>> Hello everyone. And thanks for joining us today. My name is Brent Biddulph, managing director retail, consumer goods here at Cloudera. Cloudera is very proud to be partnering with companies like 3Soft to provide data and analytic capabilities for over 200 retailers across the world and understanding why demand forecasting could be considered the heartbeat of retail. And what's at stake is really no mystery to most retailers. And really just a quick level set before handing this over to my good friend, Kamil at 3Soft. IDC, Gartner, many other analysts kind of summed up an average here that I thought would be important to share just to level set the importance of demand forecasting in retail, and what's at stake, meaning the combined business value for retailers leveraging AI and IOT. So this is above and beyond what demand forecasting has been in the past, is a $371 billion opportunity. And what's critically important to understand about demand forecasting is it directly impacts both the top line and the bottom line of retail. So how does it affect the top line? Retailers that leverage AI and IOT for demand forecasting are seeing average revenue increases of 2% and think of that as addressing the in stock or out of stock issue in retail and retail is become much more complex now, and that it's no longer just brick and mortar, of course, but it's fulfillment centers driven by e-commerce. So inventory is now having to be spread over multiple channels. Being able to leverage AI and IOT is driving 2% average revenue increases. Now, if you think about the size of most retailers or the average retailer that, on its face is worth millions of dollars of improvement for any individual retailer. On top of that is balancing your inventory, getting the right product in the right place, and having productive inventory. And that is the bottom line. So the average inventory reduction, leveraging AI and IOT as the analysts have found, and frankly, having spent time in this space myself in the past a 15% average inventory reduction is significant for retailers, not being overstocked on product in the wrong place at the wrong time. And it touches everything from replenishment to out-of-stocks, labor planning, and customer engagement. For purposes of today's conversation, we're going to focus on inventory and inventory optimization and reducing out-of-stocks. And of course, even small incremental improvements. I mentioned before in demand forecast accuracy have millions of dollars of direct business impact, especially when it comes to inventory optimization. Okay. So without further ado, I would like to now introduce Dr. Kamil Volker to share with you what his team has been up to, and some of the amazing things are driving at top retailers today. So over to you, Kamil. >> I'm happy to be here and I'm happy to speak to you about what we deliver to our customers, but let me first introduce 3Soft. We are a 100 person company based in Europe, in Southern Poland, and we, with 18 years of experience specialized in providing what we call a data driven business approach to our customers. Our roots are in the solutions in the services. We originally started as a software house. And on top of that, we build our solutions. We've been automation that you get the software for biggest enterprises in Poland, further, we understood the meaning of data and data management and how it can be translated into business profits. Adding artificial intelligence on top of that makes our solutions portfolio holistic, which enables us to realize very complex projects, which leverage all of those three pillars of our business. However, in the recent time, we also understood the services is something which only the best and biggest companies can afford at scale. And we believe that the future of retail demand forecasting is in the product solutions. So that's why we created Occubee, our AI platform for data driven retail that also covers this area that we talked about today. I'm personally proud to be responsible for our technology partnerships with Cloudera and Microsoft. It's a great pleasure to work with such great companies and to be able to deliver the solutions to our customers together based on a common trust and understanding of the business, which cumulates at customer success at the end. So why should we analyze data at retail? Why is it so important? It's kind of obvious that there is a lot of potential in the data per se, but also understanding the different areas where it can be used in retail is very important. We believe that thanks to using data, it's basically easier to derive the good decisions for the business based on the facts and not intuition anymore. Those four areas that we observed in retail, our online data analysis, that's the fastest growing sector, let's say for those data analytics services, which is of course based on the econ and online channels, availability to the customer. Pandemic only speeds up this process of engagement of the customers in that channel, of course, but traditional offline, let's say brick and mortar shops. They still play the biggest role for most of the retailers, especially from the FMCG sector. However, it's also very important to remember that there is plenty of business related questions that need to be answered from the headquarter perspective. So is it actually good idea to open a store in a certain place? Is it a good idea to optimize a stock in a certain producer? Is it a good idea to allocate the goods to online channel in specific way, those kinds of questions, they need to be answered in retail every day. And with that massive amount of factors coming into the equation, it's really not that easy to base only on the integration and expert knowledge. Of course, as Brent mentioned at the beginning, the supply chain and everything who's relates to that is also super important. We observe our customers to seek for the huge improvements in the revenue, just from that one single area as well. So let me present you a case study of one of our solutions, and that was the lever to a leading global grocery retailer. The project started with the challenge set of challenges that we had to conquer. And of course the most important was how to limit overstocks and out of stocks. That's like the holy grail in retail, of course, how to do it without flooding the stores with the goods. And in the same time, how to avoid empty