The biggest challenge in the process of setting a company's purchasing budget is the increasing market volatility. In the past, when the market was relatively stable, it was possible to set a somewhat reliable purchasing budget based on sales data from the previous year.

However, the limitations of this approach have become clear as the market environment has changed rapidly in recent years.

In particular, the reorganization of the global supply chain and the acceleration of digital transformation are further amplifying market volatility.

In fact, according to a 2023 survey by the Korea Manufacturing Association, 78% of manufacturers said that existing sales volume forecasting models are not sufficient to respond to rapid changes in the market. This shows that predictive methods that rely solely on historical data are no longer effective.

What is even more noteworthy is that market volatility is not simply a result of changes in demand, but also a result of qualitative changes in purchasing patterns.

The shortening of the purchasing cycle due to the growth of online commerce, rapid changes in consumer preferences, and fluctuations in market share due to the entry of new competitors are variables that are difficult to predict based on historical data alone.

In fact, companies that set their purchase budgets in the traditional way are facing the situation where they have to revise more than 30% of their annual budget on average. This leads to increased inventory costs and lost opportunities due to incorrect forecasts, which significantly hinders a company's profitability.

In this article, we will introduce the limitations of sales volume forecasting models that threaten the accuracy of purchasing budget formulation and the solution to overcome them, Deepflow.

‍

‍

Key factors that the existing purchasing budget formulation method overlooks

Structural changes in the industry and changes in the competitive environment

As digital transformation accelerates, the ecosystem across industries is rapidly being reorganized. The boundaries of existing industries are breaking down, and new types of competitors are constantly emerging, changing the market landscape in unpredictable directions.

The traditional method of setting purchase budgets makes it difficult to reflect such structural changes.

In the case of the domestic manufacturing industry, unlike in the past when competitors in the same industry only had to consider new product launches or changes in pricing policies, now they have to consider the entry of companies in different industries and the participation of platform companies in the market.

In fact, the scope of competitor analysis by major manufacturers has expanded by an average of 2.5 times over the past three years.

The business models of new entrants also look completely different from the past. The spread of the subscription economy has increased the number of regular customers, and the growth of direct-to-consumer (D2C) brands has diversified distribution channels, greatly increasing the complexity of demand forecasting.

According to a McKinsey report, this change has led to an average 15% increase in the annual demand forecast error of companies. In particular, the entry of digital native companies into the market is a major challenge to the forecasting models of existing companies.

These companies, which use data-based decision-making and agile supply chain management as their weapons, are quickly responding to market changes and rapidly encroaching on the market share of companies that use traditional forecasting methods.

‍

Uncertainty in predictions due to changes in consumer behavior patterns

The rapid development of the digital environment has fundamentally changed the purchasing decision-making process of consumers. In the past, consumers' purchasing behavior, which had shown relatively consistent patterns, has become difficult to predict due to real-time information and the influence of social media.

In particular, the rise of the MZ generation is further accelerating the pace of change.

According to the latest survey results from the Korea Consumer Agency, 65% of consumers consult at least three information channels before making a purchase decision, and 82% of respondents said they are influenced by social media or online reviews.

What is noteworthy is that the time it takes to make a final purchase decision has been reduced by an average of 40% compared to two years ago. As a result, predictions based on past purchase cycles or decision-making patterns are no longer effective.

It is also worth noting the tendency of modern consumers to have an 'immediate reaction'. Products that become a hot topic on social media often generate sales of several months' worth in a single day. On the other hand, when negative issues arise, sales volume drops sharply.

It is difficult to capture such rapid volatility with traditional demand forecasting models.

In addition, omni-channel shopping behavior that crosses online and offline channels is further increasing uncertainty. Consumers have complex purchasing journeys, such as checking products in offline stores before purchasing online or collecting information online before purchasing offline.

The free movement between channels is creating a new level of complexity that is difficult to grasp with single-channel sales data.

‍

Global supply chain risks and volatility in raw material prices

Instability in the global supply chain has emerged as a new variable in the formulation of purchasing budgets. The international situation, which has become more complex since the COVID-19 pandemic, has intensified the volatility of raw material supply and prices, and geopolitical risks are adding to the uncertainty across the supply chain.

According to the World Bank's analysis, the global supply chain risk index in 2023 has risen 2.3 times compared to 2019.

The volatility of the raw materials market is particularly severe. Since the pandemic, the price fluctuations of major raw materials have increased to an average of 45%, and in some cases, the annual fluctuations have exceeded 100%.

It has become difficult to respond to such rapid price fluctuations by simply establishing a budget based on the past purchase price.

Supply chain diversification strategies also complicate forecasting. Companies are diversifying their suppliers to spread risk, but suppliers in different regions have different lead times and pricing structures.

In addition, the regulatory environment and logistics conditions in each region are different, so it is necessary to establish a purchasing budget that takes these factors into account.

Rapid fluctuations in logistics costs are also a notable variable. Looking at the global logistics index, there are frequent cases where shipping rates have skyrocketed by more than 30% in a month.

In particular, whenever geopolitical risks on major sea routes increase, transportation costs rise to unpredictable levels, which has a significant impact on overall purchasing budgets.

‍

‍

Blind spots in sales volume forecasts based on historical data

Limitations in forecasting when entering new markets and launching new products

Entering new markets and launching new products, which are key to a company's growth strategy, are the biggest blind spots in historical data-based forecasting.

