Advanced Analytics For Fashion | SAP and Stylumia Possibilities

As a part of SAP.iO Foundry, Berlin data and analytics program, SAP Senior Director Jason Boyer and Ganesh Subramanian, Founder & CEO of Stylumia came together on a webinar on SAP + Stylumia joint advanced analytics possibilities for brands and retailers in fashion. Here are the full recording and transcripts.

Here is the full transcript of the advanced analytics webinar.

Advanced Analytics Webinar – Introduction

Azadeh:
Okay. Hello, everyone. Good morning. Good afternoon. Good evening, wherever you are. First of all, I want to welcome everyone to this call, to this webinar today. I want to thank the Stylumia team, Ganesh, in person for attending the webinar and being part of this program. Definitely, last but not least, Jason, I would really love to thank you for all your support during the course of the data and analytics programs and also for supporting our startups in the past months. And specifically working with Stylumia and also accepting to be part of our webinar today.

For the audience here, Jason is a senior director, solution management at Retail IBU at SAP. To give a little bit of background, today, we would like to continue the consumer journey that we have been discussing with you in the past week. Basically, we have shared with you already two parts with it. Today is the second part of that story that we have created for you. In yesterday’s webinar, we had a look at the consumer’s versus step, which was the development of the need and forming of the need. Our startup NWL from New York spoke about how to best identify and manage micro trends ahead of time.

The second step of a consumer journey will be natural to follow that trend. For that, we looked at the fashion industry. For the fashion industry, we are with Stylumia today, one of our startups from Bangalore, which helps companies to solve some of the major challenges they are facing, dealing with fashion trends today. Stylumia can help them stay ahead of the curve. Before starting the discussion, I would want to say a couple of words on the housekeeping rules. There is a Q&A section down there.

Also, if you have any questions during the webinar, feel free to type it there, or just simply in the chat. We make sure that we take all your questions. The session is being recorded. We will be sharing the recording with you with tomorrow’s report. Without further ado, I think I will just give the opportunity to Ganesh to introduce himself and then Jason, so we can start the conversation there.

Ganesh:
Thanks, Azadeh, and thanks for this wonderful opportunity. We had the privilege of being part of the SAP.iO data and analytics program. Quick introduction about myself. Ganesh Subramanian, founder and CEO of Stylumia. A fair amount of our team is watching this. I’ve been in the fashion lifestyle industry for over 20 years. It’s not that we are solving the problem outside in. We have gone through the pain of taking decisions in the fashion industry and I would say, one of the most challenging industries to manage. With my stint across offline and online, we’re using data scientists. We are using data science to solve the fundamental challenge in the fashion business, and some of that we are showing that thing.

Azadeh:
Okay. Thank you. Jason, you may want to…

Jason:
Yeah, thank you. Appreciate that introduction. I am part of our SAP solution management team with a focus on retail and fashion, and that means go to market. That means engaging with our customers, our partners, Stylumia, aligning with development and branding, bringing products to market. My focus is on planning.

Azadeh:
Okay, Great. I think then you’re the first person to give us some perspective on your areas of expertise in retail and fashion when it comes to using external data. I mean, what are either challenges that our customers or fashion companies are facing today, especially with the new changes and the corona situation?

Jason:
Yeah, so I would say, especially in fashion, planning has very much been for a long time, more of an art. What you have now, when you have lots of data is more and more fashion companies and retailers wanting to take advantage of that data and use it to make better decisions. And so, it’s balancing that art and science. Part of the story today is how SAP and Stylumia are coming together to really provide the science to support the art of planning and to really help companies make better decisions about their short mints and how much to buy.

Fashion planning, fashion forecasting is really, really tough. Your preseason, you’re maybe introducing many products that don’t have a lot of history and that could be really hard to do and to do well. And if you don’t even have the right tools, it’s even harder. You’ll find that many companies rely on spreadsheets, old tools. The data, the organization is very siloed and not well integrated. Very manual and inefficient people are working on more manual tasks rather than higher-value tasks and finding out insights that they can use to make better decisions.

