John Lewis Partnership goes Google Cloud-first to better target customers
A new data platform will help to combine John Lewis and Waitrose data, and provide a single view of the customer.
The John Lewis Partnership comprises of the John Lewis & Partners department stores and Waitrose & Partners supermarket, two of the most well-known names on the high-street. Like many others in the retail sector, the company has been striving to become more digital – attempting to bring the same type of service to its e-commerce offerings.
In addition, it has looked to bring together Waitrose and John Lewis from a technology perspective, breaking down its data silos and providing a foundation for a more tailored approach to the way it interacts with customers.
Andrew MacInnes, CTO of the John Lewis Partnership, runs a team of architects that come up with responses to business demands, helping to design and create solutions as well as educating and pushing for digitisation and changing ways of working across the partnership.
He told diginomica at Google Next 19 in London:
Traditionally, IT has been about ‘we’ll wait until someone tells us what to do and then we’ll make it’. The creation of my role, nearly a year ago, was to turn that on its head and push back and educate the rest of the leaders in the organisation about what IT can do and how to transform the business.
I think we’re at a stage now where with the digital transformation that’s occurring, it’s impossible for any business leader to understand the entire gamut of technology and be very good at their line of business, so they need a team of trusted advisors to help them and that’s effectively what me and my team are.
Step in Google
John Lewis had started using G-Suite back in 2014 as it shifted away from Lotus Notes, and this was the first experience the partnership had of Google as a technology provider. However, MacInnes explained that the retailer had a strong relationship with Google already through the ads and service side of the business – as it stocks many Google products in its stores.
The main driver for introducing Google Cloud to create a centralised data platform, was to get a single view of the customer.
Prior to this, Waitrose had its own systems and John Lewis had its own systems so it didn’t quite work. What we now know is that our best customers at John Lewis are the exact same people as our best customers at Waitrose, so we’re reorganising our organisation and our IT systems to reflect that.
His team created a new platform for data for the partnership, and the first type of data that he wanted to consolidate and analyse was customer data – giving John Lewis what it calls a golden record of the customer. This would also help the company from a regulatory perspective, to ensure it is GDPR compliant – enabling the company to forget about the person without having to go through numerous different databases to check whether there is data on the customer still stored. In addition, if the company wants to market to its customers, it doesn’t need to send several messages from different lines of business.
While the company is just starting to build up its data platform, the customer side is just the beginning, with product data, financial information, supply chain data and various other sources all likely to be shifted onto the platform. MacInnes believes that AI and machine learning will then give John Lewis the ability to change every single part of the business.
At the moment, those benefits are easy to see with customers but we will build this out to be a bigger platform. For instance, we do propensity modelling, which allows us to test how likely someone is to shop on a brand, modelling on one brand would have taken us three weeks in the past, but we’re now in a position where we can make 1000 of those models in three hours, so we can now make sure our marketing and targeting spend is going to people who it’s really going to have an impact with, and our customers are going to see something more relevant to them.
The loyalty programmes of John Lewis and Waitrose will also be combined – although MacInnes stressed that the company would do this carefully from a GDPR perspective, particularly as customers are becoming more savvy about the value of their data.
We have to come up with a really compelling proposition to make loyalty work so that customers are on board.
Google Cloud first
The relationship that John Lewis now has with Google Cloud is the smallest in terms of revenue for Google – it’s dwarfed by the advertising spend and their spend on devices that go into John Lewis shops.
It’s this relationship which means John Lewis, while not committing completely to Google Cloud Platform (GCP), will consider Google first, and predominantly use Google’s cloud.
Because of the scale of the relationship we have, there’s benefits for us to get great expertise from Google. But if you know how licensing works in cloud and how vendors can be quite protective about their software, there are things that make no financial sense to put into cloud and there’s benefits for us to put particular things in AWS and particular things in Azure.
Waitrose’s e-commerce site is hosted on AWS, and while John Lewis doesn’t have a relationship with Microsoft Azure at the moment, its SQL licensing which enables customers to get a licence on a datacentre with a free licence in the cloud, is an attractive proposition for MacInnes’ team.
Better use of data
In the long-term MacInnes wants to be able to use Google’s AI features to improve its recommendations to customers and the way it presents products to them.
If you come to our shop and you buy a Batman costume, what should the next recommendation be – is it a Snow White costume, or a Batman Lego set? What did the person actually want and what are their interests? Understanding their shopping history as well as other data – to show them the Snow White costume because they’re probably buying costumes as it’s Halloween or showing them the Lego because they love Batman, is what we want to be able to do. It’s a capability we don’t have today so we can’t do it – but it’s a default out of the box commodity solution that Google have.
In the future, John Lewis could follow in the footsteps of many of its rivals in altering the model displaying an outfit so that customers can get a better idea whether the product would suit them.
Understanding customers to the extent where we know what effectively you look like – if you’re buying in a particular size and you tell us more about you we could show you models that look like you. Some of our competitors have done this – if you have a diverse range of models by size, age and race, customers respond better.
In the long-term, John Lewis will hope that it will be leading the way in the retail and e-commerce sector in the use of AI and machine learning.