4.1.1. Mobile Site Speed Test (MSST)
Running the MSST, a set of data was centralized in the content analysis in a classified way and according to the predominant networks in Romania, the 3G and 4G.
Loading speed. Testing the touchpoints’ load speed, the MSST was running to account for how long a user has to wait for the content to load. Depending on the number of seconds recorded, is acknowledged the usefulness of the page and perceived the straightforward modus operandi.
In the 3G network, the uppermost situations are represented by the online stores that show a fluctuating loading speed from one touchpoint to another. Specifically, the waiting time becomes longer as the users leave the home page and land on the product page and it shortens as they are moving forward from the product page to the shopping cart. Even if the loading time lessens as the users draw near the checkout, the load speed of the shopping cart page is slower than the average load site speed in all situations (Figure 1
). The slightest difference between the value of the average site load speed and the value of the shopping cart load speed is showcased by the pentruanimale.ro, where the average site load speed is 4.1 s, the shopping cart recording 5.2 s.
Three out of 16 retailers expose the increment of the number of seconds as the consumer draws near the checkout and the shopping cart page is the one that loads the slowest. Two of them, f64.ro and officedirect.ro, presents substantial increases (Figure 2
Regarding the stability of the loading speed during the customer journey, only three retailers masterly managed to maintain a similar page load speed when users added the products to the cart. In this manner, the marosbike.ro, buy4baby.ro and heliumking.ro do not negatively influence the user experience, providing a constant virtual search medium during the purchase journey. This veracity does not evoke a special perception of users about the performance of the sites, but unforcedly it does not provoke a misperception about a certain deficiency of ease of use or other barriers in the journey. Thus, their experience is not affected and if there is already a purchase decision, the conversion rate does not decrease and the users will fluidly journey to the checkout. The same cannot be conveyed, however, for three retailers that show the increased loading speeds’ values as the distance between the touchpoints shortens and the impact rate becomes increasingly significant. The users that experience blockades during the loading process change their perception of ease of use, especially when they access the shopping cart touchpoint and it loads slow. In case they wanted to make a purchase, this does not happen anymore, the conversion rate is reduced by 20% for each delayed second. Only two exposed a decreasing loading speed during the journey, but none of them presented an optimal value for the 3G network (less than 2.5 s).
Improvements were observed in the 4G network (Appendix A
). The variance for each retailer is displayed in Figure 3
Additionally, the 4G network presents a page load speed that increases as the consumer approaches the checkout. Compared to the 3G network, the situation is changed in the case of helloholidays.ro, which in the 4G network, maintains the speed for the home page and the product page at 3.6 s, increasing the speed of the shopping cart to 3.9 s. On the other hand, officedirect.ro shows a very big difference when switching from the product page (5.4 s) to the shopping cart (10 s), even if this time the values are lower. The generated effect on users is negative. Within this network some retailers have a changing loading speed, oscillating between increases and decreases in loading time. Seven of them present a changing speed generated by the increment of the number of seconds when switching from the home page to the product page, followed by a decrease of the shopping cart page’s load speed.
This is maintained only in the case of two, marosbike.ro and buy4baby.ro. The ideal case in which it decreases as the user approaches the checkout is for a single online store (cafelepremiate.ro. However, every touchpoint of this retailer presents a page speed that needs improvements because of its high values. Therefore, a good example can be described in the situation illustrated in Figure 4
. There, the process of the user who accesses the www.pentruanimale.ro’s website [53
] and reaches the shopping cart page is presented. The shopping cart touchpoint has a good rating compared to previous touchpoints, presenting the recommended loading speed during the customer journey.
All other retailers have a product page and shopping cart load speed over 2.5 s, generating a poor rating that needs improvement to meet users’ expectations. Therefore, the authors considered it necessary to mention the rating that predominates in both 3G and 4G networks.
Rating. In the 3G network, there are three situations with results that show a strong need to implement improvements. These are represented by the average speeds of the sites and not by any analyzed touchpoint. This allows highlighting the fact that no studied touchpoint can meet the requirement of users to load fast and generate satisfaction in user perception. The fact that 100% of retailers have a low rating in the 3G network, creates the opportunity to improve the consumer experience by reducing loading time. Inamedicalsport.ro, marosbike.ro, and manier.ro have a rating that needs improvements regarding the site’s load speed, but when it comes to the performance of the touchpoints’ pages, the performance is non-existent. On the other hand, in the 4G network, inamedicalsport.ro and manier.ro no longer show the need for site improvements, presenting a good rating (2.2 and 2.5 s). Cafelepremiate.ro, with a loading speed of 2.5 s, and heliumking.ro, with 2.3 s, also join with a good rating, so exist four retailers that have an optimal site load speed. Only one has a slow loading speed and it is rated with a poor rating. This is the case of auto-soft.ro retailer, which has a site-level loading speed of 4.3 s. The rating and the long waiting time represented by the number of seconds discovered in the analysis, led the authors to identify solutions to improve the touchpoints pages load speed included in the study.
