As witnessed in the last few years, retail demand forecasting is vital for big box stores and small players, especially for survival and growth. Inaccurate, decades-old approaches to forecasting demand are no longer applicable in the face of a constantly evolving consumer market. The same period witnessed a dynamic shift in population from urban city centers to other suburban areas, impairing retailers’ ability to position their inventory based on historical trends. From rapidly transposing demographics to population-based demand changes, decreasing brand loyalty to an exponential increase in online shopping – these are a few examples of how the market reacted to the pandemic. Even after the pandemic ended, retailers were confronted with a severe existential crisis induced by global recession, causing income disparities and driving businesses toward off-brand and discount retailers. These bottlenecks are unlikely to faze away anytime soon. Therefore, reliable demand forecasting methods leveraging new-age technologies are gaining momentum today. In the same light, the demand sensing software solution from TradeEdge can help retailers easily capture the early signals and stay market ready all the time.
Recession-induced market trends – How do they impact retail demand forecasting?
The recent alterations in market trends have been reshaping the recession-driven, post-COVID era retail environment.
From the business standpoint, a few changes, such as the conversion of stores into mini-fulfillment centers and pickup points and restructured supply chains for catering to specific household orders rather than to large-format stores, are currently witnessed.
What do these trends mean for retail business? Firstly, retailers must forge new partnerships for continuity and success. Secondly, they are vying for meaningful relations built on trust with their consumers and other stakeholders. And, finally, bringing structural changes in the industry to thrive in the new regular retail.
But given the rising market volatility, these measures might not be enough to weather the storm. Opposing traditional orthodoxies is probably the first step, but retailers need extra pairs of helping hands to balance demand with supply in the face of evolving customer preferences. Artificial Intelligence and its extended capabilities are being leveraged to improve retail demand forecasting methods, ensuring businesses are more future-ready than earlier.
Tracking demand forecasting challenges of the retail supply chain
COVID-19 left a deep gash in the system, causing a paradigm shift in how businesses used to operate earlier. The challenges mentioned above are just the tip of the iceberg. The effect of the global health crisis and its subsequent recession has been felt far and deep in the retail supply chain. Difficulties in accurate demand forecasting are their aftermath.
Since the anticipated recession is likely to compound the myriad issues already plaguing the retail industry and its supply chains, the following are a few examples to highlight the intensity of the crisis.
A slowing economy
High inflation’s financial reality minimizes the impact on higher-income households. The pandemic-fueled savings continues to ignite their strong spending. However, their even shift towards more in-person services spending is hurting big retailers and is seen as one of the significant contributors to an economic meltdown. Goods price inflation is rising higher than service price inflation.
Poor spending or altered preferences?
Customer preferences have shifted, for example, from premium to private labels. Sales for specific categories of goods, like sporting goods, hobbies, books, and music, are losing steam. Likewise, demand for durable goods is taking a backstep, implying poor retail sales at consumer durable stores. On the other hand, inflation is denting a customer’s spending capability, with the share of users intending to delay purchases going up since mid-2021.
Softening demand
The anticipation of a recession has compelled many buyers to withdraw themselves from overspending. The purchase shift is more towards necessities, not so much for discretionary goods like clothing and electronics. Most businesses are witnessing a decline in demand for premium products in their portfolios; hence, the gradual demand shift will also affect supply chains. And when the inflation is running hot, with the inventory building into weaker demands, forecasting for the future will become a tough nut to crack.
Big inventory builds
The lingering pandemic issues resulting in multiple lockdowns in China last year compelled many companies to build their inventories. Companies fear not having enough to meet the sudden burst in demand for goods and end up ordering a lot. With inadequate forecasting estimations, big inventory builds might significantly challenge retailers.
Immobile stocks
Due to continuous product and material shortages, certain companies pivoted to bulk orders in the previous two years. Unfortunately, consumer purchasing declined further, resulting in stockpiling, pausing, or canceling supplier orders. However, if and when demand Further, immobile inventory would impact cash flow, hampering their capacity to balance supply. Similar unstable scenarios can result in poor retail demand forecasting.
