There has been a lot of hype around artificial intelligence (AI) and machine learning (ML) particularly in the last decade. Similarly, we have witnessed the rise of deep learning (DL), a relatively new machine learning technique, in the last couple of years. These three terms have been used very often in the media to reveal secrets of all the magic that has been brought to our lives by recent technological developments. Self-driving cars, voice activated assistants, intelligent robots or simply “did you mean” algorithm by Google search engine are some of the recent products of these developments. While this is all by itself a very interesting topic, today we will focus on the impact of Artificial Intelligence and Big Data in demand forecasting for fast moving consumer goods (FMCG) industry.
Intelligent Demand Forecasting in FMCG
FMCG refers to the industry where products are sold with low profit margins in high volumes. These products are easily substitutable and tend to have a short shelf life. Sales and demand in the industry are easily influenced by seasons and promotions – and very quickly. Consequently, the FMCG industry generates a significant number of transactions that forms into some very large complex big data. Due to these complexities in the industry, already the challenging process of forecasting becomes even more difficult for the FMCG sector. This is where the AI (in broad terms) and big data analysis step in.
Most FMCG companies have been collecting data for many years. In addition to demographic, geographic and user preferences data; loyalty programmes, discount cards and free internet connection at sales points are also acquiring data for companies. Therefore the main challenge for the FMCG sector is to interpret the big data to understand accurately what consumers want and convert this valuable information into more profitable business operations & strategies.
Companies who want to stay ahead in the market have already started using Machine Learning technology and new Deep Learning methods to forecast and meet market demands. Let’s see how these companies benefit/can further benefit from machine learning techniques and deep learning frameworks in demand forecasting;
Real-Time Analysis & Data Interpretation
Machine learning technologies and deep learning frameworks can quickly analyse large databases which is usually a very long process with traditional forecasting methods. This enables FMCG companies to act quickly and even real-time.
Additionally, ML and DL techniques can easily interpret unstructured data in any format such as text, image, video, electronic point of sale (EPOS) data, etc. Therefore, FMCG companies can gather all the data in one pool which enables them to plan ahead accurately with help of machine learning applications.
Since FMCG industry is easily influenced by seasons and promotions; specific time of the year or a promotional activity can have a huge impact on sales and demand. Also customer behaviour is too complex to standardise and understand. Generally, it is a very difficult task with traditional forecasting methods to interpret implication of these activities correctly but ML and DL frameworks enable FMCG companies to read & simplify data. For instance, with guidance of the deep learning frameworks, companies can predict how much stock they would need on a Black Friday. They can analyse historical, behavioural, geographic, microeconomic and demographic customer data and simulate future models to predict their sales figures. This also enables companies to tap into most demanded products as well as right consumer segment and work on the less popular goods and less engaged customers.
Some leading companies in the sector have already acknowledged the importance of data analysis. In fact; Unilever’s chief marketing and communication officer Keith Weed summarised the power of big data in as “we are already able to tell a consumer when he’s walking in the park (we know his location) on a hot day (we know what the weather is like there) where the nearest place is to buy a Magnum and send him a code for a discount.”
Also, traditional forecasting methods require human interpretation and this usually poses risk of errors whereas artificial intelligence applications minimise/eliminate these errors by deploying fully automated computerised systems.
Speed is the most crucial part of FMCG industry. Goods have short shelf life, high turnover and they are very promotional sensitive. In order to meet demand of consumers, companies have to be fast. Image based shelf management systems could open a new page here. Although this task is currently carried out by manpower, we are likely to see employment of machine learning products on the shelf management systems soon.
Supply Chain and Production
In order to meet customer demand on time FMCG companies need fully integrated production lines and supply chain which covers everything from forecasting to delivery. In this way, fully integrated artificially intelligent systems can help companies to;
- Reduce purchasing and production costs
- Reduce waste and inefficiency in production
- Minimise the risk of oversupplying or undersupplying
- Improve lead times
- Increase effectiveness of marketing efforts
- Maximise profit margins
- Improve customer engagement
FMCG companies must focus on long-term advantages when it comes to deploying AI technologies in their operations. Still, some companies don’t lean towards spending money on AI technologies. However they should take into consideration that using AI is not a luxury anymore and the technology not only improves the demand forecasting process, it also reduces the operational costs, minimises the risks, have better insights on customers and improve customer loyalty.
Soon, as consumers, we will witness artificial intelligent systems offering us what we want (even if we don’t ask), communicating our needs with companies and even getting them delivered without our involvement. Likewise, companies may start soon advertising to artificial assistants instead of us. If FMCG companies want to stay competitive in this new world they will have to keep up with the AI technologies.