Feature image for webdesign post, showing wireframe sketch for a page.

Thoughts on WordPress hosting and design

Choosing the best host for your website

Intro

This will be a rambling viewpoint more than authoritative post, but I’ve been doing web design and hosting for well over twenty years, so it’s backed up with experience. This experience includes all sorts of companies and hosting options (shared, VPS, dedicated and cloud-based servers, managed and unmanaged).

The latter varies from small start-ups upwards. It includes great companies and others that don’t appear to know what they are doing(!), hosts that get bought up and immediately result in a massive drop in service quality. Or downtimes of over a month – were the uncaring host says, “We are having technical difficulties, read the small print of your contract.” If you are a web developer with a story beginning, “There was this one host…”, well, I feel your pain.

If I name names, it will be a case of “this happened in the past, they may be better, now”, or it will be, “as at the time of writing”.

On affiliate referrals

What you will notice with a great many hosting reviews – the majority in fact – is affiliate links. Nothing wrong with them in general, some sites pull in incredible revenue specialising in it and this allows them to research and write some great articles. And some are in it just for the money, so are less trustworthy. Let’s just say ‘Trusted Reviews’ aren’t quite trusted by every. So, any links I give will be direct. I’ve done advertising in the past, I add and remove Google Adsense and the like from time to time, but mostly I am cynical about it. (Google Adsense slows your site down, as does Google Analytics; everything has a cost).

When a post says “These are great, we recommend them!” and some or all of the links end in things like /?ref=thiscampaign or /?afftrack=thiscampaign, you can almost guarantee the amount of commission for a conversion affects the choices and review, even if only subconsciously. Apparently wordpress.org do the same. If the ‘best’ site doesn’t offer them much and the most popular choice pays them for millions of sales, that’s just business. “We can’t can’t review everyone, so let’s just recommend the one’s that best pay our bills.”


Show me the money?

Apart from Adsense on this site, which makes pennies and I keep for curiosity (stats), I don’t charge nor make money from web design. Nice if I could, but I don’t try. I do play around with the idea of developing as a business from time to time, but I just don’t want the hassle of customers and tax and all the rest. I do host and maintain sites for myself, friends and family, for charities, but either at a loss or for bare costs (read: domain registration is $11.99 a year, you can pay that).

In short, I am retired. I do this out of interest and to help people, not for money. Any and all running costs, maintenance, software, I cover. It is not important; I have no staff, no offices, no overheads I can claim back, and don’t need technical support. You might want to bear that in mind if you are a reading this.

Also bear in mind as you read this, when I started doing web design, way back when

Mobile phones were like house bricks.
CMS was years away; WordPress didn’t exist until 2003
Google (1998) didn’t exist, we used Altavista and many other early search engines instead.
Even PHP (1995) was still in its infancy.
As for CSS (1994), it was so poorly supported and unreliable it would be nearly another decade before it was worth touching.
Broadband was years away and dial-up was king. 64kb ISDN leased line costs were ridiculously expensive, but a T1 line (1.544 Mbps), well, I was quoted £35,000 just for the installation. One BT manager (in the mid 90s) told me “they’d never allow broadband to interfere with their lucrative leased line business”.

(Ackadia is as old as Google, registered with weeks, and is one of the relative few continually surviving sites from the 90’s. There are an estimated 2 billion live web sites now; back then there was only 2.4 million. How many of them do you think stayed active for another 24 years?)


Hosts I’ve used: early days

Some of my reasoning is financial, some subjective, some functional. If I am not happy with a host, I move.

Some people – the majority in fact! – put up with poor service for two reasons.

The first is laziness and apathy.
“I know I should, but I can’t be bothered with the hassle.” Understandable, especially when it comes to moving hosts and changing DNS setting and downtime and all the other hassles. But people do it with electricity providers, mobile phones, etc. The savings can be substantial, so it’s a frame of mind you need to address if it applies to you.