shelves. From the perspective of the customer, it was obvious that we need to provide a very well, a very high quality of sales forecast to be able to ask for what will be the actual sales of the individual product in each store every day, considering huge role of the perishable goods in the specific grocery retailer, it was a huge challenge to provide a solution that was able to analyze and provide meaningful information about what's there in the sales data and the other factors we analyzed on daily basis at scale, however, our holistic approach implementing AI with data management background and these automation solutions all together created a platform that was able to significantly increase the sales for our customer just by minimizing out of stocks. In the same time, we managed to not overflood the stock, the shops with the goods, which actually decreased losses significantly, especially on the fresh fruit. Having said that, these results, of course translate into the increase in revenue, which can be calculated in hundreds of millions of dollars per year. So how the solution actually works? Well in its principle, it's quite simple. We just collect the data. We do it online, we put that in our data, like based on the cloud, through other technology, we implement our artificial intelligence models on top of it. And then based on the aggregated information, we create the forecast and we do it every day or every night for every single product in every single store. This information is sent to the warehouses and then the automated replenishment based on the forecast is on the way. The huge and most important aspect of that is the use of the good tools to do the right job. Having said that, you can be sure that there is too many information in this data. And there is actually two-minute forecast created every night than any expert could ever check. This means our solution needs to be very robust. It needs to provide information with high quality and high veracity. There is plenty of different business process, which is based on our forecast, which need to be delivered on time for every product in each individual shop. Observing the success of this project and having the huge market potential in mind, we decided to create our Occubee, which can be used by many retailers who don't want to create a dedicated software that will be solving this kind of problem. Occubee is our software service offering, which is enabling retailers to go data driven path management. We create Occubee with retailers for retailers, implementing artificial intelligence on top of data science models created by our experts. Having data analysis in place based on data management tools that we use, we've written first attitude. The uncertain times of pandemic clearly shows that it's very important to apply correction factors, which are sometimes required because we need to respond quickly to the changes in the sales characteristics. That's why Occubee is open box solution, which means that you basically can implement that in your organization, without changing the process internally. It's all about mapping your process into the system, not the other way around. The fast trends and products collection possibilities allow the retailers to react to any changes, which occur in the sales every day. Also, it's worth to mention that really it's not only FMCG and we believe that different use cases, which we observe in fashion, health and beauty, home and garden, pharmacies, and electronics, flavors of retail are also very meaningful. They also have one common thread. That's the growing importance of e-commerce. That's why we didn't want to leave that aside of Occubee. And we made everything we can to implement a solution, which covers all the needs. When you think about the factors that affect sales, there is actually huge variety of data that we can analyze. Of course, the transactional data that every dealer possesses, like sales data from sale from stores, from e-commerce channel, also averaging numbers from weeks, months, and years makes sense, but it's also worth to mention that using the right tool that allows you to collect that data from also internal and external sources makes perfect sense for retail. It's very hard to imagine a competitive retailer that is not analyzing the competitor's activity, changes in weather or information about some seasonal stores, which can be very important during the summer and other holidays, for example. But on the other hand, having this information in one place makes the actual benefit and environment for the customer. Demand forecasting seems to be like the most important and promising use case. We can talk about when I think about retail, but it's also the whole process of replenishment that can cover with different sets of machine learning models, and data management tools. We believe that analyzing data from different parts of the retail replenishment process can be achieved with implementing a data management solution based on Cloudera products and with adding some AI on top of it, it makes perfect sense to focus on not only demand forecasting, but also further use cases down the line. When it comes to the actual benefits from implementing solutions for demand management, we believe it's really important to analyze them holistically first it's of course, out of stock minimization, which can be provided by simply better size focus, but also reducing overstocks by better inventory management can be achieved by us in the same time. Having said that, we believe that analyzing data without any specific new equipment required in point of sales is the low hanging fruit that can be easily achieved in almost every industry, in almost every regular customer. >> Hey, thanks, Kamil. Having worked with retailers in this space for a couple of decades, myself, I was really impressed by a couple of things and they might've been understated, frankly, the results of course. I mean, as I kind of set up this session, you doubled the numbers on the statistics that the analysts found. So obviously in customers, you're working with... you're doubling average numbers that the industry overall is having, and most notably how the use of AI or Occubee has automated so many manual tasks of the past, like tour tuning, item profiles, adding new items, et cetera, and also how quickly it felt like, and this is my core question. Your team can cover or provide