The challenges of using forecasting techniques in logistics

Quantitative forecasting techniques refers to the approaches of forecasting used for examining the future trends by analysing the historical data. These forecasting techniques are applied through static methods like time series forecasting and casual forecasting (Spedding & Chan, 2010). The casual forecasting is conducted using simple or multiple regression models. On the other hand, casual forecasting is executed through the use of autoregressive moving average models. In logistics, time series forecasting focuses on analyzing the change in business strategies over a period of time. This forecasting is done using moving average and exponential smoothing which uses mathematical formulas to identify the forthcoming claim of the consumers addressed (Dombi, et al., 2018).

However, one of the major challenges of using this forecasting technique is the variation in data over a period of time. Due to this, it is difficult to examine the right patterns.

Challenges of fulfilling customer demands using forecasting techniques

The past demand and decoration of customers are equivalent to the future pattern of demand which is the primary principle of time series forecasting. So, this technique is only applicable for short term purposes that limits the outcome of the projection. However, businesses face several issues such as:

• unavailability of data,
• data assumptions,
• lack of expertise in data management,
• a rapid change in customer preferences,
• technical failure or lack of technical infrastructure and,
• numerousness among consumers (van der Laan, et al., 2016).

Lack of availability of data and data assumptions are data-based challenges which result from wrong forecasting. On the other hand, algorithm-based challenges in forecasting techniques comprise of rapid change in customer preferences that make it irrelevant (Huber, et al., 2017).

For example, a clothing company had ordered excessive designed apparels due to rapid sales. But the design can soon become outdated which result in obsolete inventory in stocks which can in turn lead to financial loss. In addition, the companies faced variable based challenges as some components of forecasting are taken as a constant .However, these components vary and affects the business process in real life.

Challenges of unavailability of relevant information

The unavailability of relevant information results in difficulties in time series model as weighted moving average and exponential smoothing is important for every factor of forecasting. The forecasting neglects the value or assumes it for final forecasting which does not matches the original market forecasting (Clements, et al., 2016). The difference between the upper and lower estimate of demands overtime is also hampered if all required data and historical information are not found. Simple projection is applied in such cases as it focuses on long term average. In these two or three parameters do not get affected, as both the factor contain equal weight.

The unreliability of information processing and information acquisition lead to the use of time-series forecasting. In order to fill the gap of unidentified data, the information is assumed for reflecting the forecasting value that differs from the practical field (Kourentzes, et al., 2017). Moreover, the analysis of external information is also required for effective forecasting techniques which improve its reliability.

Lack of technical infrastructure and technical failure has also become a threat for using time series forecasting. The forecasting values are represented with the help of software and require skills and expertise to operate (Kourentzes, et al., 2017). The value of weighting of each factor has been enhanced exponentially over time. It requires proper skill and knowledge to use technical software (Huber, et al., 2017). Trend and seasonality using enhanced forecasting method and winter’s model also focus on uncertain outcomes that include a certain percentage of deviation for forecasting value.  

Challenge of data analysis     

Businesses generally use primary data such as sales or financial information to forecast. But the wide range of customer preference and change in customer demand is also taken as a challenge for forecasting (Spedding & Chan, 2010). With the emergence of social media, the trend has transformed through the social effect that fails in terms of interpretation (Spedding & Chan, 2010).

For example, Amazon had used a customized model that works on the principles of time series forecasting of exponential smoothing and other factors such as promotions, descriptions, and sales of holidays (Amazon, 2018). Financial planning and resources planning has also become simplified with the principles of forecasting.

Lack of employees’ expertise in data management remains considered as one of the important barriers for processing the time series forecasting technique (van der Laan, et al., 2016). Lack of documentation process and casual detailing of the process disrupts the continuous flow of data. Poor handling and positioning of information set remain followed by the disruption in time series forecasting technique.

References

• Amazon, 2018. Amazon Forecast. [Online] Available at: https://aws.amazon.com/forecast/ [Accessed 25 July 2019].
• Benes, J. et al., 2015. The future of oil: Geology versus technology. International Journal of Forecasting, 1(31), pp. 207-221 DOI:10.1016/j.ijforecast.2014.03.012.
• Clements, A., Hurn, A. & Li, Z., 2016. Forecasting day-ahead electricity load using a multiple equation time series approach. European Journal of Operational Research, 2(251), pp. 522-530.
• Dombi, J., Jónás, T. & Tóth, Z., 2018. Modelling and long-term forecasting demand in spare parts logistics businesses. International Journal of Production Economics, 1(201), pp. 1-17.
• Huber, J., Gossmann, A. & Stuckenschmidt, H., 2017. Cluster-based hierarchical demand forecasting for perishable goods. Expert systems with applications, 1(76), pp. 140-151.
• Kourentzes, N., Rostami-Tabar, B. & Barrow, D., 2017. Demand forecasting by temporal aggregation: using optimal or multiple aggregation levels?. Journal of Business Research, 1(78), pp. 1-9.
• Montgomery, D., Jennings, C. & Kulahci, M., 2015. Introduction to time series analysis and forecasting. 1 ed. London: John Wiley & Sons.
• Spedding, T. & Chan, K., 2010. Forecasting demand and inventory management using Bayesian time series. Integrated Manufacturing Systems, 5(11), pp. 331-339.
• van der Laan, E., van Dalen, J., Rohrmoser, M. & Simpson, R., 2016. Demand forecasting and order planning for humanitarian logistics: An empirical assessment. Journal of Operations Management, 45(1), pp. 114-122.