The second is called “the sunk cost fallacy”.
It’s a cognitive bias, you’ve spent the money, paid upfront, and if you move, you’ve lost it. So, it’s a loss aversion bias, which all animals have. It’s hard wired into us via evolution. The problem is things like T&C; you’ve paid for the hosting, you have agreement to the contract, you can’t get your money back. So what can you do but put up with it. And then it rolls over and renews and you’ve paid up again because it slipped your mind, so you stay another year, which loops back to the first point.

It’s like the BBC in the UK, It costs £165 a year, whether you want it or not. It’s a tax, essentially. You can opt out, but they make it hard and layer it with threats, so most people don’t. We still pay the license, but if it was just me, I’d tell them to take a hike as I don’t care about scheduled TV.

Hosting is the same. You get entrenched.

I can’t remember all the companies I’ve used. Some were so awful I don’t even want to remember; So much stress, so much corporate indifference. Others I will happily recommend, with caveats.

GoDaddy
For instance, I was with GoDaddy, who I wholly, 100% recommend for buying domains from. They also do hosting, so I have had servers with them. It was a long time ago, but they implemented some support change over Plesk or CPanel and WHM and whatever it was, I didn’t like it. It made more work for me and added more costs, so I moved. This is not a reflection on GoDaddy, but an observation that T&C and service can and does change over time and if you are not comfortable with it, you have the option to move. Either write-off the investment now, or make plans to move in advance.

You can do it without overlapping, but if a backup corrupts or runs into compatibility problems, you’ve lost a site or you’ve made an unholy amount of repair work for yourself. So, you overlap. Servers, here; servers, there; servers everywhere!

HostGator (I think!)
Anyway, I moved to Hostgator for a few years, upgraded and downgraded a few times, then ran into a problem which they couldn’t sort, a hack of some sort, not sure other than it being Russian traffic. Could of been a plug in I’d added, or perhaps a backdoor into Apache, never did find out, but we couldn’t keep them out after that. Even wiped the server, reset everything. Ran it clean, no plugins, and they were inside within an hour. In general, I have no problem with Hostgator, but the experience soured my opinion of them.

HostGator (for sure!)
I wrote about these back in 2017, after my site went down. There was a technical problem which they denied for a week or so before admitting, “Hah, hah, funny story, we somehow accidentally blackholed your site’s IP.”

I didn’t laugh, I moved.


Hosts I’ve used: in not so distant past days

WordPress.com
Their standard budget offering is useless as your themes and plugins are seriously limited, so you really need the business option. In it’s favour, it is fast and fully featured; against that, it’s as expensive as heck, especially if have more than one site. For a business with one site, they are great. If you want to host several, it’s a hard pass.

Fun fact: VaultPress is a service from Automattic, the same team that owns and operates WordPress.com. Vaultpress (for wordpress.org) is (or was) incompatible with Vaultpress (for wordpress.com). So, if you’re on a .org hosted site and faithfully backup up and decide to move to wordpress.com, you might want to check on that. (The same would apply going from .com to .org).

Bluehost
These are actually a sister company to Hostgator and I can’t recommend them at all. Like their sister, they appear indifferent about your server or sites going down for a length of time. I did not stay with Bluehost for long!

This issue made me move Ackadia to WPEngine the next day, but it’s not why I moved the rest of my sites.

To be fair, I’m not terribly patience with slow computers. I can sit and happily binge through several series on Netflix, but you make me wait 4 seconds for a page to load in WordPress and I’m ready to kill someone. I think the CPU usage on one of my dedicated servers (years earlier) was 2% on a full load, but it was so fast. (Then the quote to renew comes in and all those zeroes are scary! “Honeymoon is over, son, time for your big boy pants”. When you see 4 zeroes in a renewal quote, you better run!!)

Anyway, I got to the point were updating sites for people, even so much as changing a title was infuriating. So, I upgraded. And it was still too slow! I did consider upgrading further but you get to the point were you’ve have enough and walk away.


Hosts I’ve used recently: WP Engine

WP Engine
Apart from one quibble* which very much soured my opinion of the company, these are great and I would otherwise thoroughly recommend them.

