With the holiday shopping season in full swing, poor web performance can have a big impact on revenue. There’s intense competition for online shoppers, and customers will quickly bounce to another site instead of slogging through a bad experience. The best way to track and achieve your web performance goals is through experience-based SLOs (Experience Level Objectives, or XLOs). Get it wrong, and you risk losing sales during the holiday shopping rush—something we saw happen during Black Friday.
This year, the National Retail Federation projects 2024 holiday sales will reach record levels, up to $989 billion, with growth of 2.5% to 3.5% over last year. According to a recent study, the upward trend of online shopping will continue, as more than 90% of shoppers go online, and over 40% planning to do most or all their holiday shopping online.
Most web Service Level Objectives (SLOs) are based on simple availability: Is your website up or down? But that doesn’t give you a good view of the performance users are actually experiencing at their locations. During the holiday shopping season, increased traffic can cause congestion and performance issues that don’t impact your service availability but can significantly harm customer experience.
As our upcoming SRE Report 2025 confirms “Slow is (officially) the new down.” Users won’t wait for poor web performance—they get frustrated and bounce. Setting XLOs for performance metrics provides early warnings about negative trends that can impact business long before you reach an alert threshold.
To set and track performance goals on user experience, you need to measure performance metrics, like Core Web Vitals, from real locations—not just from the data center or cloud, where your customers aren’t. This is a key requirement for XLOs and the main factor that differentiates an SLO from an XLO. It’s the best way to measure web performance against realistic goals.
To illustrate this point, consider two similar competing websites. We’ll call them Company_A and Company_B to avoid public shaming. We’ve set both an availability and a performance XLO for them. The performance XLO is based on Largest Contentful Paint (LCP), a Core Web Vital metric that helps indicate when the main content of a web page has likely loaded, making the page ready for user interaction.
The figures below show the availability burndown charts for Company_A and Company_B for the week of Thanksgiving through Black Friday and Cyber Monday. For both companies, the XLO was leniently set at 99.9% availability.
As shown in the first chart, Company_A experienced a 6-minute outage on Black Friday evening. The outage wasn’t major enough to violate the overall objective, but it was unfortunately timed. A deeper analysis revealed web timeout failures for customers in Dallas, Denver, and Seattle, as they waited for servers to respond. Customers in these locations may gone to Company_B’s website instead, resulting in lost revenue for Company_A.
The second chart shows the availability burndown for Company_B. There were no outages during this time, so no XLO violations occurred, and the 99.9% objective was easily achieved. Looking at this alone, you might think all of Company_B’s customers were satisfied—but you would be wrong.
The burndown charts for LCP below tell a different story. A good LCP score is 2.5 seconds or less. The XLO was set at 75% adherence for 2.5 seconds. That means for the majority of customers (or 75% of the time), the largest element rendered on the site happened within 2.5 seconds, providing a good user experience. This may seem lenient, but a 75th percentile is recommended for this Core Web Vital.
The first burndown chart shows Company_A easily beat the XLO objective, providing most customers a good user experience. This doesn’t include outages, so it only applies to users who successfully accessed their website.
Company_B, on the other hand, showed consistently slow LCP and missed their XLO before the Black Friday weekend was over.
A closer look at the scatterplot below shows that the LCP was often greater than 4 seconds. The black horizontal line represents the 2.5 second mark, making it clear how this objective was regularly missed throughout the major shopping week.
Note that LCP isn’t necessarily when the web page is fully interactive, only that the largest object has rendered. A deeper analysis shows the full measure of Time To Interactive (TTI) for Company_B was often above 8 seconds when LCP was high. That level of poor performance was found across the entire month of November, not just Black Friday through Cyber Monday. Users experiencing that poor performance are typically frustrated and at risk of bouncing—likely to the competitor, Company_A.
Measuring web performance from real locations is vital for setting realistic XLOs. As shown here, both companies maintained availability over 99.9% (despite Company_A’s limited outage), but only Company_A met their performance objective, offering a better user experience overall.
While Company_B might have congratulated themselves on 100% availability, the performance XLO shows they provided a suboptimal user experience for 25% of their customers for the entire high-value shopping period. This would have had a broader impact on user experience and, consequently, on revenue than a brief outage in a limited set of locations.
This disparity highlights the importance of focusing on performance metrics like LCP, which directly impact user satisfaction. Availability alone can be misleading; it’s not enough to represent user experience. While both companies had issues during the peak holiday shopping period, the performance XLO suggests that Company_B fared worse.
The best way to avoid failing the peak shopping season is to set both availability and performance-based XLOs. Burndown charts, like those above, can help you project whether you’re on track to meet or miss your performance objectives from a real customer perspective. Ultimately, that impacts revenue. Given how much is at stake during the holiday shopping period, customer experience must be a top priority.