Oftentimes, they lack the use of predictive science, or artificial intelligence to support those decisions, and this can lead to less relevant assortments. It can lead to underbuying where you miss revenue opportunities. It can also lead to underbuying where you create a lot of waste and you also increase your costs. The idea here is to really improve that and use the data that you have and bring in these new, additional forms of data that Stylumia is going to talk about today to really create relevant assortments, make better buying decisions and get the right products in the right store so you have happy customers. That’s the opportunity. That’s what we’re going to talk about today. I’ll turn the time back over to Stylumia.

Ganesh:
Thanks, Jason.

Azadeh:
Feel free to take it from here and walk us through your solution and we’re all ears.

Stylumia Solutions | Fashion AI and Advanced Analytics

Ganesh:
Wonderful. Thanks, Jason. Great introduction. Let me just take on from there. As I share my presentation and followed by a demo, just want to share with all of you that today’s session will have three parts. The first part, I’m going to talk about the why. There are so many fashion forecasting agencies, so many people today, sharing data. Why another tool? Very, very important. First, let’s get to the why before get into the what. I’m going to quickly take you through that. Just confirms that if the screen is visible.

Azadeh:
Yes, very well.

Ganesh:
Wonderful. Right. Well, we are talking about a paradigm shift. We’re not talking about an incremental shift. Now, what is that paradigm shift? The fashion industry for decades has been… Let’s look at the fashion industry at various spaces. While a lot of customer expectations have changed. Retail operations have changed product and inventory decision-making like what Jason said, hasn’t changed for a long time. We have an opportunity with cutting-edge, DeepTech to look at precision, localization, prediction and automation.

Today, we are going to talk about that, cutting across various points in the fashion supply chain. We have an inventory epidemic today leading to overbuys, underbuys and shrinking gross margins and accentuated by COVID. Now, let’s look at various stages, starting with the design. Designed research has remained the same until now. It’s researching the traditional way, storyboarding, predicting, designing. Today, a lot of data comes from designers, but the question is, what does the data contain?

Now, the data contains a lot of noise. Now, let me ask you this question, what is the trend? A lot of people come with trend reports. What is the trend? Is trend a lot of availability of a particular attribute, a particular category of products? A lot of people searching in Google, is it a trend? Now, the question here is a lot of them has a huge amount of noise or do we need an engine that will remove this noise. Let’s go to buying and merchandising.

As Jason said, this has remained the same for a long time, it’s been over years. What’s the need to change, right? We have huge demand and certainties, inventory challenges, shrinking margin, and lost sales. Now, the role of designers, buyers, and merchandisers are a lot harder than ever before with dynamic changes and a huge amount of data. The last leg is distribution. Here again, we are talking about distribution is [inaudible 00:08:55] physical activity, but also understanding the consumer demand at various levels of the distribution hierarchy, mapping the tastes. Traditional ways of clustering also don’t work well.

Well, what we’re basically saying, we need new thinking and what’s that new thinking is what Stylumia is bringing to the table. And the new thinking has to be holistic. We are looking at a holistic view of starting from consumer insight, which is going to be a large part of today’s discussion, which is getting real-time demand signals. The industry today, focusing on supply signals, while there is a lot of data, but the data only gives you supply signals. We know the amount of restaging over billions of governments are based at every year only because that we don’t look at consumer demand, but we look at the supply side of the business. That’s the first one.

And then moving into the plan, buy to distribution optimization. I’m going to now look at one of the proprietary engines of Stylumia, which runs right through all our solutions, which is predicting consumer demand. Let me just quickly get there. How do we sense demand in the market? Demand is not public data. This is the world’s fast demand sensing engine. What it does is, like Google ranks pages, we rank products all over the world for a global fashion ranking system, which tells you not trends, which tells you winning trends. I just want you to bold and underlined winning trends is what Stylumia is for. If you pick winning trends, the probability of error completely comes down.

