By Michael Feindt, Strategic Adviser, Blue Yonder

AI and data: How can retailers target the right customers, both through Covid-19 disruption and into the future?

The Covid-19 pandemic has caused major disruption in the UK retail sector, with many shops being forced to close their stores or scale back their operations. In fact, every sector – even food and drink – reported an annual fall in sales, and outlets selling clothing and footwear, furniture and carpets and recreational goods were hit especially hard. Retailers therefore need to get back up to speed, and quickly, to ensure they can keep their head above water.


Doing more with less


However, as budgets are likely to have taken a hit, many retailers are having to do more with less. Retailers recognise that technology holds the key in this scenario – with a recent report revealing that technology is at the forefront of retail strategies for the rest of 2020. For all the focus on getting the shop doors back open, more retailers indicated they would be investing in back-end systems rather than updating stores.


AI is a particular area of interest for many of them: according to the 2019 Data Science and Machine Learning Market Study, a quarter of retailers already say data science, artificial intelligence (AI) and machine learning are critical to their success. Retailers are increasingly coming to find practical applications for AI, and it could provide the extra edge needed to pull through the current period of disruption.


In particular, they are likely to feel the benefit when using AI in marketing and promotion strategies. When implementing a marketing campaign, historically the default decision for retailers is to target their most loyal customers. They’ve spent money with you already, so by sending them promotions, you’re tapping into a sure thing – surely meaning the campaign will be a success?


However, retailers don’t realise that by doing this they’re making one of the biggest and most frequent mistakes in the industry. Weren’t these people likely to shop with you anyway, and wouldn’t they have paid the full price? Targeting your most regular customers adds cost and eats into profit, when in reality you would have been better off reaching out to people who showed initial interest but shop with you less frequently. When you factor in discounts and marketing costs, the company will have lost around 20 per cent on a sale – a mistake retailers can no longer afford to make.


Using data to measure the causal effect


Simply put, rather than targeting loyal customers, retailers should be focussing efforts on engaging with more casual ones. “Causality” is a concept that excites and inflames the curiosity of scientists, mathematicians and psychologists, but escapes the marketing efforts of most retailers.


Measuring or finding causality can be tricky in practice but is possible by running A/B tests or using causal algorithms. This is where the power of data and AI can really come into its own, as retailers can use it to measure the causal effect: who your promotions will influence most effectively, and to what end. This kind of work simply can’t be calculated by human brains – you need specialist algorithms. Specifically, they need to use A/B test data as an input for machine learning algorithms, which can directly learn the influence marketing will have on sales behaviour. Once marketing campaigns have been sorted based on predicted influence, sales and profits should increase.


Of course, this represents a change from what most businesses have been doing to date. If they’ve been used to the tried and tested method of targeting current customers, it won’t be easy to persuade key decision makers to alter the approach. However, retailers should consider the cost that usually goes into marketing, and then understand the benefits a more targeted AI approach can give, pushing their marketing promotions the extra mile and helping them to increase profits.  


Not only do these algorithms evaluate the best people to contact, but also the best method to do so. Emails are cheaper, but will they be as well-received as a postal campaign? Or, if email is used too much and too many people exploit the offer, will the company lose money?


A more intelligent approach


Using AI-driven knowledge, extracted from historical data, makes it possible to identify the right type of customer to target. This will put retailers in a position to achieve optimum increases against sales that would have been made anyway, in addition to the original investment that helps initiate a campaign.


Understanding the difference between correlation and causality is necessary for retailers to increase their sales. Retail decision-makers might think they know what’s best based on past experience, but in reality, AI can help them make more informed and intelligent decisions. Placing marketing activities in the hands of a data-devoted machine will ensure more is learned, and decisions will become more precise and the probability of reaching an optimum demographic will exceed anything a human can manage.


Using AI will mean retailers can reap better results by targeting the right customers – essentially cutting the fat from marketing efforts and increasing overall profits. AI has the power to not only get retailers back up to speed in the wake of the Covid-19 crisis, but also helps them to lay the foundation for future success. Transforming marketing strategies using AI will not just serve as a short-term fix, but will set them up to make intelligent decisions long into the future.

Related articles

What's next?

Get our latest features in your inbox

Join our community of business leaders