How AWS is helping River Island to avoid the same fate as other retailers

Sooraj Shah Profile picture for user Sooraj Shah December 12, 2018
CIO Doug Gardner says AWS has helped the company transition to an agile, digital retailer.

The doom and gloom in retail  doesn’t look as if it will be disappearing any time soon; every month, well-established retailers are reporting profit warnings, falling sales, store closures, job redundancies, and in the worst cases of all, bankruptcy.

One of the key reasons for the retail woes has of course been the introduction of e-commerce, with the general consensus being that retailers were too late to go online and have not been agile enough to keep up with continual digital changes, and the associated consumer demands.

It’s for this reason that River Island, a long-standing UK retailer, with subsidiaries across the world, looked to undergo a huge transformation several years ago. River Island CIO Doug Gardner says:

The whole proposition of retail is under scrutiny and as part of our transformation, the business is really looking at what River Island stand for in a modern digital world, we do need to do something different to engage customers in-store, but I don’t think anyone right now has completely figured it out.

The big thing we’ve done is the ability to provide agility to the business, to do whatever comes up and that’s something we didn’t have before – with monolithic systems and data centers and all the traditional systems, everything was painful to do, time-consuming and was costing a fortune.

That agility has come through a transition where the retailer looked to rebuild its infrastructure, ways of working and tooling. Three years ago, Amazon Web Services (AWS) became the company’s main partner for this transition to ‘digital retailer status:

The first thing we did was migrate our website from an unreliable hosting provider to Amazon with the help of Cloudreach and that was more about scale and stability. The second part was building our core capabilities for the rest of the development team, and when we went through the selection process, Amazon was an obvious fit because it is a retailer itself, and because most of the tooling leaned heavily towards retail.


There were reports several years ago that Marks & Spencer decided to shift away from AWS because it saw it as a competitive threat – but Gardner says this was a different scenario as M&S used AWS as a hosting web platform:

They could probably see a lot of the tracking coming off of the M&S website, but the stuff we’ve got is in a secure place and I’m pretty sure Amazon has no access or interest in it – we’ve got space, tin and tooling off of them but we’re not using them as a managed platform.

AWS tools were not the only reason the cloud vendor was selected. Gardner suggests that his team spoke to Microsoft and Google, but AWS was by far the most engaging:

We weren’t like a start-up that already had digital capabilities inside us, and the fact they were willing to work with us on our transformation when we had almost zero percent of the skill set needed internally, it was a big influence in going for them. They were the most keen partner, really interested in our business and really engrossed themselves into it.

But that doesn’t mean that the likes of Microsoft and Google won’t be used in the future. Gardner feels as though the company is now in more of a mature state, and with its core base infrastructure inside Amazon, River Island can now look at adopting a hybrid cloud approach:

I think there are things that Google and Microsoft do well, so we’ll probably pick and choose individual components we like out of their clouds.

Charlie Wilkinson, head of architecture at River Island explains that this decision to use different cloud services comes with the bonus of not being locked in to one vendor, but that this wasn’t the main reason for it:

The different players have different specialisms, so it’s more about opportunistic rather than protectionist. In terms of making sure that we get things right –Google, due to the nature of their business, have got great toolsets around data insight and analytics and we would be shooting ourselves in the foot if we didn’t leave ourselves open to exploring those tools.

On the data front, Gardner says that machine learning is a big asset for River Island, as the company is in the process of using the technology to control stock replenishment and allocation, and push this into demand forecasting:

One of the biggest problems is we use manual systems to replenish stock – but when people buy online and return in-store it displaces stock all over the place, so having a good control of getting stock in the right location and satisfying any demand customer has makes a big difference for us.

The company has 308 stores, a website, as well as a number of wholesale partners and marketplaces. Gardner says for a company like ASOS which has a lot of different brands, it makes more sense to focus on search and personalisation, but as River Island is more focused on the ‘just arrived’ section, machine learning and AI are being prioritised for stock inventory rather than personalisation and search.

From 0% of the skills needed to 100%

Gardner had mentioned that before the transformation began, River Island had none the skills it required in-house from a cloud and development perspective. Roll on two or three years and things have changed dramatically. He puts this down to having the correct methodology, ways of working and engagement with the business:

More importantly, if you don’t have really creative, motivated, top-of-the-range engineers it is going to be tough. Going back a few years ago we were based in Hanger Lane and didn’t have much to shout about, it was very hard to get engineers to work for us – there was nothing in that early foundation. We put in huge effort, opened a tech office in Shoreditch and we ran events in the office, put some very interesting technology in and now we attract the best people in the industry who want to stay and work for us.

Wilkinson adds that in order to deliver real value, the company realised it needed multidisciplinary teams.

That means we have to have teams that have micro-service engineers who are great at GoLang, and cloud operations engineers, but also we have guys and girls from our core engineering functions who have a PLC background – a lot of Oracle and so we don’t see much of a divide between the new and old because that breeds hostility and if you want to deliver something meaningful you have to be able to operate over both, and be able to make the new world talk to the old world.

What next with AWS?

Gardner believes that while River Island has already come so far in its transformation, there is still work to be done over the next few years. However, he says that some of the hardest parts of the transformation are already complete, and 2019 will be for converting legacy systems into modern systems.

Wilkinson states that he is interested to see more web stacks and consumer-facing digital experiences.

I think there is a lot of opportunity for us to innovate there and I’ve asked the question of what a fully serverless digital stack would look like for our web stack. We’ve got a fairly traditional three-tier stack which is okay but there is an opportunity to innovate do something really forward-thinking on that front, so I’m excited about a serverless stack there.

For Gardner, the biggest indication of how far the company has come is that external parties are telling him that the company has built a strategic capability that its competitors don’t have:

In the last five years, a lot of retailers were entrenched in trying to ignore what’s going on and didn’t invest like we had – and now we’re coming out of the other side with maturity and acceleration.

In addition, the company now feels like it is able to take on board more quickly any new ideas that come from its peers:

We can really push the edges on serverless, whereas two years ago we would have said ‘that’s cool, we want to do that, won’t it be great when we get there’. We’ve matured a lot and we’re now on the bleeding edge.

A grey colored placeholder image