Declining Demand: Ist this the End of the Product Life Cycle?
May 6, 2025
A product’s sales curves are trending downwards – but what’s behind it? Does this mean the item is entering its phase-out stage and will soon disappear from the market? Or is demand only temporarily suppressed by other factors?
Figure 1: Sharp decline in demand over the past two years
Figure 2: Temporary decline followed by renewed increase in demand
With the right amount of context and an agile forecasting approach, it can be ensured that a product doesn’t “retire” prematurely or, conversely, that an end-of-life model isn’t held onto unnecessarily long. One thing is certain: Simply labeling an item as “phasing out” doesn’t create added value. Only by understanding the connections and background – the right context – can demand forecasting become a real success factor. Below, we take a look at how to recognize phase-out periods, why forecasts without context can easily be misleading, and how companies can manage product discontinuation effectively.
Contents
The phase-out stage in the product life cycle: timely recognition and strategic action through forecasts
The phase-out stage, the final stage in the product life cycle, is characterized by shrinking sales, increasingly irregular demand, and ultimately a significant drop-off. Concurrently, revenue and profitability often decline until the product is withdrawn from the market. It is crucial to recognize this phase early and manage it actively. Only then can companies take timely measures, such as reducing repeat production, strategically depleting inventory, or adjusting marketing and pricing strategies to optimally leverage remaining purchasing interest.
To master this critical phase, precise demand forecasts are essential. They estimate future customer demand based on historical data and form the basis for efficient planning across the entire value chain.
Technology forecaster Paul Saffo summarized the purpose of forecasts as follows:
“The goal of forecasting is not to predict the future. Rather, forecasting tells you what you need to know to act meaningfully in the present.”
This is exactly what is needed in phase-out management: making informed decisions in the here and now.
Various approaches are used for forecasting: Besides qualitative methods, which rely on expert knowledge, market assessments, or sales experience, there are quantitative methods. These use mathematical-statistical models or increasingly methods of Machine Learning and Artificial Intelligence (AI) to identify patterns and trends in past figures and extrapolate them into the future. AI-based approaches are particularly valuable here. They can better model complex, even non-linear relationships and consider many additional data sources (e.g., Google Trends, social media, weather) as well as large amounts of data (Big Data) – provided data quality and availability are adequate. Nevertheless, even these purely algorithmic approaches have limitations, especially when it comes to capturing the specific reasons for a demand decline (e.g., new competitors, strategic decisions, external shocks) or modeling the particular dynamics of a phase-out stage, often with missing or no longer representative historical data.
Therefore, the decisive added value for proactive management arises from the intelligent interplay between human and machine. Data-driven forecasting systems provide valuable insights and important clues about possible causes and potential warning signals. However, when domain experts supplement these with their domain and market knowledge, the greatest possible benefit can be achieved. It is crucial that the forecast results are transparent and understandable, and that a validation process with correction options is supported. Only in this way can the team gain confidence in the figures and make informed decisions. This human-validated and interpreted view of demand ultimately forms the foundation for economically sound phase-out management that minimizes unnecessary costs and ensures a smooth transition – perhaps to a successor product.
We will now look at what this can look like in practice and which factors can play a role in the following sections.
Reasons for a demand decline: end of product life cycle or other factors?
Not every decline in demand means that a product has reached its natural end of life. A true phase-out occurs when demand permanently decreases for product-related reasons: the product is technically outdated, has reached a saturation point, or is being displaced by competing products. A classic example from the tech industry is the rapid decline in DVD demand after 2006, triggered by new technologies like streaming and digital download options. Here, a previously popular product became virtually obsolete due to an innovation leap. In such cases, the demand decline is a long-term structural trend – the product is phasing out.
Besides this, there are numerous factors that can temporarily or uniquely depress demand without the product itself being at the end of its cycle. These include, for example:
Political factors:
Demand can be negatively influenced by political developments and decisions. New EU regulations, for instance, can lead to the promotion of particularly sustainable products, while other products, perhaps banned at a future date, are already curbed early on by regulations.
Macroeconomic factors:
Overall economic conditions such as recession or high inflation can negatively affect demand. A deterioration of the general economic framework, such as high inflation, rising interest rates, or growing job insecurity, often leads to general consumer caution. Faced with financial uncertainty or declining purchasing power, households tend to scrutinize their spending more critically and postpone planned purchases, especially for non-essential or durable goods.
Price changes:
Often, a price increase by the company or a delayed price reduction when the market level or competitor prices are falling leads to a decrease in demand. Sometimes even small price increases can be followed by large slumps if competitors offer a similar and cheaper option.
Legal regulations:
The use of some products is linked to specific regulations, for example, a prescribed fixed end to the usage period. In the agricultural sector, the use of certain fertilizers is limited by fixed time cycles, and after the window expires, another product must be used. This causes a demand decline for the respective chemical companies.
