Utilising novel data in demand forecasting to solve Fashion product launches

我们正在通过预售数据尚未开发的力量，解决时尚界的需求预测和决策优化问题。时尚业的商业模式在上个世纪基本上没有改变。大量库存是在低成本地区生产的，交货期很长(6-9个月)。这种时间滞后意味着品牌在订购时承担巨大的预测风险，商家做出数十亿美元的决定仅仅基于中学数学和直觉的预测。运用直觉，而不是分析。这种模式导致了一个行业，由于需求不足，15%的产品被毁。总体而言，不准确的需求预测导致供需不匹配，使该行业每年损失1万亿美元。对环境的成本是巨大的。到2050年，时尚业的二氧化碳排放量将达到全球的26%。解决方案需要对产品需求的准确描述。市场上有一种解决方案的缺口：1.对新产品需求的初步了解：理想情况下来自实际交易2。将这一早期需求信号准确映射到终生需求预测3。优化这一需求预测，以产生可行的建议2022年6月，我们推出了新的软件产品进行预售，使品牌能够在主要发布之前通过将部分库存直接预售给消费者来测试需求。这个项目提供了剩下的两个缺失的环节(上面的第二点和第三点)。学术研究表明，这些售前数据比目前市场上的任何东西都更能预测终身需求，预测的准确率要高出50%-70%。我们的目标是释放售前数据的预测能力，建立在我们董事会两名需求预测教授所做研究的启发下。将决策从直觉转变为严格的优化，实现利润最大化和浪费最小化。

We are solving demand forecasting & decision optimisation in fashion, through the untapped power of pre-selling data.The fashion industry operates a business model that has remained largely unchanged for the last century. Large amounts of stock are manufactured in low-cost geographies with long lead times (6-9 months).This time lag means brands take huge forecasting risk at time of ordering and merchandisers make billion-dollar decisions on the basis of little more than secondary school maths and gut based forecasting. Applying instinct, not analytics.This model has led to an industry that destroys 15% of products that get made due to insufficient demand. In total, inaccurate demand forecasts result in supply-demand mismatches that cost the industry $1trn/yr.The cost to the environment is enormous. The fashion industry is set to reach 26% of global CO2 emissions by 2050\.The solution requires an accurate picture of product demand. There is a gap in the market for a solution that provides an:1. Initial read on demand for new products: ideally from real transactions2. Accurate mapping of this early demand signal to a lifetime demand forecast3. Optimisation of this demand forecast to produce actionable recommendationsIn June2022 we launched our new software product for pre-selling, enabling brands to test demand prior to the main launch by pre-selling a portion of inventory directly to consumers. This project delivers the 2 remaining missing links (points 2 and 3 above).Academic research demonstrates that this pre-selling data is more predictive of lifetime demand than anything currently on the market, generating forecasts that are 50-70% more accurate.We aim to unlock pre-selling data's predictive capacity, building on an idea inspired by the research conducted by two professors of demand forecasting that sit on our board. Transforming decision making from intuition to rigorous optimisation, maximising profit & minimising waste.