Before we got stuck into any technical aspect we needed to do a benchmark analysis. This included site load, server load, transaction rate, weekly trends and general gaps analysis.
With the high acquisition rate of new customers and traffic we needed to design a solution that was scalable but primarily built for 10x the current members ensuring it supported the community once it was complete. This was a difficult one to say the least as predicting future load based on historical data is never clear cut.
Once we’d mapped out where we needed to be we had to land on a technology stack that allowed for flexibility and wouldn’t cause the age old issue of “We can’t add that new functionality as we didn’t design for it, we needed to know from the beginning”.