How do you measure the user journey?
The most effective monitoring approach is one that replicates real journeys by real users as they navigate your website. By taking the customer-centric perspective, you can walk in your customers’ shoes, sharing the same experiences – including any performance issues that may impact conversions and harm your brand.
Journey Delivery Time
By calculating the minimum, maximum and average delivery times between the start of the component request and its completion, you can obtain an ultra-high-level view of page performance. Comparing these statistics over time can give you a good idea of when your web performance dips, if updates affect performance or if you need to optimise pages for peak traffic periods. For example, if your average delivery time is increasing by 10 percent in the early evening, or by 30 percent during Christmas and Black Friday peaks, you’ll know when and by how much you need to increase capacity.
You can use domComplete to measure delivery time, typically after static downloading is complete: i.e., when the browser-loading spinner has stopped, even though some dynamic content may still be ongoing. domComplete has its limitations but still offers a useful metric for assessing the impact onLoad content has on overall delivery times. domComplete should certainly be part of the KPI line-up for monitoring web page performance.
Page Delivery Times
Ideally, you want the whole customer experience to be consistent across your website. However, some pages will impact the overall UX much more than others – and it’s these that require the greatest attention. For instance, your homepage is likely to have a disproportionate effect on conversions – if customers are waiting too long for the homepage to load, they may log off, regardless of how well the rest of the site performs.
Key pages are likely to include your homepage, product listing page (PLP), product display page (PDP) and cart/checkout. Your Search Engine Optimisation (SEO) ranking depends heavily on performance, which is why it’s important to pay attention. At thinkTribe, we break down performance into a series of steps – each containing a page load or key request (adding an item to a basket, for example).
This offers a valuable insight into which pages are performing badly and why. Monitoring journey delivery time and individual page delivery times will help you identify how pages are performing over time and reveal weak points.
Time to First Byte (TTFB)
Also known as server response time, TTFB describes the time between the start of the request and the point at which the browser receives the first byte of information – it’s essentially the amount of time the server spends processing the request and returning the response. Lots of factors can affect the TTFB – high load, large dynamic content, lack of caching, volume of queries – but it’s crucial to resolve any problems and keep TTFBs below 500ms. A high TTFB will be perceived as slow service by customers, even if your page loads relatively quickly after processing is completed.
DNS (Domain Name Service)
DNS lookup time is included in the TTFB – but it also warrants a specific consideration if server response time is slow. DNS takes a host name – thinkTribe.com, for example – and returns the IP address of its location. It takes only a small amount of time, but because a lookup is required for every unique host, DNS lookup time can soon start to balloon, especially if there are multiple unique hosts, internal and third parties on your page. The process can be mitigated if resources are hosted internally and if a fast DNS provider is used. It’s worth remembering, though, that DNS lookups do get cached, so users won’t have to keep referencing the same host on return visits.
Page weighting allows you to track the size – in KB – of each step in a user journey against the delivery time for that step, giving you a view on how page content is affecting overall delivery time. Perhaps you’ve just released an update to one of your pages and your delivery time has spiked. With page weighting, you’ll be able to determine if the amount of content on your page has increased and if it correlates with the higher delivery times you’re experiencing. Conversely, if delivery time remains level despite a decrease in web page size, it may point to a different performance issue.