In the absence or lack of reference historical data, existing forecasting methods are virtually neutralized. According to a report by Nielsen, a market research firm, the accuracy of sales volume forecasts when launching new products is only 40% on average.

When entering a new market, there is a problem of not being able to take into account the particularities of the target market. Even though the purchasing patterns, cultural characteristics, and competitive landscape of local consumers may be completely different from those of the existing market, many companies try to apply the existing market data as is.

In fact, when analyzing the cases of global consumer goods companies entering overseas markets, the error rate of sales forecasts in the first year is above 60% on average.

The situation is even more complicated when it comes to launching new products. The more innovative the product, the more difficult it is to predict the market response and the impact on sales of the existing product line.

In particular, it is often the case that inventory management of the entire product line is disrupted due to an inability to accurately predict the carnivalization effect.

‍

Lack of responsiveness to seasonality and special circumstances

In the case of products with strong seasonality, it is difficult to accurately predict fluctuations in demand due to changes in weather or consumption trends based on historical data alone. According to the Korea Meteorological Administration, the start and end of the seasons have been shifting in the past five years, and abnormal weather events have been occurring frequently.

Special seasons such as holidays and Black Friday also make it more difficult to predict. The growth of online shopping has dispersed consumers' buying times, and the intensifying competition for promotions has significantly changed the demand patterns of traditional special seasons.

The influence of major events and social issues cannot be overlooked. In the past, these variables had only a limited impact, but the development of social media has greatly expanded their reach.

What is particularly noteworthy is the fact that these special situations often interact with each other. For example, if an abnormal climate coincides with the holiday season or an unexpected social issue arises during a major event, it is almost impossible to calculate the impact of these complex situations on demand using existing forecasting models.

‍

‍

An innovative approach to Deepflow-based purchase budgeting

Sophisticated demand forecasting through multidimensional data integration

In the procurement budget planning of modern companies, integrated analysis of multidimensional data is essential. Deepflow collects and analyzes not only internal data such as ERP data, but also more than 600,000 external data in real time.

Deepflow has been able to dramatically improve the accuracy of its predictions by comprehensively analyzing a vast amount of data, including weather data, market environment data, social media trends, and competitor trends.

What is particularly noteworthy is the automation of the data integration process. In the past, data collection and refinement took a lot of time and manpower, but AI-based systems automate this process, enabling real-time analysis.

The AI system also automatically identifies the importance of each data source and adjusts the weighting. For products with strong seasonality, weather data is given a high weight, while for trend-sensitive products, the influence of social media data is strengthened.

This dynamic weighting system can continuously self-learn and evolve according to market conditions.

‍

AI-based advanced forecasting system

Deepflow adopts an ensemble approach that utilizes more than 200 predictive models simultaneously. It analyzes data from multiple angles using the unique characteristics of each model and derives comprehensive results.

The system's self-learning ability is also innovative. It continuously compares the predicted values with actual data to analyze the causes of errors and improves the model itself.

Innovative results are also being seen in the field of long-term forecasting. While it was difficult to predict the future beyond three months using traditional methods, Deepflow maintains excellent reliability in long-term forecasting, enabling companies to more robustly establish their mid- to long-term strategies and investment plans.

‍

Deriving the optimal purchase timing considering fluctuations in raw material prices

High volatility in the raw materials market is a key factor that directly affects a company's profitability.

Deepflow collects real-time data from the global raw materials market and analyzes price fluctuation patterns to suggest the optimal time to purchase. It builds a sophisticated price prediction model by taking into account a complex combination of economic indicators, exchange rates, and logistics data from around the world.

It also optimizes the order quantity by taking into account the inventory level and lead time of each company. In particular, for raw materials with volatile prices, the system maximizes cost savings by suggesting preemptive or split purchase strategies according to market conditions.

It also enables significant changes in supply chain risk management. The system monitors the situation of suppliers around the world and detects potential risks early.

If a potential supply disruption or price spike is detected, the system immediately recommends an alternative supply line and, if necessary, establishes an emergency purchase plan.

In addition, the system provides various scenario planning. It derives the optimal purchasing strategy through simulations that take into account various variables, such as changes in raw material prices, exchange rate fluctuations, and rising logistics costs. This enables companies to secure stable profitability even in a highly uncertain market environment.

The method of establishing purchase budgets that relied solely on past data has reached its limits. Rapidly changing market environments, changes in consumer behavior, and uncertainties in the global supply chain have become challenges that cannot be addressed using the old method.

Deepflow presents an innovative approach to solving these challenges. From AI technology that analyzes more than 600,000 pieces of external data in real time, to sophisticated demand forecasting using more than 200 predictive models, to deriving optimal purchase timing that takes into account fluctuations in raw material prices, it transforms the process of establishing a company's purchase budget into a more scientific and reliable process.

What is particularly noteworthy is the tangible results of the companies that have adopted the solution. They are achieving more than 85% forecast accuracy, reducing inventory management costs by 20%, and decreasing their out-of-stock rate by 45%, among other tangible business results.

This shows that the solution is providing real value that goes beyond mere technological innovation and leads to enhanced corporate profitability and competitiveness.

In the face of increasing uncertainty, accurate demand forecasting and efficient purchase budget planning will become core competencies that are directly linked to the survival of companies. Establish a more scientific and reliable purchase budget planning process with Deepflow.