CommerceStar represents the 3rd iteration of a solution originally designed & used by Approved Food, where I worked as their Technical Director for 7 years.
- V0 – the existing system before I joined was internally hacked Windows VB app + MS Access for their internal paper base pick system that was very loosely connected to a static PHP online store. The business rapidly outgrew the technology and needed a replacement. I was the only developer.
- V1 (2011) – My first re-architected version used a shared MySQL database with a Windows VB app, PHP based monolithic online store and hybrid paper & PHP based internal fulfilment system I was the only developer until 2013, then +1 junior.
- V2 (2014) – The internal application was migrated to a Windows C# .NET business application. The PHP based online store was migrated to a Docker micro-services to enable rapid scaling and improve robustness as the business continued to grow rapidly. The business ran multiple TV ad campaigns that pushed the limits of what could be achieved with that system. Reverted to sole developer, then +2 by the end of 2015.
This image shows the elements of the architecture I’ve been the key developer for between 2011-2017 within an online food retailer.
Key events during the evolution of that system
- Customer migration towards powerful mobile devices
- Google’s perpetual push for faster & faster websites to maintain search ranking
- TV campaigns stress tested the solution
- Integration with other businesses who were also becoming more “online”
- Technology changes, services change, customers change.
- NewRelic optimised
During that period, off-the-shelf e-commerce platforms were evaluated based on the requirements of selling food & drink online. Platforms including software-as-a-service Shopify Pro and custom installs of Magento have been evaluated, prototyped & trialled since 2011 and they continue to have restrictions that fail to meet the demands of selling online. A top UK ISP also advised us that Magento was an “expensive solution that requires high ongoing costs and won’t scale easily. Spending more money is your only option.” That ISP asked me to do a presentation on my solution as something they could build and offer their customers as a “better efficient alternative to Magento”.