How a fast fashion brand is using the predictive model to manage the lifecycle of its new launches ?

image
Story
This is an Australia based successful fashion house boasting various authentically bold and contemporary fashion labels under its brand name. Headquartered in Sydney and focused on women fashion clothing, they sell their products on various ecommerce channels and their own website. They are a fast fashion brand, who react quickly to the trends and bring the new collections rapidly across all their labels. They operate with the aim to supply unrivaled choice, quality and price-focused fashion to the female apparel market. The typical lifecycle stages of any of its products are: Identifying the trend, introducing the in-trend garment on the marketplace, growth in product sales, sales maturity and clearance sales.
image
Approach
Aays Analytics partnered with the firm to establish a cloud data infrastructure in place that helped them to systematically collect data from various sources, and analyze it effectively using machine learning and advanced analytics pipelines. This helped streamline their supply chain processes by considering various internal and external factors affecting the product lifecycle. ​Furthermore, the insights from our analytics pipeline helped the firm to make faster decisions and react more quickly to unforeseen events, helping them to reduce their losses. We were also able to accurately predict the sales of historical and newly launched products, enabling the supply planning team to plan the optimal launch inventory, prices and re-stock levels, which helped them to optimize the entire product lifecycle starting from regular sales to clearance.
image
Key Results
We have delivered ​many descriptive analysis reports through interactive Power BI dashboards, helping them to drive daily activities of supply planners. Along with the reports like stock on hand and re-order level, the predictive models suggest optimal launch inventory value and predict future sales. Together all these help them take data-driven business decisions and optimize entire product lifecycle. Currently, our entire ML pipeline helps them to accurately predict future sales and required future stock levels with over 90% accuracy.
Challenges
Being a fast fashion company, the life cycle of its products are shorter compared to the basic fashion wears, and hence there is a pressing need to plan a product life cycle extremely well. Their major challenge was to predict sales of a new launched product so that they could do production planning, inventory planning, clearance planning and an appropriate marked down price. Though the client had started to look into data to generate some meaning, they hadn’t laid any groundwork to help them succeed with data. Most of their data analysis was ad hoc and was driven by intuition without looking into the actual data or user sentiments.
The data environment that the client had, before associating with Aays Analytics
img
Integrating multisource data

The planning teams were using spreadsheets for collecting and analyzing customer and sales data. The data tracking was done on a weekly or monthly basis.

img
Data Collection in spreadsheets

Data was collected from multiple sources viz. Sales CRM, social media post, their own in-house systems etc., but there was no single point view to all this data.

img
Generating insights from data

Whatever data collected, they had the most trouble in identifying the insights from data as there was no coordination between the planning and functional teams. The decision makers did not have access to all the relevant data to make impactful decision choices.

Key Results
We have delivered ​many descriptive analysis reports through interactive Power BI dashboards, helping them to drive daily activities of supply planners. Along with the reports like stock on hand and re-order level, the predictive models suggest optimal launch inventory value and predict future sales. Together all these help them take data-driven business decisions and optimize entire product lifecycle. Currently, our entire ML pipeline helps them to accurately predict future sales and required future stock levels with over 90% accuracy.
90%
accuracy in predicting future sales and stock levels
99.8%
accurate descriptive analysis reports to drive daily activities of supply planners
Solutions that we provided using Aays Data Platform
Centralized Data Mart to Analyze data for Predicting Sales
Data was de-siloed, and a centralized data mart was created to integrate data from multiple sources using APIs, and scraping tools.
Unified Reporting
Company wide KPIs, sales reports, and product planning reports were generated on an automated basis until stakeholder trust was built over a period of time.
Impactful Visualization
The descriptive analysis reports and predictive model insights were fed into power BI to create compelling business-focused visualizations. The easy to navigate and personalized dashboards immensely helped the business planning, sales & marketing teams to carry out daily tasks with more precision and take data-driven decisions.
Machine Learning Model to Predict Sales
Proprietary ‘Sales prediction’, and ‘Inventory optimization’ machine learning models were developed using advance boosted regression algorithms and deep recurrent neural networks. These models captured key business inputs like internal promotion data, social media data, consumer sentiments, external factors (weather, pandemic) and historical sales data to predict future sales, which in turn were used to preplan and optimize the inventory. The models were not only able to accurately predict the sales of continuing products but also provided a good estimate of sales of newly launched products.
"We've been helping organizations gain visibility and deeper insights into revenue, ROI, profitability, cashflow and risk and are helping them identify areas that required attention to impact positive business outcome earlier than otherwise possible " ANSHUMAN BHAR CEO, Aays Analytics