Embracing technology to refine buying processes and retail merchandising planning is becoming crucial than ever. Celio – a leading French fashion brand operating in over 50 countries including India – is a firm believer in the impact of data analytics and artificial intelligence on merchandising decisions. As Celio expands its footprint, Vinit Doshi, Head – Product, Sourcing and Design, Celio details the brand’s adaptive strategies, from four-season buying to addressing online retail challenges in an exclusive conversation with Team Apparel Resources. Here are the excerpts.
AR: How has technology played a role in streamlining the buying process at Celio and what specific tools or platforms have been instrumental in this integration?
Vinit: I would say that tech is very important part of our business today. It allows us to be quick and more effective as compared to conventional methods. We have to use advanced technology in different processes of brand building.
For all brand like ours, there are two kinds of buying that happen. One is buying for range building and the other one is retail buying/allocation for the stores.
Ideally most brands, through trade shows, take orders from trade partners and goods are manufactured as per pre-booked orders. This reduces the risk but it increases the lead time and GTM (go-to-market) time. Our model is different from the above one. We develop the range and allocate it to stores and trade channels. This requires our people to have good understanding of every retail door in the country. So this leaves us with two options – either we hire an army of retail planners or we take tech help. In the first option, as and when we keep expanding, we would need to add people because a retail planner can efficiently handle limited number of door count. The second option is that we rely on AI technology. We realised this need in 2019, and by 2019 end, we started talking to different tech service providers who could help us in this growth journey by providing such technology where all these nuances of hundreds of stores can be addressed.
In 2020, we partnered with tech firm Increff and implemented its retail planning tool. Despite challenges due to Covid preventing a pilot, we launched it directly as retail started opening up in bits and phases. Initially, iterations were required and we had to address teething issues. Over a period of time, we achieved notable success in retail allocation and planning. Now, we plan to extend the tool capabilities to product buying and range planning.
This is how we plan to integrate tech in our buying and retail planning processes.
AR: Can you share examples of how data analytics or artificial intelligence has influenced decision making in merchandise planning, contributing to a more data-driven approach?
Vinit: Planning for large number of doors requires good understanding of every door and it means a lot of historical data needs to be interpreted. We had an option to make store clusters or treat every door as a unique retail sales point. Clustering is what everybody in the industry preferred till there was a good solution by tech companies. We have now moved on to treating every store as an ideal retail sales point and plan for that independently. This was a big change as every store had different variables and clustering them was not the long-term solution. Fashion retail has to deal with these variables effectively for better inventory management. When these many variables are analysed and interpreted correctly, it can change the way one has to allocate stocks and also finally support in range building. This is a big change and it was possible only with the help of AI and data analytics.
When it comes to integrating AI in merchandising planning, the tool helps us know the inventory we are buying, where it is to be allocated and what kind of traction we’ll be able to get out of the merchandise bought for those stores. The tool we are using is very dynamic; usually, when done manually, merchandise is allocated to doors where it is sold the most. But this particular tool goes beyond that; it tells us where it had the maximum traction and not only maximum sales.
All these variables require lot of interventions and deliberations and finally man and machine have to come to a point where limitations of each other can be addressed best within given constraints.
AI also picks the historical data, so it comes with its own challenges to align it with brand strategy. For example, data analytics uses historical data and fashion works on season. Thus, the tool compares season to season for good analysis like Spring/Summer ’24 output will be basis Spring/Summer ’21, Spring/Summer ’22 and Spring/Summer ’23 sales and growth trends. What the tool misses out is the ongoing Fall/Winter trends which could be different. This is where manual interventions are required. It also requires interventions to accommodate changes in brand strategy as per changing trends.
All the interventions from AI and changes in buying cycle have helped us improve our full price sell-through by more than 10 per cent.
To conclude, AI as a support is must and it is just the beginning and there is a lot more on the wish list for fashion industry in specific.
| Our different teams, including e-commerce, retail planning, buying, warehousing and marketing teams, are actively exploring tech-driven solutions. Specific areas of interest within buying, merchandising and retail planning include demand forecasting, where we seek tools with over 90 per cent accuracy to improve efficiency. |
AR: In the context of retail planning, what advancements have been crucial in forecasting consumer trends, managing inventory effectively and ensuring a responsive and agile merchandising strategy?
Vinit: Right planning for every store is very crucial. To eliminate problems of under or over buying for the season, we don’t do 100 per cent of our buying in the first stage. We have moved to four seasons’ buying as against two seasons’ buying in a year. This helps reducing our time period for which we need to forecast and allows us to amend our requirements very close to the season. This also paves way for relaunching styles that are trending.