the solution to not only core center store, for example, in grocery, but you're covering fresh products. And frankly, there are solutions out on the market today that only focus on center store non-perishable departments. I was really impressed by the coverage that you're able to provide as well. So can you articulate kind of what it takes to get up and running and your overall process to roll out the solution? I feel like based on what you talked about and how you were approaching this in leveraging AI, that you're streamlining processes of legacy, demand, forecasting solutions that required more manual intervention, how quickly can you get people set up? And what is the overall process of like to get started with this software? >> Yeah, usually, it takes three to six months to onboard a new customer to that kind of solution. And frankly, it depends on the data that the customer has. Usually it's different for smaller, bigger companies, of course, but we believe that it's very important to start with a good foundation. The platform needs to be there, the platform that is able to basically analyze or process different types of data, structured, unstructured, internal, external, and so on. But when you have this platform set is all about starting ingesting data there. And usually for a smaller companies, it's easier to start with those, let's say, low hanging fruits. So the internal data, which is there, this data has the highest veracity. It's all really easy to start with, to work with them because everyone in the organization understands this data. For the bigger companies it might be important to ingest also kind of more unstructured data, some kind of external data that need to be acquired. So that may influence the length of the process. But we usually start with the customers with workshops. That's very important to understand the reasons because not every deal is the same. Of course, we believe that the success of our customers comes also due to the fact that we train those models, those AI models individually to the needs of our customers. >> Totally understand. And POS data, every retailer has right in, in one way shape or form. And it is the fundamental data point, whether it's e-comm or the brick and mortar data, every retailer has that data. So, that totally makes sense. But what you just described was months, there are legacy and other solutions out there, that this could be a year or longer process to roll out to the number of stores, for example, that you're scaling to. So that's highly impressive. And my guess is a lot of the barriers that have been knocked down with your solution are the fact that you're running this in the cloud. from a compute standpoint on Cloudera from a public cloud stamp point on Microsoft. So there's no IT intervention, if you will, or hurdles in preparation to get the database set up and all of the work. I would imagine that part of the time savings to getting started, would that be an accurate description? >> Yeah, absolutely. In the same time, this actually lowering the business risks because we see the same data and put that into the data lake, which is in the cloud. We did not interfere with the existing processes, which are processing this data in the combined. So we just use the same data. We just already in the company, we ask some external data if needed, but it's all aside of the current customers infrastructure. So this is also a huge gain, as you said. >> Right. And you're meeting customers where they are, right? So as I said, foundationally, every retailer POS data, if they want to add weather data or calendar event data, or, one incorporated course online data with offline data, you have a roadmap and the ability to do that. So it is a building block process. So getting started with core data as with POS online or offline is the foundational component, which obviously you're very good at. And then having that ability to then incorporate other data sets is critically important because that just improves demand forecast accuracy, right. By being able to pull in those, those other data sources, if you will. So Kamil, I just have one final question for you. There are plenty of... not plenty, but I mean, there's enough demand forecasting solutions out on the market today for retailers. One of the things that really caught my eye, especially being a former retailer and talking with retailers was the fact that you're promoting an open box solution. And that is a key challenge for a lot of retailers that have seen black box solutions come and go. And especially in this space where you really need direct input from the customer to continue to fine tune and improve forecast accuracy. Could you give just a little bit more of a description or response to your approach to open box versus black box? >> Yeah, of course. So, we've seen in the past the failures of the projects based on the black box approach, and we believe that this is not the way to go, especially with this kind of, let's say specialized services that we provide in meaning of understanding the customer's business first and then applying the solution, because what stands behind our concept in Occubee is the, basically your process in the organization as a retailer, they have been optimized for years already. That's where retailers put their focus for many years. We don't want to change that. We are not able to optimize it properly for sure as IT combined, we are able to provide you a tool which can then be used for mapping those very well optimized process and not to change them. That's our idea. And the open box means that in every process that you will map in the solution, you can then in real time monitor the execution of those processes and see what is the result of every step. That way, we create truly explainable experience for our customers, then can easily go for the whole process and see how the forecast was calculated. And what is the reason for a specific number to be there at the end of the day? >> I think that is invaluable. (indistinct) I really think that is a differentiator and what 3Soft is bringing to market. With that, thanks everyone for joining us today. Let's stay in touch. I want to make sure to leave Kamil's information here. So reach out to him directly, or feel free at any point in time obviously to reach out to me. Again, so glad everyone was able to join today, look forward to talking to you soon.