The FMCG and consumer devices market is experiencing major transformations, fueled by a combination of various external factors like climate change, geopolitical dynamics, and the ever-evolving social landscape. These forces are driving change in consumer behavior and demands, and in turn posing significant challenges for producers. Let’s take a look at how you can unlock the next-gen FMCG appliances.
Major FMCG challenges
Social evolutions, such as smaller living, cohousing and the sharing economy, arouse the demand of consumers to follow up and allocate consumption of for example goods and power. These consumers not only want to be in control of expenditures but also increasingly expect services or products personalized to their individual or regional needs, tastes and habits. In addition, more and more consumers also want to be able to track the origin of food or goods both from an ethical and a health perspective.
To address the rapidly evolving consumer demands, producers need insights into regional and individual behaviors and preferences.
Many producers of consumer goods, on the other hand, have little to no insights into the behavior, preferences or demands of their customers. Even though they face the challenge to increase customer intimacy and their (personalized) experience to maximize retention.
And finally, the recent geopolitical evolutions put increasing pressure on capital and operational costs, both for the producers and the consumers. Think about more affordable appliances, more efficiency in the value chain, more local products and production, and less waste and energy consumption, to name a few. Moreover, the demand for more sustainable and traceable value chains rises due to the impact of climate change and evolving laws and regulations.
Overview of today’s major FMCG market trends and industry challenges ©VERHAERT
From challenges to opportunities
No need to say, many of the above challenges are conflicting, and hence demand novel, sometimes out-of-the-box solutions. That’s where new technologies, such as IoT, blockchain, AR/VR and of course AI, come in. In this blog, I’ll dive a bit deeper into how AI can render FMCG and consumer appliances smarter to address some of these challenges.
When I say AI, I mainly refer to the domain of machine learning (ML), meaning algorithms that learn (become smarter) from data such as consumer behavior, market trends, contextual and environmental information. This of course implies that this data can be gathered (by sensors, consumer feedback, external data, …), transferred, stored and processed. ML algorithms can then give insights, make predictions, advise or nudge consumers, optimize settings or automate certain activities. To bring these purposes to the consumer, your appliance should obviously be equipped with the right user interface, compute power and/or actuators.
Schematic view of a smart appliance ©VERHAERT
Trade off between cost and value
Evidently, all these functions come with a cost. Therefore, when (re)designing a consumer appliance, you should make trade-offs between cost and added value. How and where can you collect which data? How can you transfer and store this data? Where do you run the sometimes heavy computations? And how can you effectuate the outcomes? When you’re making these decisions, you shouldn’t just think of yesterday’s or today’s challenges but also the potential expectations and opportunities of tomorrow. This of course depends on the projected lifetime of the appliance (where AI can even help to increase this lifetime, for example through user nudging or timely maintenance, which I will touch on later).
When designing a new appliance, you should make trade-offs between cost and added value, taking into account today’s and tomorrow’s opportunities.
To this end, it’s even more important to explore the potential (future) user journeys during the design phase through ideation workshops and brainstorms. Hence designing an appliance that is ready to also enable new functionalities in the future, including the right sensors, computing power, communication, firmware update abilities, actuators, etc. In the remaining part of this blog, I will give some examples of AI-based solutions which can tackle some of the challenges mentioned above.
From insights to solutions
Let’s start with so-called virtual sensors. A virtual sensor is a software model, say a digital twin, replacing a physical sensor. In most cases, they have a lower cost and footprint but can serve just as well, or in some cases even better. A virtual sensor can calculate the measurements based on known laws of physics but can also be trained using machine learning, the latter opening a vast range of possibilities. A physical coffee taste sensor could be quite hard to develop. However, such a virtual sensor is conceivable, using for example bean variety and grinding, water temperature and hardness as input, and consumer experience as output, and applying machine learning to correlate these values.
Virtual (smart) sensors can not only be cost-effective, but they can also take measurements that are not even conceivable with physical sensors.
AI can also be very powerful to gain insights into individual or regional taste preferences. And consequently, always prepare and serve the ultimate perfect product under potentially varying conditions (if you have a virtual sensor that can predict the output, given the circumstances). This not only provides a great experience to the consumer but also, of course, contributes to customer intimacy and even brand lock-in. Talk about a win-win!
Another well-developed capability of AI is prediction, and so we arrive at predictive and thus optimized maintenance. This technology can nudge the user to clean the device in time or can suggest making an appointment for maintenance. Let’s take limescale in washing machines or coffeemakers as an example. The amount of limescale can be predicted quite well based on the usage and the water hardness. Call it a limescale virtual sensor. Timely cleaning or maintenance can increase the appliance’s lifetime, save costs of replacements and even save energy.
And it’s energy saving that brings us to our last topic. You can measure energy consumption itself with a physical or (again) virtual sensor. The circumstances or features can be current time and/or day, situational information, like the weather or crowdedness, ambient temperature or humidity, fill level in for example a washing machine, a fridge or a beverage keg in a dispenser, and so on. Applying machine learning to energy consumption and the circumstances, heating or cooling can be proactively adapted and users can be informed how to save energy, hence increasing sustainability, reducing costs and enabling consumers to be more in control.
Overview of AI solutions for FMCG market challenges ©VERHAERT
To conclude, AI, amongst (or in combination with) various other new technologies, can contribute in many ways to make consumer appliances smarter and help you cope with the numerous challenges the FMCG market is facing. The ultimate challenge here is to ensure new appliances are ready for the job, by well-thought, creative and future-enabled design, balancing cost and added value. And bear in mind, dumb appliances are so 2018!