How do grocery retailers unlock the true value of personalisation?

Personalisation can have a huge impact on grocery retail sales, delivering what can be a one to two percent uplift in total volumes1. Key to this is “predictive personalisation”, a data-driven process whereby retailers can make informed assumptions about what shoppers will want tomorrow based on how they’re behaving today. Those predictions can then be used to deliver tailored promotions to shoppers that meet their specific needs at that point in time.

While that’s beneficial in terms of sales, the true value of predictive personalisation goes much further. While encouraging the shopper to make a purchase might be the ultimate goal, predictive personalisation can also help retailers create better, more convenient shopping experiences that drive greater satisfaction and loyalty as well.

That’s particularly important in an environment in which many retailers working tirelessly to try and differentiate themselves from the competition. Let’s look at how they can use predictive personalisation with that goal in mind.


Creating better customer experiences


LINE’s Shopping app provides a great example of why predictive personalisation can help to set a retailer apart in the eyes of the customer.

In my experience, when I sign up to receive messages from retailers through LINE, I’ll get between 10 and 15 notifications per day. Whether or not that’s too much is largely down to personal tolerance, but it’s tough to argue that this is providing a better service to the shopper. Because those messages aren’t personalised, retailers are actually asking the customer to do the hard work for them – digging through a potentially huge number of deals to find an offer that’s relevant to them.

When we combine that situation with the culture of flash deals and discounts in APAC, even good and highly-relevant promotions can get lost amongst the noise. That’s bad for the customer because they can end up missing out on a deal that they’d otherwise benefit from, but it’s arguably worse for retailers and brands who have spent the time and money to organise that promotion only to have it go unnoticed.

Predictive personalisation tackles both of those problems head-on. Rather than delivering untargeted offers like these, retailers could instead deliver uniquely personalised and highly useful deals to every one of their customers – through LINE, or via any other digital marketing channel.

That also has implications for the ever-present issue of price perception. When retailers are reliant purely on mass-market promotions, the only real way to improve price perception is to opt for deep discounts across the entire store. With predictive personalisation, they can instead move to a model in which price perception can be tailored on a shopper-by-shopper basis.

So, predictive personalisation can help retailers make shopping easier, maximise the effectiveness of their promotions, and improve price perception. But how do they know that those things are actually happening?


Measuring predictive personalisation


The critical measure for any personalisation programme is incrementality. Whatever your KPIs happen to be, you need to be able to say with some degree of certainty that the actions you took caused a customer to respond in a way that they otherwise wouldn’t have.

In terms of actual performance indicators, sales can be a great measure for this. At the same time, sales can also be broken down into numerous sub-categories, all of which help us to understand the true value of predictive personalisation.

Behavioural changes are a big focus here. Maybe a personalisation programme delivered a 15% uplift in sales, but was primarily effective with shoppers in the north of the country. Perhaps the campaign you activated didn’t resonate with price sensitive customers. As well as tracking sales figures themselves, you also need to be able to understand which customers are responding and why.

By doing so, retailers can actually start to refine their personalisation models. The more that they learn about which “types” of personalisation work for which customers, the more effective their promotion strategies become. It may be that a customer is more attracted by offers in a certain category, for instance, or that they respond better at a particular level of discount. This can provide valuable insights to marketing teams, but also helps to train AI and machine learning models on which techniques work best for individual customers.

In many ways, predictive personalisation can actually help retailers get better at evaluating the success of customer engagement programmes. Rather than relying on arbitrary KPIs like email opens or voucher redemptions, they can instead focus on more nuanced measures that reflect a broader range of actual behaviours.


Achieving balance


Any retailer putting this kind of programme into place will naturally want to ensure that they still have control over their promotional mechanisms. This is where business rules come in.

Predictive personalisation is something that needs to be an “all company” effort. Commercial teams need to be involved from the start, not just by helping to source and manage offers from suppliers, but to help set the business rules that the programme operates on. This can range from the maximum discount that a retailer is prepared to offer within a category, to the number of incentives that one customer can receive in a set period of time – all of which can also be supported and guided by data science, of course.

One of the main challenges here is that of balance. Business rules are clearly important, but they can also stifle a retailer’s personalisation programme if they exert too much control.

Rules around loyalty provide a good example of how this can happen. Very often, loyal customers are seen as being those who spend the most with a retailer. That isn’t always the case; a family of five might spend more than a single person, but that one person may also spend 80% of their grocery budget with a retailer.

One of the things that predictive personalisation is very good is actually spotting the difference between those two examples and ensuring that “good” customers don’t fall through the net. As a result, it’s important to ensure that business rules don’t filter them out.

Ultimately, unlocking the true value of predictive personalisation is about more than having the right mechanisms and controls. It also depends on a retailer’s willingness to lean in – to let the data do the work and shape the best possible experience around every individual shopper.


1. Personalizing the customer experience: Driving differentiation in retail – McKinsey, 28th April 2020

The latest insights from our experts around the world

customer first data science analytics & machine learning services
Ready to get started?

Speak to a member of our team for more information

Contact us