Currently, there is a $210 billion disconnect between what consumers want to buy and what retailers have in stores. This inventory distortion reflects a severe misalignment within the retail industry, but, even more disturbingly, it also reveals the harrowing impact inventory stocking practices have on our environment. As retailers attempt to align stock with potential demand, many find themselves with substantial inventory pile-ups due to overstocking stores, eventually lending towards unwanted markdowns or wasted inventories. In fact, every year retailers produce over 150 billion garments – of which a whopping 30% are never even sold. Not only does this discrepancy result in huge margin losses for retailers, but it also contributes to the 12.8 million tons of merchandise tossed in landfills each year.
From a sustainability perspective, these numbers show the retail industry’s impact on our environment, as well as the social responsibility retailers must assume to minimize inventory waste. However, inventory waste is not the only culprit as other resources necessary for inventory production also force a hard hit on Mother Nature. For instance, a single t-shirt requires 2,700 liters of water for production – the same amount the average person consumes in 2.5 years. Even more upsetting, a basic pair of jeans needs 7,000 liters of water to produce.Yet, despite these statistics, a recent report shows the fashion industry is actually slowing down its efforts in sustainability. With the retail world in a constant battle to adapt to changing consumer expectations, sustainability efforts continue to take a back seat to existing business operations and processes.
However, those who do seek to adapt a more environmentally conscious approach to the business are tasked with answering the most difficult question at hand — how?
How can retailers produce merchandise more sustainably?
For retailers to sustainably run operations, there needs to be a stronger focus on obtaining an accurate view of demand, and effectively aligning inventory to meet such demand. If retailers could accurately predict consumer demand and sequentially reduce the amount of merchandise produced to match demand, then wasted inventories and over-abuse of environmental resources could be significantly curtailed.
The Current Dilemma
Adapting a more sustainable model of business where inventory is more accurately aligned with demand is much easier said than done. Understanding what to buy, how much to buy, or how much to produce to meet consumer demand continues to elude retailers. Today, the challenge is many retailers struggle to understand how much an individual store will sell of a particular garment. Not to mention the more complex issues of how many garment choices should be made available in a particular selling location (online and in-stores) and, subsequently, how many of those yet-to-be-produced garment choices must be manufactured to cover demand. With a perplexity of important decisions to consider, it is vital retailers respond to these inventory dilemmas with precision and efficiency.
The reality is traditional methods of demand forecasting fail to address many of the data challenges retailers face today, which make it difficult to align the right inventory, in the right mix and quantities, to potential demand. The ineptitude of traditional forecasting methods is primarily due to the nature of retail data – which is often volatile, sparse, and noisy. Difficult-to-read data is especially frequent with fashion, seasonal or short-life products, as these items commonly have little to no history. In addition, new pressures to meet customer expectations in today’s complex retail environment (where customers can buy, receive, and return anywhere) make it challenging to accurately forecast at the level of granularity necessary to improve inventory decision-making.
Without being properly equipped to accurately forecast demand, many retailers have resorted to a templated approach, where deciding what and how much to buy (for example, all store buys and top store buys) are typically established several months out and based on summarized data. This method of decision-making is unproductive as it lacks the capacity to act in real-time and provide relevant insight to decisions at hand.
Three Areas Widening the Consumer-Inventory Gap
There are three key areas driving the disconnect between consumer demand and the available inventory:
1. Number of customer choices available
Retailers are often over-assorting by making too many choices available to consumers at any one selling location. More choices generally do not equal more purchases. In fact, too many buying options often confuse customers or simply overwhelm them. Additionally, product assortment is decided at a summary (store group) level, which is ineffective when managing inventory as individual stores in a store group seldom behave the same way. Different stores have different customers who have different preferences - rendering this method as less-than-optimal.
2. Quantity produced of each customer choice
In addition to an overinflated number of choices, the quantity of each item produced is also drastically overestimated. Retailers typically build in contingency to account for the inability to accurately predict consumer demand at each selling location. As a result, retailers end up overstocking to avoid under-stocking, since lost sales opportunities are more costly than marking down overstocked merchandise.
3. Inaccuracy based on timing
Finally, large quantities of merchandise are committed upfront in order to capitalize on lower unit costs. Retailers attain products this way because they know merchandise can be heavily discounted in the future, if necessary. However, this method of overcompensation is lagging and highly inefficient as it does not align the timing of products placed in stores to the existing demand of consumers – leading to an increase in markdowns and unbought merchandise, all at the expense of the retailer.
How to Match Inventory to Consumer Demand
When it comes to operating sustainably, AI helps address the inventory distortion problem by improving the accuracy of quantities needed prior to production, as well as how much inventory should be placed in stores, post-production, to maximize full-price sell through (thus reducing the need for contingency stock). AI’s ability to provide an early read on trends and selling propensity also makes small batch manufacturing more feasible than ever before.
AI and machine learning technologies allow retailers to proactively anticipate fluctuations in demand and optimize inventory use – from managing omnichannel retailing to demand forecasting to uncovering the latest consumer trends. AI transforms planning and buying by incorporating granular insights into what consumers want to buy, how they want to buy and receive, and how consumer preferences are impacted by existing assortments and context. By answering the questions of what to buy, how much to buy, how much to produce, and where to place inventory, AI empowers retailers to anticipate changing consumer demands and accurately align inventory.
This innovative approach addresses all three of the key areas driving the existing disconnect between consumer demand and inventory by opening branches of opportunity to make decisions differently. First, retailers can use AI to recommend the number of choices to carry in a store and how much to buy for each selling location. Second, AI enables retailers the ability to test and learn quickly (and often) through early reads and insights buried in sparse data – delivering a more flexible, proactive, and agile approach to decision-making. Lastly, AI adjusts decisions based on the latest and best view of demand, allowing previously silo-ed inventory decisions (for example, around what to buy, where to place merchandise, and where to ship from) to weave together, forming a clear, holistic view around how consumers shop.
For the sake of the environment, it is critical for retailers to accurately align supply with demand. Less than 1% of the material used to produce clothing, globally, is recycled into new clothing and only 12% is recycled into other products, such as insulation or mattress stuffing. There is an incredibly small chance wasted products will be recycled and put to productive use, and an overwhelming chance they will instead be carelessly discarded at the expense of the environment – again, reiterating the necessity for retailers to manage inventory at the optimal level. Additionally, 1.2 billion tons of greenhouse gas emissions stem from textile production, which is more than all international flights and maritime shipping combined. Whether those in the retail industry are aware of it or not, mismanaged inventory holds extreme consequences for our environment, which cannot afford to bear additional waste or exploitation.
With AI, retailers can improve inventory decision-making while also significantly decreasing the negative effects of inventory distortion on the environment. However, at the current rate, the fashion industry remains far from sustainable. Retailers must increase their pace towards systemic change within existing business operations through the use of innovation to help overcome the technological limitations contributing to constant environmental degradation. In the end, retailers can use AI to help plan production cycles more sustainably by making smarter and more accurate inventory decisions from the get-go.
Access the full article, Wasted Inventory: What It Costs and How to Prevent It, here.