Labor shortages and an aging workforce
Demand and sales forecasting also depends upon the availability of labor and an adequate workforce. But, an expected economic meltdown could easily exacerbate the ongoing supply chain labor crisis. Labor shortages and an aging workforce are already impacting the logistics and manufacturing industries, compounded by the new rounds of layoffs during the recession. As many of these points are estimations of the growing uncertainties, retail demand forecasting methods will fail to paint a near-accurate picture.
Fragmented approach
A holistic understanding of customer demand across all product categories and channels benefits the entire end-to-end supply chain, supporting efficient demand forecasting for retail. But a fragmented approach and existing siloes prevent the timely availability of data from the supplier end, which can easily falsify the estimated demand figures or increase businesses’ inability to match.
Elevating demand forecasting for retail with AI
The world market looks decisively different today, but the worst is not over yet. The supply and demand shocks of the global economy shutdown unearthed vulnerabilities in the original production strategies and supply chains of firms everywhere. In addition, distribution issues worldwide and complexities of supply chains’ global scope hampered companies’ ability to forecast demand and determine the best ways to fulfill them. Furthermore, the previously indicated market for in-store sales was greatly impacted by the growth of online purchases. Therefore, retailers need to adopt a more omnichannel approach, meaning retailers should provide accurate inventory estimations online to induce more in-store visits.
But clinging to manual systems and antiquated software solutions only adds to the rising difficulties.
Therefore, investors are embracing next-generation technologies and software solutions incorporated with AI, ML, and advanced analytics to pave the way for autonomous planning. Demand forecasting for modern retail in the face of unprecedented challenges requires the assistance of granular data in real-time – data reflecting changing buying patterns, customer buying behaviors, supply chain, and logistic channels disruptions, macroeconomic factors, weather patterns, and so on. All these factors and their interrelationships significantly affect demand daily. AI technology can easily leverage such unstructured datasets to improve forecasts.
With its extended machine learning capability, AI-enabled demand forecasting methods can quickly learn from the past and deliver better, more accurate predictions every time. Daily forecasts further enhance stakeholders’ ability to respond promptly to sudden changes in demand patterns.
AI in retail demand forecasting – Three real-world examples
#1. A Swiss food and drink processing giant struggled while handling massive volumes of sales and inventory information for over 2000 brands from siloed and disparate internal systems, multiple channel partners, e-commerce, and syndicated data providers. This resulted in sub-optimal sales planning, supply chain inefficiencies, and sluggish business growth. However, with the help of TradeEdge Demand Sensing solution, the client was able to unlock the value of data and identify $200M in potential cross-sell and up-sell opportunities, thereby winning an edge over competitors.
#2. The consumer goods industry comprises hyper-competitive markets, where high customer expectations, a wide range of products, and complex supply chain networks are primary challenges faced by CPG companies. Therefore, the latter needs efficient demand planning to ensure the availability of products at the right place and time to retain their hold over consumer markets. Unfortunately, timely placed data, critical in demand planning, must be solved owing to their multiple, disparate sources. Further, companies also needed more consistency and questionable data quality. This significant issue troubled a multi-billion-dollar global consumer goods company. Data inconsistency impacted the client’s sales efficiency, partner relationships, and productivity. However, with the help of TradeEdge Demand Sensing, the client overcame data challenges by streamlining data acquisition, improving data quality and time to insights. As a result, the client achieved 5X growth in five years than previously witnessed.
#3. The sports merchandise market comprises a network of intermediaries, wholesalers, distributors, and retailers. Each follows its point-of-sale (POS) system that affects real-time visibility into sales performance across regions and products. Without real-time harmonized data, end-to-end supply chain visibility across the ecosystem remains a persistent challenge affecting adequate demand planning for companies. Our client, a multinational corporation with a global footprint in the sports merchandising and apparel business, wanted to redesign its Point of Sale (POS) Business Intelligence system to gain insights into their retailers’ POS and inventory position. As a result, they could increase market share, collaboration, and effectiveness with better demand visibility. Also, near real-time, high-trust data could enable them to make predictive business decisions. With TradeEdge’s support, the client was able to increase demand visibility by 60% and gain access to near real-time harmonized data.