OK, they are not cheap, and if I’m honest, after the first site (e.g. hosting 10 instead of 1), they are really stingy with the allowances compared to similar competitors. Still, the speed, service, support everything is great. I did consider upgrading and moving all my sites to them, but nothing is as scary as nothing – when that ‘nothing’ is a zero added to the annual fee. Besides…

*It was something and nothing, but annoyed the heck out of me. It’s an honour thing. I bought a lifetime agreement with StudioPress for their Genesis framework and plugins. WP Engine bought out StudioPress and promised to honour this. They don’t, or at least not in any honourable way. I paid something like £240 for this several years ago, when StudipPress were still a new company. WP Engine want $360 a year (+tax) for it.

I just wanted to play with the Genesis Pro plug-in to assess it. Chris Garret (now Managing Director of WP Engine) announced StudioPlus customer could have full access to Genesis Pro, as agreed. But when I tried to use it, it wants a WP Engine code, which they refused to hand out, saying there was no such agreement and if I (we) wanted it, we had to subscribe. Naturally I vehemently disagreed and they relented, saying yes, he did say that, we will honour the agreement.

Me: “So, can I have the code?”
WP Engine: “No!”

Corinne Olson (WP Engine Support)

Apr 12, 2022, 7:37 AM CDT

Hi Paul,

I am sorry for the confusion and that this did get overlooked by some of the support team. We will be sure to send out a reminder for the support team.

As a Pro Plus member you will not have access to a subscription key at this time to use Genesis Blocks Pro or Genesis Custom Blocks Pro. The key is only used to enables auto-updates it does not affect how the plugin works and Pro Plus users will have to manually update by downloading a new copy through their StudioPress account when an update is released.

They are more or less honouring the word of the agreement, but in no way honouring the spirit of it. A plug-in that you can’t auto-update and is recalcitrant in the extreme, is not secure and is absolutely NOT something you want to design a website with. I have issues with money-grubbing corporates and, to me, this reeks of it!

Besides, if you were going to spent that sort of money on a plug-in framework, you want Elementor (Pro)!

Update Genesis Blocks Pro
Updating Genesis Blocks Pro can be done automatically or manually depending on where you accessed it (through your WP Engine hosting account, a Genesis Pro subscription, or your StudioPress Pro Plus account).

If you accessed the Genesis Blocks Pro plugin through your StudioPress Pro Plus account, you will need to update it manually.

WP Engine: Update Genesis Blocks Pro

c.f.
Many Pro Plus all-themes-package customers have lamented that Genesis Pro seems to be “double charging” Pro Plus customers for themes since the Genesis Pro package includes themes that Pro Plus customers also receive. In response, Pro Plus customers will now be getting Genesis Blocks Pro and Genesis Custom Blocks Pro for free to download in their StudioPress portal on or around June 8th, 2021.

You asked. We listened. Announcing Big changes to StudioPress! (Chris Garrett, May 11, 2021)

What does this have to do with hosting?
OK, I will admit, I am paranoid, both about security (rational) and trust (less so), but bear with me.

Old customers can – with hassle and security concerns and error messages – use the plug-in, so what is the problem?

“Well, Bob, I’m glad you asked that!”

I can use the software as if I was a lifetime subscriber (which, in fact I am/we are), but they have intentionally made it as awkward as possible to do so. They want that $360+ a year. Period. On face value, it costs them nothing to give Pro Plus customers autoupdates, so why not? Why? Because, I believe, if even one person gives in and pony’s up that $360+ a year, they made buck. It’s all about the money. Show me the money! They clearly did not and do not actually want to honour the agreement, so, in typically corporate lawyer spiel, went with: “We have done nothing wrong, legally, within the word of the law.”

So, it’s about trust.
Would I trust them with hosting my website? Yes. No question about it. Can’t fault their hosting provision. (But see later, what people with higher traffic have to say).

Do I trust them as a company? Would I give them my business again?

Hint: I moved servers again, twice, one far more expensive that WP Engine. I did not and will not give them any more money or send them more business because – in my eyes – they are dishonourable. I cannot trust them.

I can state as an absolute fact that this one arguably spiteful decision has already cost them several thousand dollars in lost sales. It’s a business thing. You do something great, a customer will tell a few of their friends, if you do the opposite, they will tell all of their friends.

