Re-platforming: Notonthehighstreet’s Monolith to Microservices Journey
As notonthehighstreet looks to expand its tech team, the company’s CPTO Richard Zubrik talks about his experience of re-platforming, and how the brand’s exciting digital transformation
Re-platforming, often known as a ‘digital transformation’, is something a lot of online businesses have either gone through or are thinking about going through in terms of their technical stack. Technology advances quickly, and so ten years ago when these businesses’ websites and applications were first built, their tech stack was current and modern and did the job it needed to do at the time. Now, these tech stacks can be difficult to work with and manage as they’ve been built on top of, over and over again. Small changes become difficult to make and large changes virtually impossible. Therefore the ‘re-platform’ becomes the focus of many online retailers and marketplaces..
So, you’re probably wondering why I’m writing about this. I wanted to share the experiences as I’ve seen, been part of and led over the last couple of years. While they were all different, it’s fair to say that through each one I learnt something new & important. My specific focus will be on the world of online commerce where businesses either have bespoke in-house built platforms or one of the ecommerce packages hosted in either physical data centres or public cloud.
My beginning with re-platforming
I first came across re-platforming back at my first ecommerce tech (well, ‘IT’ at the time) job. I had very little experience within ecommerce tech but still managed to get myself involved in the build of a new online food platform for international markets. While it wasn’t exactly re-platforming (we only had an ageing UK platform to look at as the basis), the idea was to use the new international platform later in the UK if it worked well.
To my, and probably a few other people’s, surprise the new international platform was a success, we managed to launch online food shopping to 7 new markets in about 24 months. However, it became clear due to the scale, features and reliability it wouldn’t have been able to handle the volume and complexity of the UK market. There is a good lesson to be learnt from this though — the importance of the Minimum Viable Product (MVP). If your MVP is small enough, you have a really good chance of getting something into production. It’s a lot easier to get something over the line and build on top of it than it is trying to get the whole thing done on the first day. And you’ll be surprised by how much the rest of the backlog changes once you do go live. Things you thought you’ll need to do next won’t be that important and a whole new different set of features or bugs get to the top of the backlog.
I then took part in a completely different re-platforming venture at the same company. The outcome was meant to be a feature rich, own product, dropship and marketplace platform, competing with the likes of Amazon in the UK. I was lucky enough to be handed the responsibility of running the platform in production which was an incredible experience and one I learnt a lot from. It took us a year to get it to a reliable state, but once we did we were able to start building new features which was a major improvement. However the issue came down to the fact that by the time we had finally got it into a good place, it was almost time to re-platform again. There is another big lesson to be learnt here: unless the tech stack you use enables you to keep things up-to-date and current, you will be re-platforming every few years, slowing down the pace of change for customers.
Moving to a bigger scale
My next move was to an online fashion retailer and another new ecommerce platform to be built. At the time I joined, the business was 15 years old and their platform was about the same age. The business had been on a re-platforming journey which started as the ‘new checkout’ but expanded to most of ecommerce. The legacy platform was an in-house built .NET based platform hosted in a traditional data centre.
When I joined I decided the best thing to do was to take advantage of being new and to take a step back with fresh eyes. I looked at the plans and how they matched what we were actually delivering, the code and what were we focusing on. I knew, from previous experience, that the sooner we broke things down, the sooner it would get done. So our focus went to smaller increments and most importantly to the present; looking at what we’re doing this and next week rather than worrying what will be in a few months’ time. When we got closer to our MVP we start looking at the amount of risk we were willing to take, accepting that things might go wrong but we will learn from it and start working on bugs that really matter quicker. Near the finish line, I decided to turn the traditional project management model approach upside down, and instead of looking at what there is left to do, adopted a very different approach. Our focus went to — what is actually stopping us from going live today? What would we worry about if we turned it on now? It focused all of us including Product Owners, Engineers & QA’s on the real MVP.
As you can imagine, not all of it went smoothly and there were things that slowed us down: our old platform being one of them. However, what was good was that we were winning more than we were losing, and this really helped to keep the team motivated and driving forward. We managed to get the platform out before peak; having a stable platform and significantly improved customer experience. The success of this should definitely go to the small wins and the amazing hard working team who made it happen!
The future at notonthehighstreet
My latest move has brought me to another ecommerce marketplace and the home of 5,000 incredible small creative businesses, notonthehighstreet. Since I started back in January 2019, my task has been to undertake another exciting re-platforming challenge. Our current platform is over 10 years old but there is good news — it’s already hosted on AWS. It’s reliable and doesn’t require a great deal of maintenance or support (shout out to the fantastic Tech @ NOTHS team for keeping it this way!) but unsurprisingly some of the features, ok most, need updating and bringing to 2020. Our current tech stack is Javascript with React on the front-end, and then some services using GraphQL/Node.js and Java for others. All of this is then hosted in containers or lambdas in AWS. We also have both Native Apps & Data functions.
My plan for the team is to use all of the learnings from my previous roles to guide the team into making some great improvements. I’m now a year and a half into my NOTHS journey and I’m excited to say we have made some great progress already. We’ve already got a new search being tested with customers on our site, new basket, CMS, navigation, CRM and we have launched an Android app! I’m excited to see what the future has to hold for us and for what we’re going to be able to achieve together as a team. We’ve got a fantastic foundation and made some brilliant progress so watch this space to see what we can do and find out more about our journey…
Want to be part of notonthehighstreet’s digital transformation? We’re currently recruiting for a number of roles across our tech team. Check out all our latest positions here.
