We have been focused on improving the performance of our Website for the past 11 months to end of September 2020.Â
This journey started in November 2019, leading to the publishing of an eNsight titled “7 key insights about eNitiate Website stats you can learn from“.
It is time to share the 11 optimisation lessons from the 11 months of improvements that resulted in eNitiate Web traffic growth, using data from Google Analytics for support.Â
Table of Contents
Our definition of optimisation
There are 3 main elements to our definition of optimisation:
- development of high quality content;
- discoverability, alternatively called search ranking optimisation (or SEO); and
- positive user experience.Â
For the purpose of this eNsight (commonly called a blog), I loosely use the term optimisation to refer to the second element – discoverability.
And now, let me take you through the 11 optimisation lessons from the previous 11 months of focus on our Website.
1. Tech optimises, but humans decide
As part of our development of eNsights, we test for the most search ranking optimised titles, using a tool called Headline Analyzer.
The tool scores each tested title out of 100, based on an algorithm that incorporates various factors, including a combination of common and uncommon words, sentiment inducing words and the length of a title.
The higher the score, the more optimised the title for search rankings.
Here below are 2 titles that we tested recently, for an eNsight we will publishing in due course.Â
As can be seen from the scores above, the title in Figure 1 yielded the higher score of 72 (out of 100).Â
However, my gut brain tells me that the title in Figure 2 is more engaging, despite its low score of 48.
The example above is one of many that demonstrate that humans must ultimately decide, and at times even disregard optimisation recommendations made by Web applications they use.
2. Good looks don't make up for lack of substance
In the beginning, I always wondered why big sites such as Google use limited to no design graphics on their various Websites, until I discovered that colours add to the size of a Web page, and can adversely affect its load speed.
But importantly, the blinding glimpse of the obvious is that an average visitor lands on a site looking for content related to a search term or phrase, not looks.
Simply put, a good-looking Website with poor quality content will not get decent traffic, or will suffer from high abandonment rate by disappointed visitors.
- Jan-Sep '20 vs Jan-Sep '19
We deliberately went with a minimalist look and feel for our revamped Website that we launched in January 2020, and upped the development of high-quality content instead.
The result? We almost tripled Web traffic from organic search engines since then, compared to the previous year.
In my mind, the growth of “free” traffic that our Website experienced should be every Web-based business’s dream!Â
3. Mobile first? Yes, but not at all costs
Mobile-first Websites have become the talk of search engine town for years now.
eNitiate effected major changes as part of its latest Website revamp, and as part of ongoing content development, to move in this direction.
However, ignoring desktop visitors is not a sound strategy for us, for 2 reasons.
First.
A substantial number of visitors to the eNitiate Website are still using the desktop device.Â
Second.
Of the sources feeding traffic to our site, there are discernible device-influenced skews for organic search and referral (and mainly Whatsapp) sources.
See the 3 Pie charts below, based on the Jan-Sep ’20 period.
This observation must be read together with lesson 2, with specific reference to growth of organic search for our site’s traffic.
NOTE: Google Analytics combines tablet and desktop devices for the traffic source metric, but the tablet’s contribution is negligible (Graph 1) and thus it does not have a bearing on this finding.
Based on the provided 2 reasons above, the importance of both devices will be treated equitably on an ongoing basis.
4. Not everything that can be optimised, should be optimised
The golden rule in blogging is to write for humans and not machines.
For this reason, too much emphasis on optimisation can actually impact negatively on the user experience.
If Web page optimisation was the single-most important measure used for our site, we would not be using images and embedded videos and tweets in our eNsights to the extent that we do.Â
Why? Because these two features, which enhance user experience, affect page load speeds.Â
Here are 3 examples to make the point, where we calculated Google Pagespeed scores (out of 100) of the following 3 different Web pages:
Figure 3
The Homepage:
This Web page has a limited number of images that are also light. Also, the page does not have Youtube and tweet embeds.
The Google Pagespeed score is in the amber zone for the page, which is acceptable by our standards.
The page takes 4.1 seconds to fully load, accordion to GMetrix, and this load speed is also decent.
The existing third-party code that is mainly made up of Google fonts takes only 480 milliseconds to load.
Clearly, this page is adequately optimised.
Figure 4
eNsight about South Africa’s digital adoption rate:
We used 14 carefully selected images with mainly large sizes in this eNsight, and there are no Youtube and tweet embeds.
The Google Pagespeed score is in the red zone for this page.
The existing third-party code that is mainly made up of Google Fonts takes 1,040 milliseconds to load.
The page takes 5.5 seconds to fully load. Not great, but not too bad either.
Figure 5
The 2nd eNsight in the “O Jewa Ke Eng” series.
We used 30 highly relevant images that are mainly tweet screen grabs, and 1 Youtube and 15 tweet embeds.
The Google Pagespeed score is in the red zone for this page, and it is the worse score of the 3 examples – not a surprise given the combination of images and third-party embeds as described above.
This page takes 16.8 seconds to load, which is way too slow.
As I indicated at the beginning of this lesson, there would not be so many images plus the third-party embeds in the O Jewa Ke Eng eNsight, if optimisation was the primary motive.
Notably, and despite the poor pageload scores, both eNsight examples in this lesson are in the top 10 eNitiate Website pages for the period Jan to Sep 2020, and in fact the worse performing of the 2 – O Jewa Keng – ranks two places higher, at top 5 spot.