For instance: Harvard Business Review (2002): Pricing and the Psychology of Consumption

Communications Pro (n.d.): What Does a Bad Business Reputation Cost?:

"Sales Revenue Will Fall
According to a BrightLocal survey, 93% of consumers read online reviews. A separate study found that four out of five customers won’t buy from companies with negative reviews."


Hosts I’m currently using: Siteground

Siteground

I have mixed feelings about this company. I moved to them from Bluepost as the best low-end hosting offer I could find at time. They keep making changes (to make them more money) which makes me twitchy, but they are not a bad host to try.

The “try us” option is just £7.49 (plus vat) per month for the all you can eat Geeky option. 75% off, it is a great deal. After your taster (of up to a year) the honeymoon offer is off the table and it jumps up to £30 a month, plus vat. A few quid a month is a no-brainer, when you want to renew, it jumps to over £430 (inc vat). Not so cheap. Is it good value at that price, with the changes they are doing? Well, I have my doubts.

For a start, they have pulled the plug on their partnership with Cloudflare. If you like or rely on Cloudflare, tough, look elsewhere. I would note they dropped Cloudflare after I subbed for a year. Trust issues. They do have their own CDN and will (snatch your hand off) offer you a slightly better "Premium" CDN for the princely sub of £115 a year PER SITE. For a 10-user hosting, with premium, the hundred quid a year is now at well over £1,500 a year (£130 a month) and now terrible. Bait and switch comes to mind!

{ Siteground drop Cloudflare support }

They made more changes again a few days ago; I didn’t pay a great due of attention to it because it was “Our new address (for you) is London, UK, this means <insert boring corporate reasons>” I’d already decided I was fed up with them by this point.

Plus, I had questions about TTFB and, in technical circles, found alleged claims of VERY questionable business practices. I cannot saying they are true or not, but I considered upgrading to a business cloud to improve page speeds (‘cos their Premium CDN did nothing!) but the more I read, the more wary I got. I am, as I said, paranoid, and if you look for a thing, you will find it aplenty. But, no smoke without fire. As I said earlier, bad reviews hurt a company’s bottom line (whether they are true or not).

From YOUR point of view though?

Bob: “How was Siteground to use compared to say Bluehost or WPEngine?”

Me: “Well, there are two ways you need to consider this, as a developer, operating inside the engine, and as an end-user, accessing a site. For both, Siteground, even at the Geeky level, is leaps and bounds ahead of Bluehost. If you are using Bluehost (or similar) and not enjoying it, absolutely, give Siteground a try (unless lack of Cloudlflare is a deal-breaker, of course).

Page loading speed is a different matter. Don’t get me wrong, Siteground can be good, great even, but it ain’t WPEngine. If actual web speed matters (as in, you as losing conversions), then you have to ask yourself if the losses outweigh the hosting savings. Absolutely no disrespect to Siteground here, but they are more at the budget end of the market. Within that niche, they are the best I’ve used and I would – do – recommend them for that use.

However, if your hosted page take 8s to load on mobile and the host your nearest competitor is using only takes 3s, you are losing most of your customers as they walk away. According to Google (2017), 53% of visits are abandoned if a mobile site takes longer than three seconds to load. By 6s it’s over 100% abandoning your site. It just keep getting worse after that point.


Hosts I’m currently using: Rocket.net

Bearing in mind the previous paragraph…

I made some changes to my theme (swapping from an mobile-ready response theme to a visually similar Astra theme) and moved a site to Rocket.net. The difference was dramatic! I had added NitroPack to the site (on Siteground) but it was still just too slow for my liking. Maybe I should have tried Astra on Siteground, and maybe Siteground are or will do something about their slow TTFB, but those are maybe’s.

The fact is within a week I saw the mobile usage jump 2,000%!

Graph of web traffic improvements with mobile optimised theme and fast hosting
{ Web traffic improvements with mobile optimised theme and fast hosting }

That figure is down to 365% now and will continue to fall in line over time as things balance out, but 50% of my traffic is now from phones. It was about 10%, then I told Google, “Hey, guys, check out my new body!” Well, <body> and such, but it got their attention. Page speed (still dragged down ‘cos of all the slow and ugly Google javascript!!!) dropped from 8s to under 3s. Google’s opinion and rating jumps up from 57% to 97%.

