FOLLOW

FOLLOW

SHARE

Why Retailers Are Turning To Machine Learning

They are struggling, but machine learning could be the light at the end of the tunnel

23Aug

The environment bricks and mortars retailers operate in today is undeniably tough. Already so far this year, we have seen more retailers close their doors than in 2008, the year of the financial crisis. However, it could also be argued that rumors of the demise of brick-and-mortar are exaggerated. Online sales still don’t represent the majority of retail spending overall. The demand and need for physical stores is absolutely there, and they have a critical role to play in an omnichannel world. Indeed, the importance of physical stores has been demonstrated by Amazon’s decision to buy Wholefoods, with the move seeming to suggest the king of e-commerce believes physical stores still have a major part to play.

What is changing is the way that people shop. For instance, many consumers now research a product online before buying in store - 58% of people do this for non-grocery good while 78% say they use multiple channels to make a single purchase. Google has also found that there has been a 40-45% increase in local shop inquiries through their Maps app over the past year, with three in four of these people actually choosing to go to their destination.

So technology, rather than destroy the physical store, is in many ways helping it. Retailers are simply having to adjust rapidly to incorporate it into their strategies. Those who don’t are soon finding their stores financially unviable. Among the most important of the technologies they are looking at are machine learning and AI, which are already proving to be game changers across the sector.

According to one Gartner report, 85% of customer interactions will be managed by AI in retail in 2020. According to another Persado study, 45% of retailers plan to start using AI in the next three years. The technology has many benefits, and it is already being used to solve a number of issues that should see them continue to compete with the online-only rivals.

The primary application for machine learning at the moment is analyzing the wealth of big data that all retailers now collect. The digital trail left by customer interactions throughout their journey to purchase means data is being collected around almost every facet of operations, including products, prices, sales performance, costs, availability, logistical activities and consumer behavior. In research conducted by Accenture Analytics, 58% of retailers described big data as ‘extremely important’ to their organizations, while 36% called it ‘important’. Additionally, 70% said that big data is necessary to maintain competitiveness, and 82% agreed that big data is changing how they interact with and relate to customers. And they are right to place such a high value on it. Another recent study by McKinsey found that ‘US retailer supply chain operations who have adopted data and analytics have seen up to a 19% increase in operating margin over the last five years.’ The majority of retailers are now highly adept at collecting information from their consumers, particularly as IoT sensors now allow them to monitor their behavior in-store.

The real challenge for retailers today is crunching this data. This requires machine learning algorithms, as it is impossible for human beings to consume and analyze such a high quantity of data. This analysis enables retailers to look into the future, to better understand demand so they can better understand inventory levels, where products should be placed in store so that they will sell, what marketing campaigns will be most effective, and myriad others way.

The supply chain is another area that will be revolutionized thanks to machine learning. In Gartner’s report ‘Predicts 2016: Reimagine SCP Capabilities to Survive,’ the research firm said that supply chain organizations they had spoken to expected the level of machine automation in their supply chain processes to double in the next five years. The level of automation that supply chain managers could, theoretically, embrace, will completely revolutionize the speed and accuracy at which they can operate. Traditional models for handling inventory can be used in conjunction with sophisticated algorithms to increase the speed of computation, and it will likely augment this process by generating new features to run such models on. Autonomous vehicles will also mean upheaval in logistics and shipping. For retailers, this all means that they can not just predict spikes in demand and adjust supply, it means they can react far more quickly to any shocks. One of the major stumbling blocks for physical retailers in the past has been that if something a customer wants is not on the shelf, they will go elsewhere to get it - likely online - and they will not come back in the future.

The in-store experience is vital and remains a key reason many still prefer the high street to the computer. Machine learning will enable chatbots, that should prove a real boon for this. In a 2016 TechEmergence survey of AI executives and startup founders, 37% said virtual agents and chatbots were the AI applications most likely to take off in the next five years. Apple’s Siri, Microsoft’s Cortana, Google’s OK Google, and Amazon’s Echo services are already far advanced relative to where they were a couple of years ago, and developments made in speech analytics and natural-language processing means that they are getting better all the time. Even business app giant Oracle is creating chatbots for its apps. Market research firm, TMA Associates, estimates that the chatbot and digital assistant market will reach $600 billion by 2020 as a result of such conversational user interfaces. This has huge implications for the nature of customer support. According to IBM, 65% of millennials prefer interacting with bots to talking to live agents, and as we get more accustomed to it, this number will only go up. Automated customer service could mean an end to waiting in line and therefore happier customers.

In retail, Target has already partnered with startup AddStructure to develop an Alexa-like assistant, which a Target spokeswoman told the Chicago Tribune ‘works with customers the way they naturally talk and the way they search for things. Anything we can use to create a better experience for our guests, that’s what is most appealing to us.’

Machine learning has been used successfully in e-commerce for a number of years. Amazon, for one, see 55% of sales come from personal recommendations made by machine learning algorithms. Such algorithms are now being used in shops, as stores integrate tablets to essentially enable the online experience alongside physically being able to touch the products themselves, particularly in fashion and make-up where customers can get a better idea of how a certain product will look. Machine learning’s move offline has already seen some significant successes among physical retailers. Target Corporation has achieved 15 to 30% growth in revenue with the help of machine learning predictive models, and where there is success, others will follow. And they must - quickly. The technology is advancing so rapidly that those who take a ‘wait and see’ approach will be left burdened with an outdated business model. In an industry as competitive and fast-moving as retail, this is simply not an option. 

Comments

comments powered byDisqus
Science small

Read next:

​So, You Think You’re Cut Out For Doing Data Science?

i