Seasonal and weather effects:
In the consumer goods industry, seasonal (and related weather-dependent) demand fluctuations are a dominant factor. Many products show large swings depending on the season or occasion. A classic example is the consumption of ice cream, which sells in huge quantities in the summer, while demand drops sharply in the winter – but of course, the product is not phasing out, it’s just taking a seasonal break.
Supply bottlenecks or logistical problems:
Sometimes the trigger lies on the supply side, when global supply chain disruptions, raw material shortages, or logistical bottlenecks mean that products are unavailable or only available with delay, despite high customer demand. In recent years, bottlenecks in semiconductors (computer chips) have increasingly occurred, severely limiting the production and delivery of cars or game consoles, for example.
Customer preferences & trends:
Changing tastes, needs, or societal trends among consumers can reduce demand for a product or initiate its phase-out. If a product no longer meets current expectations – for example, due to changing fashion and lifestyle trends – demand can decline structurally.
Internal factors:
Decisions and actions within the company itself can negatively impact sales figures, regardless of external market conditions or seasonality. Strategic decisions like cutting marketing budgets or shifting focus away from a particular product (e.g., in favor of a successor) can cause a sales decline. But even basic things like a company holiday can have an influence.
Sometimes a demand decline cannot be explained – neither by external or internal factors nor by the natural phasing out of the product. In such cases, one might assume that with a prolonged period of low demand, the probability of a subsequent increase in demand rises.
Typical mistakes in handling low demand
Crucially: Is it a permanent decline in demand signaling the end of the product, or just a temporary slump? And what causes might be behind the demand decline in each case? A common mistake is to prematurely interpret a short-term sales decline as a signal for the product’s phase-out. Such a misjudgment can be costly: If demand unexpectedly picks up again – for example, after an externally caused pause – the company might find itself without sufficient inventory or production capacity and lose sales. At the same time, a typical error is always attributing low demand to sales or pricing, even though the market might have long been saturated. Blindly holding onto formerly high-revenue products can tie up resources that would be better used elsewhere. Such errors in dealing with demand weaknesses can often be traced back to a lack of context, as figures are viewed in isolation without asking about the causes. Therefore, a root cause analysis should always be carried out for noticeable declines before strategic decisions are made. Only in this way can incorrect decisions – whether it be writing off the product too early or holding onto it for too long, for example – be avoided. This is best achieved through the clever use of data-based approaches, combined with human experience and the intuitive foresight of an experienced planner.
Root cause analysis for demand changes
How can companies recognize in practice whether a product is actually phasing out?
→ The key lies in the context-aware analysis of all available information.
First, the sales history should be examined in detail: Is there, for example, a sudden trend break, level shift, or a gradual trend change? A gradual, long-term downward trend, for example, is more consistent with the end of the product life cycle:
Figure 3: Gradual decline in demand in a smooth time series as a possible indication of the end of the product life cycle
An abrupt decline, on the other hand, suggests a one-off event (e.g., loss of a major customer, delivery problem, unique special effect, e.g., due to an awkward pricing strategy):
Figure 4: Abrupt decline in demand in the form of a level shift in a smooth time series
Dampened trends, where a positive upward trend increasingly flattens, stagnates, or turns into a downward trend, can indicate market saturation for this product. Recognizing patterns like gradual trends or abrupt changes presents different challenges depending on the time series type and other data characteristics. Even with smooth patterns, as seen in the last two figures 3 and 4, it can be difficult to distinguish trends from random noise and accordingly recognize a change. For sporadic time series, as in figures 5 and 6, the frequent zero values can complicate the identification of an underlying pattern and require specific analysis methods to reliably detect changes in trend behavior based on data.
Figure 5: Slow decline in demand in a sporadic time series as a possible indication of the end of the product life cycle
Figure 6: Abrupt decline in demand for a sporadic product
Furthermore, a comparison with similar products or markets should be made: If demand collapses only for this one product, but the specific market overall remains stable, this points to a product-specific problem (possibly poor quality or similar). If, however, the entire market or demand in all regions declines, an external factor could be at play. A comprehensive analysis of the market environment, including competitors and customer trends, can provide clues, e.g., whether a competitor recently launched an alternative or whether customer preferences are changing. Because customer feedback also plays a role – negative reviews or lack of online interest can explain sales development. Some companies use web analytics tools to capture and evaluate, e.g., Google search trends, social media mentions, or webshop click rates.
So, if unusual sales slumps are detected, one should always investigate the cause and conduct a comprehensive analysis with an interdisciplinary team (Forecasting, Sales, Marketing, Supply Chain) to take the right measures and not prematurely adjust the forecast downwards. Here too, a good forecasting system that considers all potential influencing factors can ensure early detection and timely action.
Warning systems for demand changes
A central element and practical tool is the establishment of early warning systems that draw attention to demand declines or increases. Such a system continuously monitors forecasts and detects emerging downward trends early on. Statistical Process Control (SPC) concepts are often applied to forecasts for the technical implementation of such warning systems, or Machine Learning models for anomaly detection are used. The latter learn what “normal” demand behavior is and raise an alarm when demand falls outside these normal patterns. Even simple approaches like comparing current demand with that of the previous year can provide valuable information. An important factor here is that such early warning systems consider the trend development of the time series and sound an alarm if a trend shift, level shift, or other structural break occurs.
