Load testing – getting the metrics right
At thinkTRIBE, we know that load testing can be transformative for the businesses we work with. Often, load testing is the only way to be sure you can confidently deliver a smooth and responsive user experience (UX), even when your website is at its busiest. But – and it’s a big one – while load testing may be the best way of ensuring your site is operating at optimum availability, it’s only ever as good as the data it analyses.
One of the most commonly used testing metrics is that of counting concurrent users – after all, it seems to make sense to understand what your site’s ‘maximum occupancy’ might be, right? However, if your questions are limited to ‘how many concurrent users can my website handle?’, you’ll necessarily be excluding some important factors – including the potential distribution of web visitors across the site and the variety of requests that are being made. Counting concurrent users tells you nothing about where, how – and if – your users are interacting with the website, just that they’re there; if half of them have gone off to make a cup of tea, they’ll still be counted.
This is why bottlenecks can still occur, even if your site is experiencing visitor levels that are ostensibly well below peak. For transactional websites, the payment gateway can be a particular hot spot. The data gleaned from your concurrent user metric may show that your site can operate comfortably with a couple of thousand users – so long as they’re evenly distributed – but if they’re all queuing to check out, your site could still crash. Looking at the simultaneous number of users on your website can be helpful, but it’s important to understand that this only provides a snapshot of activity at a moment in time and isn’t the detailed analysis you need to effectively plan resources.
Taking a more dynamic approach
The most pertinent question, then, isn’t how many users you can accommodate but how many can complete their individual journeys in any peak hour. Only by looking at your traffic as a number of individual journeys, rather than as a fixed capacity, will you be able to analyse how your site handles real-life customers taking lots of different journeys.
The thinkTRIBE test engine differs from other approaches in that it was designed from the bottom up for realism. It doesn’t simply ‘replay’ a fixed sequence of URLs and claim it as a ‘journey’, but rather dynamically interrogates the live page at each step to determine the link for the next page to follow, choosing randomly from the options – options that are available to real users taking their own journeys.
These virtual users behave and interact with content in the same way as real users. By taking a wide range of paths, they are not only able to emulate stress conditions accurately, but can also highlight problems that may only present themselves when certain combinations of options are selected, enabling the tech team to intercept and resolve glitches before they impact the customer.
Gaining actionable insights
This kind of in-depth load testing analysis makes it possible to gain important insights about how customers’ behaviour impacts web performance and provides accurate data that can predict whether your site is equipped to handle peak traffic. Because testing protocols are based on real user traffic data, the results offer unique insights into user activity – including whether users are accessing the site via mobile or desktop devices and what the ‘drop-off’ ratios are.
Joined up retailing requires a multi-dimensional approach. Load testing has an important part to play in understanding customers’ browsing and buying habits so you can analyse and better understand buyer behaviour – whether you’re measuring the load of journeys with similar goals across different platforms, or those that include multiple devices and take place over more than one session.
thinkTRIBE has the expertise to help create effective tests for such complex scenarios and to discuss the implications with both business and technology teams. thinkTRIBE’s managed service approach means that we work extremely closely with clients to understand all their needs from testing, and to meet their operational and commercial goals.
thinkTRIBE provides a managed service precisely tailored to each client’s business and designed to optimise web performance. If you’d like more information about how taking a more realistic approach to load testing could help improve your customer experience, download our ebook here.