Now, with that quick introduction, I’m moving into a demo of our solution. Let’s look at this. We’ll walk through a use case of a customer. Well, let’s assume that this is a brand-based order of the US and Europe and it’s a woman’s wear brand coming into Stylumia. The first thing is I would really like to know what’s happening, right? As you come into the system, you are thrown, what’s new, very relevant to your interest globally. Whether you’re interested in UK, US. Based on your profile, you’re hit with what’s new. It’s very important because that’s where the inspiration starts.

Now, you also get to see what’s trending. What’s trending means what consumers are buying more of. They are liking more of, across the hierarchy of fashion, from luxury to value, from fashion shows to retail multi-brand to mono-brand retail globally. And then personalized recommendations as to the designer for you. Now, this designer… And then say that I would like to understand before I deep dive into my brand, what category is I would like to see high-level insights?

Then she jumps into category level insights in her area of interest, and suddenly she sees in Germany, I see in women’s wear dresses are the top-performing category. It does not supply. You’re talking about a top-selling category mix. Dresses are the top one, followed by t-shirts. She’s also looking at what are the various brands that are doing well in that particular territory. Then she says, “Okay, this is what is happening in the market, but I’m looking for inspiration.” So what she then does is, she gets into her favorite designer in Farfetch, UK. She’s looked for Farfetch, the US and she’s looking at now designers trends and winning trends, right?

These are all the products which are performing really well in that market. Then she comes here and then looks at that particular… She likes this zebra print and she wants to go deep into it. All in few clicks, she knows that it’s doing well, it’s on the top hundred percentile of performance of dresses. In her area of interest, she’s saying that, “I would like to know whether this is a mass product or it’s a niche idea where there is less competition.” But in this particular case, we also provide your with product positioning. It’s performing well, but it’s in the mass market, high demand area.

Now, she will decide whether I want to take it depending on her brand DNA. Once she has done that, she also can look at how has been the price performance of this particular product. She can also look at the availability of sizes. What are the winning sizes? What are the sizes doing at the stock? The sizes I should not miss. Now, she gets that in-app trend and then she moves to say that, “While I like the zebra print, I want some ideas, is zebra print in general, working very well?”

We use our computer vision engine and immediately you’re getting similar zebra prints in the market. Not only that, she looks at which kind of zebra prints are really working well. The first one is doing really well. The second one’s a good seller and she’s also seeing an average seller. She’s just picked up new ideas from the designer space, which is inspirational for her to put into that range.

Now, at this point, the designer is left with, “I’ve got some idea. Now, I want to start generating some designs before that.” There are various options available for you to analyze what are the color trends, the key things, whether it’s fabric trends. What you’re seeing on the screen right now are all the colors that are winning and which are not winning. Each one of them comes with Pantone colors. We have a partnership with Pantone where every single product you get to see what’s the Pantone color here. For very actionable, you’re spotting the winning trends, winning product ideas, similar ideas globally, and ready to execute.

Now, the question you’re left with is, how do I now start using this in creating a design? Here comes a very innovative opportunity for a product that you’re just about to get in the market is we can generate designs using gender adversarial networks. This is one of the solutions all imagining. Now, the design you’re seeing on the screen is generated by the machine with all the ideas that you have got from the platform. Now, she has an opportunity of fine-tuning this, but let’s say we now want a data-generated image, select the attributes of our choice, colors of our choice so you get a design.

Now this design, she has an opportunity to change attributes and convert the way she wants. Now suddenly it has become a long sleeve. Now you want to look at a pattern weightage. You can change the weightage of the pattern. You want to change the sleeve length. Now it has suddenly becomes almost a sleeveless product. You want to change the color weight. Now you’ve changed the color weight of the product or you’re going to the other extreme, now you’ve got another dress.