The monthly trend. The unfavorable situation of the retailers studied from the point of view of the satisfaction of connected consumers is underlined by one that was discovered, registering changes in the last month regarding the performance of the site in terms of speed loading This is f64.ro, which shows a fast loading speed, the performance at the site level is improved by 6 s in the 3G network. Unfortunately, in the 4G network, it no longer records this performance, like the others that do not show any change in both the 3G network and the 4G. This result convinced the authors that they do not continuously increase performance, to seek to improve the user experience and thus optimization of the customer journey.
The evaluation generated reports that all retailers received recommendations in two directions, for marketers and developers. For marketers, the report presents opportunities to improve the loading speed, to deploy customization, and give access to a fast purchase process. A checklist was exported for developers based on how they can make improvements to the studied pages.
In addition, 40% of users are willing to spend more than they planned if they encounter personalized experiences at a high level. Therefore, the report generated presents four opportunities that retailers have to improve the user experience. Use the Optimize platform to continuously improve a personalized user experience, implement experiments using A/B tests, re-engaging users with push notifications with relevant content, and use Progressive Web Apps (PWA) to facilitate site access, even offline.
According to reports, to ensure fast access to the purchase process, the online stores must allow authentication with a Google account, using the Autofill option for an easier checkout process. How the payment is made can be a challenge for retailers, being often the main barrier in mobile transactions, determined by the massive abandonment in the checkout process. By creating an easy environment for concluding transactions during this touchpoint, users will convert more.
Developers can use the available checklist exported after running the PSI test. It consists of seven key points in increasing mobile loading speed:
Optimize the quality of your user experience by focusing on the metrics that matter most according to Core Web vitals
Improve the site speed with the custom recommendations from report
Check if the site is an installable and exemplary PWA
Implement Google Optimize to run server-side experiments against variants.
Enable Google Identity Services and Autofill to build a system that allows seamless sign-in
Add GPay to your existing payments processing stack to offer simpler, more secure checkout
Connect the web pages to the app pages with deep linking for both Android (App links) and iOS (Universal links).
4.1.2. Page Speed Insights (PSI)
The PSI test ran to identify the characteristics related to the loading speed, especially the one perceived by users, interactivity, visual stability, and perceived utility. All data were centralized in the content analysis and were interpreted according to its structure.
The performance score generated by running the test includes almost the entire set of retailers analyzed in the range 0–49, which means that they have a poor rating of site performance. The only pentruanimale.ro has a performance score related to the need for improvement, with a performance score of 53. This score reflects the performance of the following features.
In the case of all studied retailers, FCP is over the limit of users’ expectations, registering between 1 s (cafelepremiate.ro) and 4.8 s (auto-soft.ro). Given the role for which this parameter was introduced in the analysis, namely that it analyzes the perceived loading speed, measuring the time from the beginning of loading the content until its appearance on the screen, the results show that most studied touchpoints are perceived as less useful. For example, auto-soft.ro records the following timeline when loading content (Figure 5
Compared to cafelepremiate.ro, which has the performance of displaying content faster, it happens that auto-soft.ro loses conversions from the first contact with users. On the other hand, sdhouse.ro manages to display the content almost instantly (Figure 6
), at the first touch. The chronology of the content loaded is shown in the following figure.
LCP. By completing the FCP and measuring the perceived loading speed when the main content of the page has been loaded, the analyzed indicator guides the authors to identify the necessary improvements so that users perceive from the first interactions the usefulness and ease of use of mobile online store pages. The values under 2.5 s of this indicator places the f64.ro, inamedicalsport.ro, and heliumking.ro in favorable positions from this point of view towards users, while the rest of the retailers face the opposite situation, in which the pages are difficult to access.
FID. Although most of them face the challenge of displaying the content on the accessed page as soon as possible, the first impression that they provide to users is a good one, assuring them of the functionality and usefulness of the page. The experience that users feel at the moment of interaction with the first pixels is the factor that determines them to wait further, until the moment of displaying the content (FCP). This was concluded by the authors, because the FID values obtained are in the optimal range, representing an opportunity in the customer journey.
CLS. Unfortunately, 12 retailers recorded CLS values outside of the optimal range and one of them (f64.ro) shows a very high value (3.32.). This indicates the need to implement the opportunities and solutions that the online stores need. The best ones in terms of this indicator have values between 0–0.6. The range provided the opportunity to conclude that users consider that the site’s cellini.ro, pentruanimale.ro, and marosbike.ro are visually stable and attractive.
These being mentioned, according to the Core Web Vitals assessment, no retailer out of the 16 studied passed the evaluation. Taking into consideration the distance of each touchpoint from the point of purchase and the impact rate assigned, the authors state that there are online stores that do not carefully drive the users to the checkout. Their attention to the performance of technological features must be amplified. If the performance of the loading speed improved, the results regarding the completion of the customer journey would not hesitate to appear. Moreover, the number of conversions would increase, and the bounce rate, but also of the shopping cart, would decrease, which would be reflected positively on the business income.
To propose concrete solutions in the Discussion section, 14 recommendations given by Google were highlighted to improve the experience on mobile devices by reducing the seconds required to display the content, their implementation determining a certain saving of the specific loading time (Table 1
Creating the Pareto diagram, on the horizontal axis were located the resulted opportunities after running the tests. According to the information presented in the table above, Figure 7
illustrates that relatively 80% of the effects produced by the implementation of opportunities are caused by 20% of them.