More than that, which would just a change is usage rather than numbers, the actual traffic jumped up 60%. As is still rising. People were staying longer, exploring more pages. Basically, fitting in line with Google’s research, getting my mobile page speed below 3s resulting in roughly 60% of visitors NOT abandoning my site. Unique visitors jumped from under 300 to around 700 a day. For a business or eCommerce site, that could translate to doubling your turnover.

You can’t tinker with it, nor get the statistics, which is a shame, but Rocket.net use Cloudflare Enterprise with WAF. This normally costs a bucket load of money. It’s several step up Cloudflare price plans, which jump up like this: Free, Pro (£20 a month), Business (£200 a month), Enterprise (call us). As in “If you have to ask, you can’t afford it.”

With caveats, they also handle migration for you. I’d rather do it myself, but it’s how they roll. Less work for me. The caveat is it’s not simply a Nameserver edit, you have to roll your sleeves up and delve into the DNS, possibly ripping out unwanted A and MX entries, adding in CNAME and TXT entries etc. I’m fine with that, most experienced developers will be, and there are instructions, but less experienced users will be mortified. (It can break your site in so many ways if you mess it up!)

If I’m honest, I’m ambivalent about promoting Rocket.net. Do I recommended them? All day long, 100%. At least from my experiences so far. OK, not ambivalent then, selfish! They are great; I don’t want a bunch heathens piling in and turning them into another Hostgator! Get you gone, this is my spot!


Other hosts

Bob: "This is great, but why haven’t you mentioned…"

What, Kinsta? Dreamhost? Flywheel? Cloudways? I am aware of them and others, but haven’t used them, so I can’t comment. Also, if 10 sites review 10 hosts, you will not get 10 sites agreeing on the top one, or the top 3. Too many variables, for a start. Theme and site tested with, chosen hubs, monetary biases confounding choices.

A quick Google and Matt tells me WPX is THE best (in his opinion). I close the tab and a screen-filling pop-up appears (how rude!) and suggests I give them a try. Yep, that guy is on commission.

Another says the top three are *drumroll*:
SiteGround
Bluehost
and WP Engine

Having used all 3 – and moved away from all – I would have to disagree, though my comments for and against all three stand. Actually, I don’t think I had a good word to say about Bluehost. They are cheap; there you go. I was Googling for their prices and instead got a Cloudways page. Obviously they are going to be biased in their own favour, but here’s what they had to say and – from personal experience – I agree with every point:

Siteground, I would still say, is decent enough at the low end of the market, for a year, to try. Bluehost I wouldn’t touch with a bargepole. I lost all trust in WP Engine as a company, but I can’t say a bad word about their hosting, it’s top notch.

Bluehost websites fall apart when you try to scale up

Bluehost is cheap when you are getting started and doesn’t charge you for traffic. Sounds like the perfect fit for small businesses, but here’s the downside…

  • You’re locked into a 1-3 year contract, with a price hike upon renewal.
  • Your site goes on a shared server that’s slow, risky, and not built for high traffic.
  • Support is hard to reach and slow to respond when you need critical help.
  • Cloudways: Bluehost vs Cloudways


Location, location, location

Annoying repetitive but popular sound byte, that. Location matters. But what does location matter online, it’s all a click away, right?

Wrong!

If you are a UK-based business and set your host to London (assuming you have that option) and all your customers are in the home counties, that’s great. But if your business is selling beer to Australia, after the speed drops from all the hops from London to Perth, Australians won’t give a Fosters for your slow-loading website.

I have a number of links (and opinions) on my design statement, one is a test for TTFB: SpeedVitals TTFB. It’s just one of a score of measures, but one were slow servers and server location drag your page speed down, costing you visitors. Knocking even 0.1s off your page load times can increase the number of people stopping on your site.

