Is predictive shopping about to take over retail?

One of the first real-world developments to be born from AI and machine learning is likely to be predictive shopping, but will it catch on?

1895

“Five years back there was a massive discussion around whether having stores is a source of competitive advantage or a liability. There was a big debate around if there is any space for brick and mortar infrastructure in the new retail landscape. Where we are right now, there is a generally accepted answer.”

Although, according to Adobe senior director Vijayanta Gupta, the industry has stopped squabbling a come to the conclusion that there is room if used correctly, a more pertinent question has since been raised: what comes next?

The next five years are set to be an exciting time for retail. It is undergoing an unprecedented level of evolution at a rate even the most experienced in the industry are struggling to keep up with.

This rapid transformation has been driven by huge leaps in technological capability, and this very technology is set to dictate the shape of the industry in 2023.

“Big data”, “artificial intelligence” and “machine learning” are fast becoming overused buzzwords, but little is said about how this frighteningly complex subject will affect the way we shop day-to-day.

One of the first real-world developments to be born from this virtual process is likely to be predictive shopping.

In theory, this means that through analysing your purchase history, online demand and stock levels, items you need will appear on your doorstep without you ever having to think about it.

“The idea of predictive shopping (or programmatic commerce, as we have termed it) is not as far off as you might think,” Salmon’s global head of innovation Hugh Fletcher told the Retail Gazette.

“Our research found that while 10 per cent of shoppers were prepared to use fully automated devices in 2016, this rocketed up to 46 per cent last year.

“It can only be a matter of time before similar capability is available for online grocery shopping and home delivery”

“And the reasons are simple. Consumers’ attitudes to shopping – and the way they shop – is rapidly changing. Their loyalty to the brand is diminishing in the face of convenience and speed of delivery, in no small part due to the ‘Amazon Prime effect’ which is dramatically changing the retail model.”

Iterations of this technology are beginning to appear already. Amazon’s Dash buttons largely pioneered the concept of one-click ordering, and Samsung’s Family Hub Fridge took this a step further offering wireless connectivity.

No-click ordering is also becoming more common thanks to the rising popularity of Alexa, Siri and Google Assistant.

Former Halfords head of logistics and current BearingPoint director Stuart Higgins said the concept has already evolved past these initial trials.

“Predictive shopping is already with us – with many online retailers already providing customers with options to subscribe to regular deliveries of consumable items for everything from toothbrush heads to dogfood,” he said.

“At the same time, luxury retail already uses their relationships with high net worth customers to re-order consumable items, such as high-end cosmetics and skincare products, seamlessly as part of their clienteling services.

“Predictive ordering is also starting to emerge in the restaurant industry where a number of platforms are already offering predictive ordering solutions, tying together weather and major events with historical data from their existing restaurant POS systems to predict customer footfall, help restaurants prevent shortages, and avoid over-buying of perishable foods.

“It can only be a matter of time before similar capability is available for online grocery shopping and home delivery.”

So, predictive shopping is absolutely possible within today’s technological landscape. However, being physically possible and being the next major shift in retail are two very different things.

Removing the process of actually having to shop from retail is an incredibly problematic concept, and one that the industry itself may be very opposed to adopting.

Global digital retail marketing spend is predicted to hit $362.11 billion (£270 billion) by 2020, predictive shopping would largely make marketing a redundant concept.

As Fletcher explained: “This kind of future would start to remove brands from the equation because an AI is making purchasing decisions on behalf of the consumer.

“If they don’t specify a brand of dish cleaner, then the AI will make the decision for that consumer based on their preferences, not on what ad just showed up on TV.”

But the key obstacle for this technology is the fact that people, for the most part, enjoy shopping.

Despite the notable shift towards experience and leisure among millennials, the number of hours people spend transfixed by their smartphone everyday is still rising, and an increasing portion of that time is spent shopping.

Were this concept to be adopted over the next half-decade, it is likely to only take hold in a handful of sectors.

“This kind of future would start to remove brands from the equation because an AI is making purchasing decisions on behalf of the consumer”

Pragma’s associate director Jacob Gascoigne-Becker believes any categories where there is an emotional reaction to making a purchase will reject predictive shopping.

“The race to perfect predictive shopping aims to deliver the functional benefit of convenience in exchange for emotional process of exploration and choice,” he said.

“This means it’s best suited to categories like groceries and consumables, where the excitement that comes with the shopping process, and the emotional benefits received from the actual products, are pretty limited.

“We believe that reducing shopping to a purely functional, benefits driven exercise, will only work for a limited number of categories.

“In fact, for others, we’re seeing a radical shift in how shoppers graze and consume products, facilitated by new business models. Fashion is a great example, with Zara’s success hinging on its strategy of stocking less merchandise, and frequently updating its ranges to encourage visit frequency and exploration.”

Predictive shopping is a tantalising idea seemingly plucked straight from science fiction. However, as with so many exciting concepts which are technically possible, its implementation into the real world is unlikely any time soon if at all.

Although it promises to remove the monotony of buying the same fast-moving consumer goods repeatedly, it fundamentally challenges notions of how we shop.

Moreover, it threatens the way in which various multibillion dollar industries are structured, and they are not likely to encourage its adoption.

Though its implementation is undoubtedly inevitable in limited applications over the next five years, even the rapidly changing industry will be hard pushed to adopt such radical change in that time.

Click here to sign up to Retail Gazette‘s free daily email newsletter