complete flexibility whether they would want to opt-in or they don’t want to opt-in, they want to share, they don’t want to share. Those kinds of concepts which becomes part of it and when we are defining functional requirements, it transpired into the architecture how we want to tackle those scenarios.
Device functional requirements are generated from this which are of end-to-end in nature not just doing the, “Yes, the POC works.” As I mentioned, can I monitor them remotely? Do I know if they are failing? If you had to maintain hundreds of thousands of sensors in different hotels, it becomes a mammoth task as a maintenance itself. We went to the hardware specification from that specification we had to keep that in mind.
Do I have capability of early alerts in case things are not working? Do I have a capacity to put futuristic self-healing concept? In case the hardware is powerful enough, can I scale and change our firmware remotely? A lot of people must be knowing that now Tesla does that where you can change your car software remotely. Similarly, here we have to bring those concepts while doing this solution.
When we went through the whole product design, which is more user-centric experience, how can we think of robustness? How can we see to it that they are practically non-existent to the end-user? It has to become a seamless experience. It cannot be that I have to do many steps or many things to activate that function. To give you an example, I am entering at a hotel can it check-in on its own? If I had opt-in for it, I don’t have to do anything, the system should recognize me, the system should actually do everything for me and give me a seamless experience. Multiple technologies would be used to design this, at the same time implement it, but it is extremely a successful scenario if you opt-in to give a seamless check-in experience.
Similarly, other types of scenarios are, “This is a large hotel complex. I am at this area and I want food, can somebody deliver me the food or do I have to go to the restaurant in the hotel?” That concept is also implemented where you can be anywhere and food can be delivered to you. This can be multiple technology-based solution where we have to use Beacon wireless functions in identifying where you are with complex algorithms, but today, that is possible. Now, these are the types of concepts and designs we had to bring in together as part of this.
Now, second phase immediately was, how do we do prototyping? This is where we as a team came with a mini hotel in our office, we call it the lab, where we implemented most of the scenario within our lab. We call it a lab as a service concept where we brought all the things together and see to it that we get a seamless experience as what an end-user would like to have it within our lab. There are multiple iterations by this whole development happen including bringing the right customer base and try it out. At the same time constant changing and managing the churn of technology upgrades we had to go through to make it happen.
Second phase, of course, as we upgraded from prototyping we created different streams of development and integration. Every hotel brings tons of different systems which we had to integrate so we had to create a nice adaptor patterns where we could integrate them seamlessly and much faster. Most of the hotel industry has standardized their own part of how we can bring a hotel, all the systems together but each hotel has its own evolution also. It becomes complex scenario where we have to integrate them, especially when we are forward-looking of new generation scenarios and see how it impacts their back-end systems also.
That, again, created a very different outcome as a project where we had to create multiple streams just purely functioning towards integration. Integration happen from multiple sides, from device till how we can implement in the hotel, till integration of back-end systems, till new generation things like wise enable, mirror-based scenarios retail mobile-based wallet.
The third one is very important part we had to go through. To give a right experience, from the beginning, we had to do automation, in such that we can implement and do the automation. What we implemented is the in-sprint automation for most of our use cases. Hardware had its own lifecycle but on the software side and integration side we had to create in-sprint automation to see to it that as the things are changing and progressing, as the user stories are getting implemented, they are getting tested thoroughly. If one of the PC is not working, it doesn’t help in end-to-end experience so the robustness of the system, we had to look, go through as if we are creating a product rather than a project. That’s a different mindset we had to bring in to make this happen.
Different types of testing, of course, everybody knows that we have to do it. It’s such a complex scenario. The biggest challenge in testing was doing an actual implementation in a real-life hotel and make it happen. Then you have to see to it that you test out in conjunction with customer to see to it that they are actually experiencing what we had thought of or we ought to yet design for. From the testing part and verification part it becomes very important that we not we do not forget the customer thought process insight.
Now, last one is where the scale happens. You deployed at one place. How do you deploy it again and again in multiple places? How do you centrally control this from your central command for all the hotels? Do you keep them in isolation or do you do it on your own? There are multiple thought process which we had to go through that each hotel has its own way of controlling it or each hotel participates in our centralized solution and centralized solution monitors them, manages them and controls everything. Both the models were implemented so that we can bring the scalability.
As any lifecycle of technology, we’ll find issues, we find defects, and we fix them, we push them. On the back-end side, we push them on the sensors of the world, we push them on the firmness of the world, and of course the application of the world. We choose a standard lifecycle we have set up right now and is going on. I would say next six months you will start as a user you’ll start seeing this difference. You will start seeing that how the new generation experience will create a real a-ha situation for any customer. Again, this is from the front-end perspective or the user experience perspective for any users. When we look at the behind the scenes, there is much more complexity by which we have use IoT and we’ll talk about that in a next webinar.
Now, were there any hurdles while doing implementation of this complex nature? Of course. There are multiple teams, multiple partners ecosystem which was involved in it. Some were ready-made solutions which had to be customized, some were just pure technology which we had to bring it, some we had to do on our own, multiple streams of devices, our development, some multiple streams of software side of the world, integration side of the world.
Project governance became one of the biggest challenge. We knew that from day one that it is going to be complex from implementation perspective but it’s very hard to estimate how many partner challenges you are going to get when you have this multi-partners and other. At any point of time more than 20 to 30 partners were working on this project so you can imagine the complexity of it.
Second biggest challenge is everyone has a say in any of this new generation experience. From your internal management till customer management everyone has a say how to do it, what to do it, what would they like to have it. How do you manage your requirements? That became another challenge because everyone thinks experience differently, everyone thinks that they have an idea, and everyone thinks they are the expert.