To learn more about how we can help you stay resilient during peak traffic periods, check out our Internet Resilience Program
With the holiday shopping season in full swing, poor web performance can have a big impact on revenue. There’s intense competition for online shoppers, and customers will quickly bounce to another site instead of slogging through a bad experience. The best way to track and achieve your web performance goals is through experience-based SLOs (Experience Level Objectives, or XLOs). Get it wrong, and you risk losing sales during the holiday shopping rush—something we saw happen during Black Friday.
This year, the National Retail Federation projects 2024 holiday sales will reach record levels, up to $989 billion, with growth of 2.5% to 3.5% over last year. According to a recent study, the upward trend of online shopping will continue, as more than 90% of shoppers go online, and over 40% planning to do most or all their holiday shopping online.
Most web Service Level Objectives (SLOs) are based on simple availability: Is your website up or down? But that doesn’t give you a good view of the performance users are actually experiencing at their locations. During the holiday shopping season, increased traffic can cause congestion and performance issues that don’t impact your service availability but can significantly harm customer experience.
As our upcoming SRE Report 2025 confirms “Slow is (officially) the new down.” Users won’t wait for poor web performance—they get frustrated and bounce. Setting XLOs for performance metrics provides early warnings about negative trends that can impact business long before you reach an alert threshold.
To set and track performance goals on user experience, you need to measure performance metrics, like Core Web Vitals, from real locations—not just from the data center or cloud, where your customers aren’t. This is a key requirement for XLOs and the main factor that differentiates an SLO from an XLO. It’s the best way to measure web performance against realistic goals.
To illustrate this point, consider two similar competing websites. We’ll call them Company_A and Company_B to avoid public shaming. We’ve set both an availability and a performance XLO for them. The performance XLO is based on Largest Contentful Paint (LCP), a Core Web Vital metric that helps indicate when the main content of a web page has likely loaded, making the page ready for user interaction.
The figures below show the availability burndown charts for Company_A and Company_B for the week of Thanksgiving through Black Friday and Cyber Monday. For both companies, the XLO was leniently set at 99.9% availability.
As shown in the first chart, Company_A experienced a 6-minute outage on Black Friday evening. The outage wasn’t major enough to violate the overall objective, but it was unfortunately timed. A deeper analysis revealed web timeout failures for customers in Dallas, Denver, and Seattle, as they waited for servers to respond. Customers in these locations may gone to Company_B’s website instead, resulting in lost revenue for Company_A.
The second chart shows the availability burndown for Company_B. There were no outages during this time, so no XLO violations occurred, and the 99.9% objective was easily achieved. Looking at this alone, you might think all of Company_B’s customers were satisfied—but you would be wrong.
The burndown charts for LCP below tell a different story. A good LCP score is 2.5 seconds or less. The XLO was set at 75% adherence for 2.5 seconds. That means for the majority of customers (or 75% of the time), the largest element rendered on the site happened within 2.5 seconds, providing a good user experience. This may seem lenient, but a 75th percentile is recommended for this Core Web Vital.
The first burndown chart shows Company_A easily beat the XLO objective, providing most customers a good user experience. This doesn’t include outages, so it only applies to users who successfully accessed their website.
Company_B, on the other hand, showed consistently slow LCP and missed their XLO before the Black Friday weekend was over.
A closer look at the scatterplot below shows that the LCP was often greater than 4 seconds. The black horizontal line represents the 2.5 second mark, making it clear how this objective was regularly missed throughout the major shopping week.
Note that LCP isn’t necessarily when the web page is fully interactive, only that the largest object has rendered. A deeper analysis shows the full measure of Time To Interactive (TTI) for Company_B was often above 8 seconds when LCP was high. That level of poor performance was found across the entire month of November, not just Black Friday through Cyber Monday. Users experiencing that poor performance are typically frustrated and at risk of bouncing—likely to the competitor, Company_A.
Measuring web performance from real locations is vital for setting realistic XLOs. As shown here, both companies maintained availability over 99.9% (despite Company_A’s limited outage), but only Company_A met their performance objective, offering a better user experience overall.
While Company_B might have congratulated themselves on 100% availability, the performance XLO shows they provided a suboptimal user experience for 25% of their customers for the entire high-value shopping period. This would have had a broader impact on user experience and, consequently, on revenue than a brief outage in a limited set of locations.
This disparity highlights the importance of focusing on performance metrics like LCP, which directly impact user satisfaction. Availability alone can be misleading; it’s not enough to represent user experience. While both companies had issues during the peak holiday shopping period, the performance XLO suggests that Company_B fared worse.
The best way to avoid failing the peak shopping season is to set both availability and performance-based XLOs. Burndown charts, like those above, can help you project whether you’re on track to meet or miss your performance objectives from a real customer perspective. Ultimately, that impacts revenue. Given how much is at stake during the holiday shopping period, customer experience must be a top priority.
To learn more about how we can help you stay resilient during peak traffic periods, check out our Internet Resilience Program
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。