I can confirm that the bulk of traffic to the two eNsight examples came from organic search engines.
For us, quality of content and user experience trump poor load speeds.
This lesson is to be read together with lesson 10.
By the way, this very eNsight’s Google Pagespeed score is 39, the page loads within 3.8 seconds, and there are 12 images and no Youtube and Tweet embeds.
5. Once is never enough
It has taken us 11 months to get to this stage where we grew our Web traffic by 56.71% in the last 9 months vs the previous year.
And the journey is still far from over.Â
Read more about the process we are still on here:
6. Aggregated numbers don't tell the full story
Aggregated results can send you into a lull of false sense of satisfaction, or unnecessarily raise your blood pressure into dangerously highly levels.
What we have found is that the real story lies beneath the aggregated numbers, and it can be calming and reassuring if you have been doing the work.
Let me demonstrate, using a comparison of average time spent on the site and average time spent on top 10 pages by visits in minutes and seconds, for the Jan to Sep 2020 period.
01:57
Average Time on Site
Per Visitor
04:52
Average Time
Viewing a Top 10 Page
The comparison of the times above show clearly that visitors spent more time on the top 10 pages, as compared to the time spent on the site overall.
You can read more about the issue of aggregate Web stats here:
7. The 80:20 principle is the best approach to effective optimisation
We have a relatively large Website, using the more than 550 published pages as a measure.
The perfect way to optimise the site would be to test every published page, 3 of which we used as examples in lesson 4, so that we can determine optimisation issues in each and fix them accordingly.
However, this is not the most effective way to optimise a large Website.Â
The return will not justify the time and resources invested.
The answer? Optimise the pages that are responsible for the largest contribution, based on a key Web traffic metric.
In the case of eNitiate, the top 10 most visited pages have been our focus of optimisation efforts, as reinforced by the two stats here below:
It is more effective to optimise 2% of the site’s published pages that contribute over half of the pageviews.
8. Measure what matters, not what appeals
There is a plethora of Google Analytics metrics that can be used to measure the performance of a Website.
However, a careful analysis of a Website’s traffic, informed by Web performance objectives, can guide the selection of metrics that matter. Â
We share our selected Web metrics and ROI goals in the related eNsight below.
I invite you to check out how we use our selection to measure the performance of our site in the related eNsight.
9. A solution for one technical issue can be a poisoned chalice
Third-party applications, which have become the standard for Website development and maintenance, are funny like that.
Adding or updating one Web application can lead to a conflict with others on the site, which can trigger a malfunction that persists for days and weeks at a time, or even crash the Website in the worse case scenario, as we experienced in 2015.
The last few weeks have been especially hair-raising for us, with a flood of software updates by the WordPress platform that we use for our site, and some of the main Web applications (popularly called WordPress plugins).
The latest updates in the last 2 months, some which were major, resulted in incompatibility between some of the applications, and this affected the functioning of the site, which had an impact on user experience.
At least one of the applications – called Essential Addons for Elementor – has a bug that is still being fixed as I publish this eNsight.
Unfortunately, the issue of conflicting third-party Web applications cannot be prevented, but early detection and resolution are possible through vigilance, regular site backups and use of Web applications that constantly monitor the functioning of a Website.
10. At times, a work-around may be the best answer to a technical issue that may not have a solution, such as a page speed hump
In improving the performance of the eNitiate Website, we have kept a healthy and sometimes delicate balance between positive user experience and optimal search engine optimisation.Â
As I have indicated in lesson 4, we have taken the decision to live with some of the top 10 pages that load at sub-optimal speeds where we felt quality of content and user experience may be adversely impacted by (further) optimisation efforts.
But we continue to explore work-arounds for the pages with slow load speeds, including:
- using content distribution networks (CDN’s) such as Cloudflare to bring the Website closest to physical locations of our visitors across the globe;
- using image compression applications such as Smush;
- using preloading screen applications such as Preloader Plus to bridge the gap with the long waiting times, and
- using lazy loading of images with Smush, which allows for loading text first to reduce visitors’ waiting time.
11. Only a great host will give your Website the best party
- Value for money.
WPMU’s hosting services care bundled with features such as SSL certificate and staging site facility; and optimisation applications – some of which are highly recommended in the industry, including Smush and Hummingbird.
- Great uptimes.
A Website is a 24/7 online office.
It is important that this office is open all the time, so visitors can come in and check what a business offers at a time that is convenient to them.
Only a reliable server hosting service provider can guarantee 100% uptimes, and WPMU does.
- Great technical support.
In addition to a ticket-based service, there is a 24/7 live chat service that is staffed by people who really know what they are doing and are always willing to go the extra mile.
What’s more, the live chat queue waiting times are reasonably short.
Our Website was previously hosted by a different service provider that shall remain nameless.
We are absolutely happy that we switched to WPMU for the 3 reasons I advanced.
And just to be clear, we are not paid to mention our current service provider.
The road ahead
In echoing lesson 5, optimisation of a Website, balanced with improvement of user experience, and both underpinned by continuous development of high quality content; is a never-ending exercise.Â
We are constantly keeping a close watch on the performance of our platform, and effecting ongoing improvements, which have so far led to increased traffic.
Clearly, we have created a virtuous cycle, and so can your company too.