Integration of context and early warning system
More complex systems can additionally consider potential influencing factors to form a particularly good basis for decision-making.
The influence of price changes on time series
The following figures show how a price increase, for example, can affect demand. Such a development can evolve constantly over a longer period or lead to a real level shift:
Figure 7: Negative trend in demand due to a continuous price increase
If one only considers the highlighted area of the demand plot in the period mid-2018 to 2020, its development can easily be misinterpreted as the product phasing out without the given context of the price increase. Furthermore, it becomes clear from the further course of time from mid-2021 that the demand decline is reversible by adjusting the price.
In the next example, the demand decline triggered by the price change is also shown, although here it does not lead to a slow decrease, but to a sudden level shift. This obviously too strong price increase may have led, for example, to important customers migrating to the competition, only returning after the price was lowered.
Figure 8: Sudden drop and subsequent increase in demand due to price changes
In both examples, the respective company recognized the negative development far too late. With a data-based analysis considering all potential influencing factors, they could have reacted early and averted the development.
The influence of macroeconomic factors on time series
Price is a very important, but only one of many potential influencing factors, as we saw earlier in the section “Reasons for a demand decline”. We see a significant external factor in the following illustration: the macroeconomic environment. Identifying and considering the most important business indicators for one’s own business, such as the Consumer Confidence Index (EU), can provide a decisive advantage, especially in periods of trend reversals. As can be seen in the graph, the index has a lead time of three months and can thus point to turnarounds with this time advantage, even if they are not yet visible in one’s own history. Thus, a company that has integrated the crucial indicators into its forecasting system can react in time and, for example, expand production capacities, procure materials, or hire more staff, or in the opposite case, postpone a planned investment in new warehouses or production facilities.
Figure 9: Influence of macroeconomic indicators on demand
Of course, such effects are not always as clearly visible in demand as in the examples shown. A visual analysis can often contribute to gaining insights in the root cause analysis, but especially with more complex relationships, such as the interplay of several factors, these can no longer be easily captured visually.
We have now seen that gradual declines or changes do not always predict the phasing out of the product, but can also be caused by a multitude of other factors.
Demand changes in sporadic time series
Let’s now look at two sporadic time series examples.
Figure 10: End of the product life cycle
Figure 11: Temporary decline in demand
The left graph shows a real phasing-out, provided that the price is constant, all other relevant factors have been analyzed and excluded, and it is known that the product shown can indeed tend towards a natural phase-out or has reached this point in the product life cycle. Early detection and prediction of very low requirements (or identically zero) is elementary here so that production or purchasing can react accordingly in time and a sell-off with targeted promotions, etc., is initiated before even a low price no longer induces anyone to buy.
In contrast, the figure on the right shows a clear decline in demand due to the price being raised too sharply at the beginning of 2020. After the price was corrected downwards in 2022, demand picks up again with a slight delay. In this example too, the connection between the collapsed demand and the raised price was only discovered and corrected after almost 2 years, which could have been avoided with an implemented warning system.
Demand is changing: how to optimize responsiveness
For ongoing planning control, companies can, for example, create overviews to enable quick monitoring with key figures on forecast accuracy and demand development. Additionally, an implemented notification system for unexpected deviations, which informs the responsible Demand Planners or S&OP teams in time, supports rapid responsiveness. Holistic warning systems monitor both internal and external factors. Due to the rapid advances in Large Language Models or Generative AI, systems are now also possible that, for example, monitor news feeds and provide hints for messages like “Competitor launches new product” or “Regulations are changing” that could be related to sales. This type of tool also helps to react to demand declines in time. It is crucial to understand that the function of an alarm system is not to make decisions, but to provide a basis for decisions. (→ alert, then diagnosis). Overall: context-aware forecasts require not only the initial consideration of much data but also ongoing monitoring of demand and intelligent alerting to react dynamically and informedly before it’s too late.
Conclusion: context for your forecasts with futureEXPERT
Demand forecasts always need the right context. Only with the relevant background knowledge about why a demand trend is rising or falling can companies distinguish whether a product is actually phasing out or whether the decline is only temporary. Through holistic approaches that connect data, people, and processes, forecasts can be significantly improved and the right measures taken in time – so that costly mistakes in terms of shortages or overstocks are avoided. In a world of fast-moving markets, this contextual view of forecasting becomes a competitive advantage more than ever.
Context is crucial – futureEXPERT delivers it for your forecasts. Through the integrated POOL, you have a range of potential external influencing factors directly available, and the MATCHER identifies the influencing factors that best fit your data.
Get to know futureEXPERT now and create your own demand forecasts.
To start directly with your demand predictions, we provide a ready-to-use template for implementation with futureEXPERT and more information on Demand Forecasting!