What we had to do is we have to go through the concept of customer first vision. Whatever we could be passionate about as a technologist we had to see what would customer like and what would take an a-ha moment about. We put those customer first vision first and actually made the whole customer centricity oriented system, which is the one concept I would like to push to everyone, has worked beautifully.
Security. Everyone knows that security becomes the biggest concern when there’s IoT. That was one of the drivers that some of the solutions or the hardware we had to implement was on our own, on top of which we manage the whole security layer which were industry-graded solutions rather than just picking any hardware out of the box. It becomes very hard if where this device has come from, what is the security layer on top of it which they implemented. When we go through a whole governance model of this, we would like to see to it that the security becomes number one priority and see to it that there are no issues in case there are scenarios which are security breach-oriented.
One big challenge, of course, were interoperability and integrations with different systems. Having different protocols, having different ways to communicate became a big challenge but that is where the expertise of Apexon has worked beautifully to solve those issues. There were some out-of-the-box solutions, at the same time there were homegrown protocols we had to implement on our own to solve this problem.
What happens is that because we had multiple data sources which are producing data, for example, each sensor produces its own data, you have your back-end systems which have tons of data, how do you manage them? How do you aggregate them? How do you make a sense out of them as a business? That was another challenge or a good challenge which emerged out of it.
On top of that when you put a layer of rooms, it has its own data and people have their own data as a person, and sensors are producing data. Now, one layer above each hotel is producing these data and then when you aggregate everything it becomes a very different beast altogether which you have to manage it, but that gives so much of intelligence to you that you can put a real-time machine learning algorithms, at the same time a trending algorithms on top of it to find out what are the trends happening, what are the ways people are using our system. The same time one can give real new generation business aspect that what can I cross-sell on top of it. Dependency remains the same that if data became huge, how do you test remains the biggest problem.
By saying this I would like to conclude that it’s good to see that new generation technologies are now getting adopted by the mainstream industries. They are all going towards how we can change the user experience. I would say that in next six months you will start seeing this transformation and you will experience on your own.
Sarah-Lynn: Great. Thanks so much, Manish. All right. At this time we would like to get your questions answered. As a reminder, you can at any time ask the question utilizing the chat feature at the bottom of your screen. Manish, we do have a couple questions for you that came in. First one, what were the decision drivers for in-house first ready solution?
Manish: This is a typical scenario which an old industry, or I would say, a traditional industry has to answer themselves, “Do I want to build something on my own?” or, “Should I buy anything ready-made?” It’s a build versus buy question has happened for most of it. Now, having those visions from top down helps a lot because do I want to create IP? Can I sell it to different industries of similar nature? Can I generate separate business out of it? When we start this project those drivers actually help us that whenever you had a chance to develop on your own, develop it. In case it’s a mammoth task and ready solution, which is meeting your requirements. For example, is it scalable? Can I monitor it? Can I support it remotely? Is it secure? All those answers are, if it managed, we would go towards a ready solutions, but whenever we thought that any of those criteria are falling short, we will go after our own implementation. That helps to distinguish as a methodology from day one that any new solution, any new feature go through this process, then you will have an easier decision whether you want to buy it or build it.
Sarah-Lynn: Great. Thank you. How did you manage to test this large solution end-to-end?
Manish: As I mentioned, it’s one of the largest complex problem. We had to literally implement a lab within our organization to mimic a hotel, to bring in various systems together, create a whole experience as if you’re in a hotel experience, bring in the right people to test it out. That’s one side of the world. The other side of the world is automate, as much as possible, automate. Automation has helped us a lot in bringing the in-sprint automation concept, has shortened our SDLC lifecycle, at the same time, it has given us predictable that any changes are happening in the system we know they are working fine. Bringing automation and creating lab has been the one of the big driving factor in creating and solving this end-to-end testing of this such a large solution.
Sarah-Lynn: Thanks so much. I think we have time for one more question. Can you give some use cases which are solved for the behind the scene?
Manish: I knew this question would come because we have not touched upon that subject yet, but I’ll still give some glimpse of it. Behind the scene, a lot of complex logistics goes after. For example, do we have the inventory of the items we require? How do you check that? You want to create those inventory-based system very smart. Food itself is a bit logistics which hotel industry goes through where, “Do I have enough food to cater to enough people?” Do I have raw material for that? How do I track how much? Do I have enough raw material or not?”
Hotel industry has many items which are floating around from room to room, to place to place, to conference room to conference room, so that brings into- especially the expensive items becomes a large or mammoth task for them to track, where they are, and who is using it. We’ll put smart systems around it, on top of each and every items to track them. Staff management is another. It becomes another challenge, where staff is constantly on the move. Now, how do you track who’s where, who’s the nearest to answer or take care of the customer needs? Those smart concepts together has given a tremendous amount of good experience which are behind the scenes as what we call it. It has helped give the customer the ultimate a-ha situation.
Sarah-Lynn: Great. Thanks so much, Manish. I believe that’s all the questions we are able to answer at this time.
Also, be sure to check out and subscribe to DTV, it’s a new digital transformation channel that brings in new industry experts. Again, many thanks for our speaker, Manish, and thank you, everyone, for joining us today. You’ll receive an email in the next 24 hours of the slide presentation and a link to the webcast replaying. If you have any questions, please contact us at firstname.lastname@example.org, or call 1-408-727-1100 to speak with a representative. Thank you all. Enjoy the rest of your day.