The point here is, not only you got all the intelligence from the market. Now today, designers are bombarded with a lot of requests for creating design on an ongoing basis in a season-less merchandising. Here is a tool to enable designers to come out with winning ideas and the machine… Again, models can be trained using both uncontrolled and controlled inputs from the designers. Now we have got a design which will win. I think the next question is, well, how do we ensure that we make the right quantity of these designs?

We are now getting into the assortment, optimization, pre-season and in-season (advanced analytics) Very simple, I’ll just take maybe a minute or two to tell you where you can load all your data into the system and you get an immediate heat map on how I allocate my resources across various brands in the portfolio if you’re a multi-brand retailer. Or if you want to say that you’re a single brand, but I want to optimize across my category, then it will tell you what is good and what is bad. Well, let’s go back to the brand view. It gives you a dynamic OTB optimized on constraints that even fit in. Not only that, you can get to see planned values of what should be our OTB to maximize your revenue.

You want to see a trend view of this. Like Jason was saying that “How do I control my overstock and under stock,” the system automatically comes back and say, “These are the predicted overstock problems across various times in a season.” It can tell you three, four actions that you could take, be it a marketing activity or a deal. Or if you have an opportunity to return to the window. If you want to see that in detail, you could do that. Or if it’s an understock opportunity and you say that, “I’m losing sales here, how do I reorder or replenish?” With that bit of a brief, I would just leave it for Jason to tell us what are the opportunities for working together? We see these are the days of partnership and with the architecture and capabilities of SAP that we can add value to brands and retailers in the fashion lifestyle space.

SAP + Stylumia Combined Possibilities

Jason:
Thank you. All right. Just a quick check, if you can see my screen. Okay. Thank you, Ganesh. It’s really exciting to see what Stylumia has been working on and the innovations and ideas that they’ve put together. I just want to talk for a minute and give you just a brief overview of SAP assortment planning and how together with Stylumia. We bring that together for a really powerful, innovative, assortment planning process. Before I do that, I also wanted to highlight where SAP is in the market when it comes to retail planning.

Last year, SAP was actually ranked as a leader by Forrester in retail planning. Part of that evaluation included one of the solutions that we’re going to highlight today. Well, that’s SAP assortment planning, but SAP has been investing in the space for some time now, and that investment is paying off. The planning solutions, actually, we designed them with a very strong focus on fashion. We also have a suite, a collection, an integrated end-to-end planning process and a suite of planning tools that work together from end-to-end. This can start from the strategic financial planning that feeds into the merchandise planning, setting metrics and targets for the business sales margin across all your channels and stores and categories. This feeds into the assortment planning and as I build out my assortments, what products, what stores, what time, how much to buy.

I can also do that in line with the corporate targets, the merchandise planning targets as well, and then execute on that, execute the buys, managing season and allocate and replenishment depending on the type of merchandise. Today, we’re just focusing on the assortment planning piece there in the middle, and just give you a brief overview of this tool, and then how SAP and Stylumia work together. This tool does actually quite a bit of thing. It begins with the location clustering to use predictive analytics, to figure out the best way to group my stores together.

Ultimately, I can have a localized assortment that meets the needs of the customers that shop in that store. The ability to actually plan attributes and add attributes for planning purposes and how that’s used both for planning and also for greater insight and trends. Something that Stylumia can help with. Also, the ability to define an assortment strategy or an option plan, the number of customer choices I’m going to offer by key planning attribute. And then I build out that product mix, according to that strategy.

Through this, there is a lot of predictive analytics that are leveraged in order to optimize the assortment, provide decision support about products, keep drops, where to sell them, and how much to buy. Also, surrounded by business rules that take into account your business strategy, provide guardrails for the optimizations and things like that. Also, this is a very visual tool, and especially for fashion, you want to be able to visualize that assortment, how your collections look, how they look together before you execute on that. That’s part of it. Also, the ability to introduce new products and do so when you don’t have the master data in place yet and if you want to take those ideas provided by Stylumia and realize those and start planning those early on, share that information with vendors and suppliers and collaborate so that you can secure capacity. You can secure procurement in advance.