AI in retail demand forecasting – Business benefits
Today, the global AI adoption in the retail market is growing at a CAGR of 28%, with an estimated future value of above US$ 127.09 billion by 2033. Driven by the growing popularity of online shopping, tech-savvy, and mobile-friendly Gen Z population, retailers are investing more in AI technology to improve areas such as demand forecasting, inventory management, and customer experiences.
AI for retail demand forecasting offers the following benefits:
Common pitfalls of retail demand forecasting
There are some common pitfalls associated with legacy retail demand forecasting methods.
Factoring historical sales data: Using historical data would eventually lead to repeating past mistakes for demand predictions. The global health crisis and current economic volatility have devalued historical figures. It condemned retailers to consider new datasets reflecting factors like cyclicity, seasonality, inflation, geo-political scenarios, price changes, promotions, holidays, and weather patterns.
Lacking granularity in demand decisions: Traditional forecasts aggregate past data to give a surface-level view of demand. But daily patterns, trends, and variations in customer behavior remain untapped. The legacy approach enables decision-making on product categories but not at SKU levels. Hence, decisions about replenishing seasonal stocks stay out of the drill.
A top-down approach to demand planning: Demand predictions are made at the highest level of the supply chain while changes in demand at the granular levels remain unaccounted for. Unfortunately, inventory operations will never improve unless predictions at granular levels are inaccurate.
Bloated safety stocks: Safety stock in the buffer improves availability and service levels and eliminates chances of out-of-stock situations. Since most companies operate in a system where suppliers’ lead time is longer than buyers’ tolerance time, the buffer stock is a safety valve to provide uninterrupted services. Unfortunately, this condition can lead to locked-up working capital and bloated demand numbers.
No scope for strategic building: Yet another pitfall for traditional retail demand forecasting is the lack of crafting stellar strategies to improve market share and beat the competition. Low levels of data transparency and forecast accuracy further hinder retailers’ ability to identify plausible risks and mitigate them or create a buffer plan to meet a range of outcomes.
How to get started with retail demand forecasting
By utilizing predictive analytics, AI, and ML capabilities, demand forecasting methods for retail have improved manifold, driving the accuracy and reliability of outcomes. Here are a few tips to get you started.
Large-scale data processing: Machine Learning can easily automate planner work and processing of large datasets. Evaluating each data set potentially impacting demand can elevate forecasting accuracy.
Predicting the effects of decisions: Commercial choices like promotions can affect sales volumes immensely. Machine Learning algorithms are leveraged to process large volumes of data demonstrating the impact of promotion and marketing activities, price elasticity and visibility of brands, and impacts of price changes on other products within the same category.
Considering price changes and their impact: Cannibalization or demand shifts occur when product prices are reduced, impacting the demand for goods in the said category. This means demand shifts immediately to cheaper products. To prevent over-ordering of low-priced goods, retailers should adjust forecasts for non-promoted goods and adjust figures in their replenishment plan.
Collaborating with vendors/partners: With a regular data-sharing routine developed between partners and suppliers, retailers can gain an end-to-end overview of how their network operates. Data sharing improves collaboration and provides enough bandwidth to arrange for risk mitigation when a range of issues arise to impact demand signals.
Accounting for new product introductions: New product introductions can challenge sales forecasts because there is no historical data to refer to. But, often, reference products with historical data can serve as a blueprint. AI-enabled demand forecasting tools can automatically select reference data and quickly update SKU forecasts when actual sales patterns emerge.
Evaluating external factors: External factors include those not under anybody’s control. These factors change the direction of demand for goods/services, including weather forecasts, local events, or competitors’ business decisions. Machine Learning can easily automate extracting such factors and intergrade them into your forecast. This reduces forecast errors on the product, group, and location levels.
Key takeaways
Accurate demand forecasting for retail helps reduce risks and supports efficient decision-making for higher profit margins, seamless cash flow, proper allocation of inventories, and scale opportunities for expansion. This way, retailers can reduce lost sales and ensure customer returns while lowering safety stock to avoid spoilage. Further, strategic and operational plans are formulated around accurate forecasting. With the increasing use of new-age technologies, predicting future demand has become more streamlined, data-driven, and accurate, enabling retailers to easily keep up with unprecedented changes. It helps them catch the early demand signals accurately and stay prepared aforehand.