SBN

Get the message: Your shopper data is trying to tell you something

This article originally appeared in Internet Retailer

As we head into the thick of the holiday shopping season, web teams must make sure the path from discovery to purchase is as easy and painless as possible. Knowing where potential fail points are ahead of time makes it easier to prepare, while real-time user data during the rush can help quickly spot problems.

In my house, the weather forecast always seems to be a topic of conversation. “What’s the weather going to be like this week?” I’ll ask my wife. But even with my eye on the forecast, when I wake up in the morning one of the first things I do is check the local temperature from my iPhone to help me choose what clothes I’ll wear, and also mentally prepare for the chill. Even with the benefit of weather stations all over the map giving me the temperature data and expected precipitation, I still find myself opening up the front door and taking one step outside to see, how cold is it really and is it raining/snowing yet? You know you’ve done it too, when someone comes in from outside as you’re deciding whether or not to grab a jacket and you ask, “hey, is it cold out?”

Monitoring web performance is very much the same. There are some cases where synthetic data, measured from test agents out on the Internet, is incredibly useful for setting a baseline for website performance. But to really understand the experience your users are seeing, web teams need to collect and analyze real-user performance data. Only data collected from actual site visitors and the devices they use will allow retailers to uncover the key patterns and actionable insights needed to keep website and application performance at optimal levels for a smooth visitor and buyer experience.

It’s critical to understand if there is a gap between when your user thinks the application is ready and when they can actually interact with it.

For example, conversion rates may reveal that shoppers are abandoning the landing page for the deal of the week. But it is only with a granular real user data that the web team can figure out why–was there a micro-outage and what caused it?

As we head into the thick of the holiday shopping season, web teams will feel the pressure mounting to make sure the path from discovery to purchase is as easy and painless as possible. Lost sales due to poorly performing or unavailable sites are unacceptable in the cut-throat competition of the holiday season. As page load time goes from one second to five seconds, the probability of bounce increases by 90%. Fortunately, capturing real-user data in real time will help web teams detect when and where users are stumbling or falling down in the process and allow them to determine what to do to ensure they have a well-performing site.

Unpacking real user performance monitoring

The first thing to know is that real-user monitoring does not replace synthetic testing. Instead, it acts as a complementary process for different stages of the development and deployment cycle for retail web sites. Synthetic testing in browser is still important to help web teams make sure everything is working as intended, especially during the pre-deployment phases of development. Synthetic tests help set a baseline for performance optimization and can be useful in running experiments. Using multiple test agents for script tests that mimic what a user would see can reveal some very helpful points of optimization for the website.

However, these scripted tests are only as good as the test plan. Real users are much less predictable, and scripted tests are likely to miss conditions that will happen out in the real world.

They also can’t see everything. Real-user monitoring creates a more complete view of your site traffic and user experience by giving web teams visibility into 100 percent of the users that come through your site. It can provide very granular insight into how the performance of the page and specific components on it impact user behavior and business metrics. Image and JavaScript files are the heaviest elements of a typical web page and contribute the most to slowness. In querying the HTTP Archive Paul Calvano found that 27% of sites have greater than 90% of 3rd party requests. These third parties can be problematic for performance.

For example, a web page may appear visually ready to interact with, but heavy elements such as third-party JavaScript require additional loading, causing major slowdowns and users to “rage click” or click repeatedly in frustration when the page is unresponsive. It’s critical to understand if there is a gap between when your user thinks the application is ready and when they can actually interact with it. It’s also important to understand what is causing the disruption.

Season-proofing with real-user data

With the insights from real-user data at hand, it is now time to put them to use. These insights can help web teams prepare for potential crises and make improvements that meet business goals when the stakes are high during the holiday rush.

It is helpful to have these monitoring processes in place before a peak traffic season. The initial insights during the weeks leading up to the busy season can reveal some areas to watch during the rush, such as a slow-loading JavaScript component that could fail under added pressure. Knowing where potential fail points are ahead of time can help web teams prepare by running predictive scenarios like, what would happen to my business metrics if this JavaScript slowed down, sped up, or was removed altogether? But, even if you missed the boat on monitoring in advance, having real-time user data in the moment during the rush can help you keep your finger on the pulse of the customer experience.

It is also important for web teams to make sure they have a clear understanding of business priorities and KPIs and incorporate those into their strategy for optimization. For a retail site, that includes increasing conversions, reducing bounce and maximizing time on site.

According to Akamai’s 2017 State of Online Retail Performance Report, data showed that conversion rate decreased drastically as page load time went up. For example, a 3-second load time cut mobile conversion in half from  about 2 percent to 1 percent and continued to decline sharply from there as load times got longer. Being able to correlate those metrics to performance is valuable. Further than that, being able to monitor specific page elements and third-party objects and their impact on these KPIs gives you the power to prioritize the enhancements you need to make in preparation for and during the big rush.

As consumers start their holiday shopping earlier every year, or for those that race to the web at the last minute (me), retailers need to prepare with testing and experiments but also keep close watch by monitoring real-user performance during the holidays to give shoppers a seamless and satisfying user experience on any device, all the time.

*** This is a Security Bloggers Network syndicated blog from The Akamai Blog authored by Anthony Larkin. Read the original post at: http://feedproxy.google.com/~r/TheAkamaiBlog/~3/8MW29m3Tmak/get-the-message-your-shopper-data-is-trying-to-tell-you-something.html