All of this includes as well, some embedded analytics and the ability to then take this, execute on the assortment and manage that in season. The flow looks a little bit like this, and then also how Stylumia fits into that flow. Merchandise planning, I established my targets for my channels, my departments, my categories, even down to my stores as needed. And those can flow into then assortment planning. So as I build out my localized assortments, I’m in alignment with those targets. I’m doing the things that are going to help me achieve those financial targets.

Establishing, or building out the assortment can start with strategy first. And some of the things that you’ve already seen from Stylumia is their ability to identify ideas for new styles, the ability to identify white spaces and make that a part of your strategy. “Now that there’s a white space, okay, how do I fill that white space with the right products?” It’s really about taking those insights from Stylumia and being able to then execute on that and build out that assortment with those ideas as well, and introducing the support for introducing those new products.

Also, the ability to enrich those products with additional attributes that help not only with the planning, but being able to analyze and see which attributes contribute most to certain trends, to sales, to margin, what really drives those things in the business, and then taking actions on those. Also, really important, which we’ve mentioned already a little bit, is the ability to improve fashion forecasting. In fashion, oftentimes many of your products are new with no history. What’s the best way to really come up with a more accurate forecast? So bringing together both the forecasting capabilities of SAP and also the forecasting insights and trend analysis from Stylumia to make a better fashion forecast together. We can take those trends, insights, and competitive information and other things into the SAP forecasting capabilities.

Also, the advanced analytics, which you saw some as well to help provide decision support in the assortment planning process, and also to be able to do things like competitive intelligence as well throughout the process so you are again creating the best possible assortment. And then creating the best possible forecast for the best possible buys. This can bring a lot of benefits by bringing these tools together. Again, one of those is to really improve forecast accuracy for fashion and seasonal products. It’s a tough problem and Stylumia has brought a lot of innovative ideas to address that problem and improve that. That has huge financial benefits, both in terms of not under buying or buying a lot less so that you don’t miss out on revenue and margin on great products and to avoid and minimize over buying and the tremendous amount of waste that causes, the increase in costs and the reduction in margin from many markdowns that happens.

Advanced analytics, to really leverage the data that both you have inside the organization, but to also look outside of your four walls, your organization. What’s happening in the marketplace? What are competitors doing to really scour the internet and identify insights that you maybe weren’t able to capture in the past, but now you can? And then identify those assortment wide spaces, those opportunities, and not only identify them, but have the ability to execute on them through SAP. And being able to earlier spot trends, what’s happening. The early you sense something, the better you can respond to that.

Again, reducing that waste and costs through better buying. Part of what Stylumia brings that is really innovative is not just leveraging the history and making improving decisions, but really bringing those forward-looking insights, what’s happening now in the marketplace? What are the trends? And then using those both pre-season and in-season to make better decisions around the assortment. And to identify and validate new product ideas before introduction to really come out with products that customers want, that are really in line with trends, what customers are buying, hot sellers and so on. As you do this, you will have happier customers through more relevant assortments. This is truly win-win and together with SAP and Stylumia, we can offer these benefits and improvements to retailers and fashion companies. I’d like to turn that over then to some Q&A time.

Azadeh:
Thank you very much, Ganesh, for the presentation on the solution and, Jason for laying out so amazingly and comprehensively, how this combination work and what value we can bring to SAP customers with this combination. It was amazing. Thank you. Then I would say that I have to thank everyone for attending the call, Ganesh, Jason. Ganesh, it was a pleasure having you in the corporate and it was truly amazing, but it’s only the start of our joint efforts. Jason, thank you once again for joining the session and presenting how this combination works and thank you for helping us through the course of the program and also mentoring Stylumia. Thank you, everyone, have a great evening and until next time.