That’s how we try to manage inventory. Is this a solution to 100 per cent of our requirements? As of now, it doesn’t solve everything. As I keep saying, there needs to be a tool that can forecast all these things much in advance so that there’s not a lot of stress in the entire supply chain. Buying close requires everyone to stretch and brings in a lot of stress to the entire supply chain process. That’s where technology will play a big role going forward. If it can forecast, if not 100 per cent but with at least 90 per cent accuracy to start with, it will facilitate right planning for everyone in the chain. Robust AI solution is way forward.
| Over the past two years, Celio has been expanding its presence through its own stores in India. Additionally, it is present in all major retail giants across India, including Lifestyle, Shoppers Stop, Pantaloons and Reliance Centro, positioning itself in the premium segment. While Celio currently doesn’t operate in the value segment, it has a substantial presence in both large-size and mid-size Multi-Brand Outlets (MBOs). It also has a distributor network aiding penetration into smaller cities, including Tier-1 and Tier-2 towns. In total, Celio’s footprint now encompasses over 600 stores, including 65+ Exclusive Brand Outlets (EBOs) and collaborations with large retailers and MBOs. |
For example, AI now has to predict and forecast keeping these subjective nuances of fashion retail. This industry is very subjective and relative. Let’s take an instance where I might find that my yellow polo is doing very well today, but suddenly other new seasonal colour comes up and starts trending well, this is where the yellow colour Polo may automatically slow down and all earlier forecasts may go wrong. All my AI forecasts would have said that yellow will be able to sell an ‘x’ quantity. Now, the new colour which was never in the picture when this forecasting was done takes over yellow colour, leading to forecasting error. That’s where tech partners have to come forward with solutions if they can find one. This was just one example and there are several such nuances that decide performance of the style/option. If this gets solved, lot of cash will be freed up from working capital.
AR: With the rise of e-commerce and digital platforms, how has Celio adapted its buying strategies to align with online retail trends and what role does technology play in this transformation?
Vinit: There are conscious efforts to create a range that meets e-commerce requirements now. Earlier, what was happening was we had a one standard range and it was all offline-focused. On e-commerce, it used to go only once it was launched. And it used to be presented as a product and didn’t represent the product story in totality. We realise that e-commerce is a different ball game altogether. If you have to grow online, there are multiple things required. One is your range should be big enough. At Celio, we work on a very curated range and to meet e-commerce demands, we are working towards increasing our width.
Also, as I said, what we require is, we have to do a lot of things for the marketplace. So how do you bring out stories to life? How do you call out all those innovations? How do you make sense for the consumer? Celio is working on that front, and as a brand, we introduce a lot of newness and innovations. Otherwise, with too many products on the same platform, it gets lost, especially when most other brands have collections three to five times larger than ours. So, this is what we are working on.
But if you ask me honestly, there is a lot of work that Celio has to do to gain shares in online business. The advantage that online partners have is they have a plenty of data. They can tell you immediately what’s working, what’s trending, what’s not trending. Now we have to use that data to understand changing consumer preferences and quickly adopt them in our range. Celio, as a group, is well-positioned to bring about changes quickly.
AR: Can you share any upcoming tech-driven initiatives or innovations that Celio is exploring to further enhance efficiency and customer experience in the areas of buying, sourcing and merchandising planning?
Vinit: Our different teams, including e-commerce, retail planning, buying, warehousing and marketing teams, are actively exploring tech-driven solutions. Each team collaborates with different tech partners to recruit more customers, enhance customer experience and improve customer loyalty. Specific areas of interest within buying, merchandising and retail planning include demand forecasting, where we seek tools with over 90 per cent accuracy to improve efficiency. Retail planning has shown positive results, with sell-through increasing from 48 per cent to 60 per cent.
Another area to explore is accurate prediction. Challenges arise in obtaining accurate input and output, particularly when feeding data to tech partners close to the season. Despite past attempts, we believe advancements are possible with time.
Additionally, we recognise the need for tools that can provide insights into offline consumer trends. Currently, online data is readily available, but it comes with challenges such as discounting and promotional interventions. To address this, we hope to hear from tech partners about how AI tools can capture consumer behaviour, and as some industry leader aptly said, consumer is leaving traces and if AI can capture those traces, it will be a game changer.
Simply put, our focus areas for enhancements include demand forecasting accuracy, precise predictions and tools that provide insights into offline consumer trends.
