Artificial Intelligence and ‘big data’ have emerged as the most powerful FMCG tools around.
That’s not hyperbole. In March 2015, the Harvard Business Review claimed organisations with above-average grasps on customer data and analytics were outperforming competitors by two or three times in all the important spots: sales, margins and profits.
FMCG’s major players were already ahead of the game - Unilever and P&G were both committed to serious data-driven strategies in 2014. P&G CEO Robert McDonald says: “It would be heretical in this company to say that data are more valuable than a brand, but it’s the data sources that help create the brand and keep it dynamic.”
Meanwhile, Unilever CMO Keith Weed has been candid about the power of big data in practice: “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.”
Target’s enhanced analytics game has been even more impressive (if a little intrusive). When the company started specifically targeting pregnant women in an effort to shift their buying habits before the blitz of post-natal marketing begins, they ended up knowing a customer’s 16-year-old daughter was pregnant before he did. That’s a bridge too far - the father in question wasn’t best pleased, and early trials resulted in a backlash from women who felt “spied on”.
The point is, FMCG brands can do so much with the data they already have. So what does the next 10 years look like for big data and AI in FMCG?
Data
FMCG brands have been collecting data for years.
Loyalty cards don’t just earn discounts; they log purchases across a customer’s spending lifespan and identify customers who aren’t engaging with the brand. They put phone numbers and addresses into the hands of brands which would otherwise struggle to collect them.
The data situation has only improved with the rise of free Wi-Fi and geolocation technology in people’s pockets. Coffee shops, supermarket cafes and retail infrastructures (a Wi-Fi network spanning a whole shopping centre, for instance) are all able to collect data and obtain consent while offering a service on (or even before) demand.
The devices they’re using, meanwhile, enable automatic ‘geofencing’. When customers with their GPS switched on enter a store, retailers can be alerted to their presence automatically and data can be marshalled to support customer service.
Data collection even takes place at home. Amazon Alexa and Google Home are, basically, data collection devices which people welcome into their homes and leave switched on all the time for convenience’s sake.
Consumers have never given away so much data so freely, and they’re going to give away more. We predict the next ten years will see more and more devices becoming more and more integrated with daily life. VR, already popular as a way to explore environments regardless of distance, will expand into retail, perhaps building on developments in heritage like the Smithsonian Museum’s virtual tours.
Wearables, already gearing up to send data directly to doctors, could soon sync with nutrition apps and, through them, with FMCG suppliers and sponsors, feeding real-time data on consumers’ nutritional needs. IoT devices could arrange deliveries of staple goods, or even predict likely purchases based on harvested information.
The challenge for FMCG and other retail firms isn’t “getting the data” but “processing, interpreting and applying the data”. That’s where AI comes in.
AI
Analytics are the belt to data collection’s braces. Processing as much data, as quickly and accurately as possible, is the key to achieving its full potential.
AI began to succeed off the back of FMCG product recommendation systems, which couldn’t tell why customers wanted to buy an item, but could identify trends in their previous purchases and make predictions based on those.
Now we can feed more and more data into artificially intelligent apps and algorithms - not just what people buy, but how long they spend in store, the order in which they look at products, what they do with the free WiFi while they’re in there, and maybe even how they use the product at home.
AI now has so much data that it almost understands customers - it certainly has enough correlating data to automatically find them the product they want as and when they want it. It even has the capacity to tailor product designs. An algorithm has created seven million different versions of Nutella’s graphic identity, based on a database of colours and layouts. Marrying that up with data collection means products can be custom-designed for more precise appeal.
In the decade to come, this enormous potential will bring about more reactive retail businesses.
Activity-detection data will allow forward-facing questions - “will this be on sale soon?” - to be predicted and answered, and decisions about promotion and offer timing to become automatic. Image recognition will shift searches away from typed words; the spelling of a brand name will be less important than a distinctive, machine-recognisable visual identity. In-store assistants will help shoppers gather information about products while navigating stores - in the flesh or in VR - down to the level of walking them through environments to the products they want.
Behind the scenes, the use of big data to inform supply, demand and manufacture decisions - already a key part of P&G’s digital domination strategy - will lead to increasingly automated production, delivery and restocking processes, with supervision and market research taking on a more supervisory role.
This does come with a challenge for brands, though. Users have preferences.
Savvy users will have set up favourites and filters to manage the messages that reach their eyes and ears. As tech becomes more screenless, brands will need to communicate their values to people who might never see their website or packaging until after point of purchase. “Brands are disadvantaged because there is no way for them to show the investments they’ve made online,” says Scott Galloway, founder and CEO of L2: “Brands will not matter at all, because of voice.”
This isn’t quite true. Brands will have to shift their approach toward soft power; aesthetics that appeal and directly address the customer’s preferences, and language that positions you in the customer’s shopping basket without forcing yourself there. The words Alexa recites back to customers when they complete an order are part of brand communication, and they’re what brands will have to control and design in order to stay relevant.
Big data and AI can’t replace design talent. The goal is to augment rather than automate, harnessing data to understand what customers want, how they feel, and what factors encourage them to buy. The middle stage in the process from ideation to execution is where data comes in. At this testing stage, aspects of the process - such as generating visuals, as with the Nutella example, and seeing how well they perform in front of human eyes - can be automated, and the designs tailored and tweaked for final release by human decision-makers.
AI and big data won’t change what we do - but they will help us do it far, far more efficiently.