The Co-op is a food retailers steeped in tradition. Founded in 1844, the now wide-reaching organisation has endured a reputation in Britain for being slow to change.
However, after four years, the company has transitioned all 3,000 of its stores to SAP’s cloud platform, making it one of the biggest SAP deployments in Europe.
ASDA, a larger supermarket chain than the Co-op, spent six years trying to do the same and backed out due to the roadblocks it faced.
Scott Robertson, principal cloud solutions architect at the Co-op, joined Cloud Tech News to talk us through the insights he gained during such a largescale digital transformation journey.
Cloud Tech News: When and why did the Co-op decide to transition to the cloud?
Scott Robertson: We started our cloud journey about four and a half years ago with two aims. One was to become slicker through digital transformation. The other was to wind away from the legacy risk associated with the old ways of doing IT by closing down our data centres and adopting a new model.
CTN: What stands out as the key insights you gained from this four year long migration to the cloud?
SR: Above all, it’s that cloud transformation isn’t about tech. Cloud adoption has three pillars – people, process, and technology – and everyone focuses too much on the tech, but it’s largely irrelevant.
Whether you’ve got the blue cloud provider, the red cloud provider, or the yellow cloud provider, they all work in a very similar fashion. Obviously there are some nuances when you get down to the details, but they all do very similar things for you.
What matters is how it changes the way you do things as an organisation. The biggest challenge in any cloud transformation journey is changing your company’s culture – how does your organisation and its employees think? The tech itself is brilliant, but how are you going to use it? How are you going to change the mindsets of employees that need to do their job differently now?
Another critical problem I noticed is when organisations dip their toes into digital transformation without knowing what they want to get out of it. Too many companies are solution focused rather than outcome focused in asking what they want to deliver. “We want to have stuff in the cloud.” Why? If you just lift and shift what you have, without changing your architecture, processes, or tooling, it’s going to cost you more money.
So, you have to rearchitect or re-platform, which is possible if you’re in control of your software development and build your own apps. If you’re not, you have to buy a packaged application off a vendor, leaving you stuck with their prerequisites.
This can start to erode the benefits of moving to the cloud, so I always recommend thinking about the applications and business processes that the cloud can support instead of the infrastructure.
CTN: What were some of the major challenges the Co-op faced during its journey?
SR: Some of the biggest hurdles we’ve had to face weren’t from the cloud providers but our own internal processes. For example, I don’t know any organisation to this day that has fully solved adjusting its internal financial mechanics to allow staff to work in the way cloud providers allow you to work.
We also still have a lot of education that needs to be done with our business as usual (BAU) support service teams to adapt them to our SAP deployment. Working on Azure entails a lot of changes, like we no longer need a storage person and a network person, all we need is a platform person. Someone who may have once sorted wiring in our data centres may now look at data metrics for capacity and cost of cloud storage.
CTN: With your SAP deployment complete, is it time to kick back and relax?
SR: Ha! If only. We’re currently in the process of consolidating some of our on-premises data infrastructure by turning a large scale Hyper-V deployment into an Azure Stack HCI.
It’s taken us down from sixteen racks to two, which saves the Co-op about £200,000 a year in electricity alone, and that’s not even why we did it. We did it to get off the old Hyper-V kit that’s going to be dated in a few years and because Azure Stack HCI is an extension of Azure, it’s not a new on-prem.
This means we can bring all the tools, processes, monitoring, security, and policies that we developed for our Azure cloud deployment and bring those on-prem.
CTN: You mentioned on-premises data. Does the Co-op still use data centres?
SR: For now, yes. Our current hybrid cloud model is data centre centric, but we are looking to shut down our data centres in the future.
To achieve this, we’ve been working closely with one of the world’s largest data centre companies, Equinix. Amongst other things, they are providing our on-ramp into Azure, which means aggregating our network out of our data centres and into Equinix’s two UK data centres in London and Manchester.
Through Equinix, we have been able to start moving away from our data centre centric model towards a cloud hub that aggregates our services. In a way, that makes Equinix the glue that holds our on-prem and Azure worlds together.
What Equinix facilitates is enabling us to have specific policies for specific applications, instead of a stringent catch-all policy written from the security perspective. With this model we can, for example, give developers with more experience greater access on a per application basis.
CTN: How close is your relationship with Equinix? What else do they support the Co-op with?
SR: A lot, actually. Without the Equinix cloud exchange we have in operation, we could not have delivered SAP. End of.
It goes back to when we were trying to find a UK front door into Azure. Equinix was one of only two companies with such an offering but we found Equinix to be the better fit and the better product.
After a proof of value, it became clear their processes work well and the setup was pretty slick. So the Co-op went with them as our product of choice. It essentially came down to the two cultures of our organisations aligning incredibly well.
CTN: What made Equinix so essential to the operation?
SR: Whilst in this phase, I was working on our request for proposal (RFP) with SAP for our supply chain. This involved lots of time with Microsoft figuring out how we wanted our protective to look like, how we might back it up, and how we can make it resilient.
We started off by migrating simple services, but at some point you have to bite the bullet and start migrating business critical services.
To do that you need an excellent service owner who’s happy to provide for you, which is quite hard because nobody wants to be the first to migrate critical services.
We ended up not hosting through SAP, but instead we essentially fronted an internal RFP that was backed by Microsoft. I’d never been involved in an operation quite like it and I fully expected it to fall flat on us.
However, from the technical perspective, it was the right answer. We had a great relationship with Equinix so we were confident we could build a resilient network path and we knew what we needed to do with Microsoft.
Ultimately, it paid off. I estimate us taking that risk shaved nine months off our delivery time. Had we deployed with SAP, a project that size would have cost around a million pounds a month, so that was a massive payoff.
Looking to learn how to establish a strategic hybrid cloud? Learn more about the virtual Hybrid Cloud Congress, taking place on 18 January and explore how to optimise and unleash the power of your hybrid cloud.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.