Why localised assortment is now the key to loyalty and growth
In late 2022, Gartner published a market guide to assortment management applications. In it, the consulting firm clearly noted that “the traditional approach to building assortments through ‘clustering’ is being challenged by customers’ demands for more ‘local’ site-specific assortments1”.
Gartner wasn’t wrong in that assertion. Since the release of that paper, a growing number of grocery retailers have started to move towards localised assortments. In the US, for instance, Kohl’s, Kroger, and Dollar Tree have all announced localisation plans2. In the UK, meanwhile, Tesco is developing a “range optimisation tool which automates bespoke product selection based on store location and demographics3”.
Why the trend towards localisation, then? While there are several factors at play here, this is an undoubtedly customer-centric trend. Creating a range tailored to the needs of specific customers doesn’t only help to make shopping faster, easier, and more convenient; it also reintroduces the element of surprise and delight that’s so important from the perspectives of engagement and inspiration, too.
Right now, that’s a valuable capability to have. In our own Consumer Pulse studies, we’ve seen that customer behaviours have changed considerably, with many beginning to prioritise value4. That many discount retailers have enjoyed incredible growth at the same time is no coincidence5. Against that more competitive backdrop, localised assortment can serve as a valuable differentiator for mainstream retailers.
At the same time, localisation isn’t only a good response to current economic circumstances. It also paves the way for wider gains in terms of operational efficiency and revenue optimisation, too. So, the real question is less “why localisation?” and more “why localisation now?”. And to answer that, we need to dip a little deeper into the science of assortment.
Smarter science, better decisions
If we wind the clock back a decade or so, assortment was a relatively linear activity. Retailers would typically have a couple of variations in place, with some element of store clustering used to determine which products would appear in which locations.
At that time, anything more than this kind of ‘light touch’ localisation was largely impossible due to the sheer complexity involved, presenting a real challenge for retailers.
As noted above, one of a mainstream retailer’s key differentiators is that of scale. A large grocery retailer might be able to draw from an overall catalogue of around 30,000 products, compared with nearer 1,500 for a discounter, for instance6.
Clearly, while that depth is a huge advantage in terms of being able to offer customers what they want, it adds an undeniable element of complexity into proceedings as well. Think about it this way; to create a truly localised assortment, retailers would need:
- Granular insight into customer behaviours. As well as knowing how customers shop in a particular category, they also need a firm grasp on how they make their decisions and what they’re likely to want in the future.
- Deep understanding of product performance. That means being able to forecast the likely success (or failure) of a product in any given store – even if that product has never been ranged in that store before.
- An appreciation of various practical considerations. Each store has its own nuances – layout restrictions, different fixtures, etc. – and these factors also need to be incorporated when planning an assortment.
Doing this for just a single store would be a time-consuming process. A huge amount of data would need to be analysed, forecasts created, and planograms drawn up. But when you multiply that task across tens, hundreds, or even thousands of stores, what was just difficult instead begins to feel impossible. And 15 or so years ago, it almost certainly was.
Today, however, things have changed. As the industry’s data science capabilities have evolved, retailers have gained the ability to be much more granular when it comes to creating their ranges. As a result, we’re moving to a world in which hyper-localised assortment is a very real possibility – an area in which dunnhumby is leading the way.
AI-assisted decision-making
As is the case in many other fields, machine learning, or AI,plays a key role here.
Capable of taking on much of the mathematical ‘legwork’ involved in localised assortment, AI has made it infinitely easier to analyse data in the kind of volumes required to create a store-specific range. A well-trained AI algorithm can quickly create space-aware ranges and associated planograms ensuring the range fits to the space, something that would normally require hundreds of hours of manual efforts, and much back and forth between teams.
That’s not all. As we’ve found in our own work with AI-powered assortment, machines are also extremely good at optimising space. Given the right parameters, AI can tell us not just which products should be ranged, but how best to display them and how to fit more of them on the shelf, too. Naturally, the ultimate decision remains with a human – with AI augmenting, not replacing their decision.
For retailers, this creates an opportunity to fundamentally rethink the way they approach assortment. Critically, though, that opportunity doesn’t have to start with revolutionary, wholesale change.
Localised assortment isn’t about making every store’s line-up of products completely unique. To do so would eliminate one of the key reasons shoppers head to chain stores in the first place, after all: the ability to get a reliable variety of goods. Nor does a localised range have to be rolled out on day one across every store. Often, it’s a case of starting small, before gradually increasing the number of store-specific items. Subtle differentiations can still have a big impact.
While the shift towards localisation may be gradual, however, it’s certainly the direction of travel. And at a time when shoppers are used to getting almost anything they want – wherever and whenever they want it – localised assortment will give forward-thinking retailers the ability to respond to (and exceed) those expectations.
For more information on dunnhumby Assortment, visit dunnhumby.com/dunnhumby-assortment
Read more about the ways in which AI is changing the retail industry:
- Personalisation, products, and price: some of the ways in which dunnhumby is using AI
- AI: three breakout applications for consumer brands
- Understanding the why: Causal AI and its value for retail
- Opening doors unimagined: the next generation of assortment science
1 Market Guide for Retail Assortment Management Applications: Long Life Cycle Products – Gartner, 8 December 2022
2 The Localization Playbook: How to Develop Targeted Merchandising Strategies & Win Repeat Customers – A&M Consumer Retail Group
3 Preliminary Results 2023/24 - Tesco
4 dunnhumby Customer Pulse, October 2023
5 Discounters forecast to be fastest-growing grocery channel over next five years – The Grocer, 14 June 2022
6 The battle of traditional retailers versus discounters – Journal of Retailing and Consumer Services, 2017
TOPICS
RELATED PRODUCTS
Create customer-centric ranges using AI-powered science
AI-powered Assortment OptimisationThe latest insights from our experts around the world
![customer first data science analytics & machine learning services](https://www.dunnhumby.com/wp-content/uploads/2023/03/customer-first-data-science-analytics-machine-learning-services.webp)