Google considers anything under 200ms for TTFB ‘good’, more than that is less desirable. The page gives some great advice, and the utility runs tests across 35 locations, aggregating Europe, Asia and America, then aggregates the three, giving a global average. If your global average is over 999ms, that is not good. It’s awful!

That said, if it’s 200ms in your country/town and you are a local butcher, who cares how long it takes someone in Perth to load your page, they ain’t buying a pound of lamb chops from you!

Most of my traffic actually come from North America but I set the server to London instead of New York anyway ‘cos it’s better for locals. Plus, CDN.

Again, SpeedVitals covers this well, explaining CDN vs TTFB and how it all matter. Cloudflare is the best known, but others are considered as good, such as Akamai. You have to do your own research there, but the more global your traffic targets, the more it matters, and that – amongst other factors – comes down to the number and location of their datacentres. Ease to use is a factor too, but casual bloggers are unlikely to stride into this territory.

Features matter tool, notably security, but one things to look for is the number of pops. Figures very with reviews and over time, but as a vague guide, KeyCDN has something like 35 to 40+, Cloudflare is put at 180 to 200+, spread across the world, covering Europe, North and South America, South America, Europe, Asia and the Pacific, and Australia and New Zealand.

So, when you pick a host that supports (say) Cloudflare and out of the blue said host says, “Nah, we’ve changed our minds, you have to use ours instead” it’s time to take a really hard look at your host!


I’m late, I’m late…


Disney White Rabbit
{ Every millisecond counts! }

I was trying to explain to my son why this matters in web design, but also why it is so absurd that it matters (but it does!) and this is the best I could offer:

Take a typical post on Facebook. Let’s say a political post. Or one on climate change. Something to get the blood boiling. And it’s popular, with hundreds of comments. And what is one absolute guarantee? Every. Single. Time.

One commenter will say, “What are you talking about? Did you even read the the article?”
And the person will reply, “Nah, I don’t have time for that…”

… and spend the next hour arguing with a stranger over their opinion on something they don’t have time to read. Then, they will go off and binge on Netflix for the next 8 hours.

Now, let’s look at what Google and co have to say on the matter:

Improving your load time by 0.1s can boost conversion rates by 8%.
Google/Deloitte
Milliseconds Make Millions

People are 40% more likely to spend more than planned when they identify the shopping experience to be highly personalized.
Think with Google
Google/BCG, U.S., Business Impact of Personalization in Retail study, 2019.

77% of smartphone shoppers are more likely to purchase from companies whose mobile sites or apps allow them to make purchases quickly.
Think with Google
Google/Ipsos, U.S., Playbook Omnibus 2019.

You have to ask yourself, “What really annoys me?” And then put yourself in your customers or visitors shoes.


Well colour me pink!

I could talk about say colour and accessibility, note how your graphic designer has a perfect eye for colour, has memorised 30 Pantone shades of ‘purple’ and can recite all the HTML colour names. What a vivid, shiny site you have!

However, I’m a bloke, we tend not to care about things like the difference between Dark Orchid and Dark Violet; they are both just purple or some purple shade to us. Fact is something like 1 in 14 men are colour blind to some degree, though the figure is much lower for women. W3C and many others are particularly interested in this though, and so should you be.

1 in 14 men. In the UK, that’s about 2.5 million men. Will they take the time to struggle with your soft but unreadable pastels, or will they leave immediately?

I’ll generally put up with a slow loading page, if I want an answer, but after a few seconds of waiting, nah, I’m done. We live in a world conditioning us to instant gratification. We might binge for 8 hours on Netflix or Disney, but we want it to load NOW. Before the Internet, before we had so many TV channels clawing for our attention and money, for advertising revenue, people were more inclined to just sit and read, or lie in the grass staring at the clouds, seeing what shapes they made.

Now? Now it’s a meme: Old man yells at clouds

Talking of angry men yelling at clouds, this was a factor in helping me decide to move again: SiteGround vs. WP Engine 2022: Why You Shouldn’t Use Them (Both Have Declined). I didn’t have the same poor experience with WP Engine as some of these, but despite it’s size, Ackadia is not a heavy load. 300, 700 visitors a day is not taxing. If it was 7,000, 70,000 a day, I’d probably yell more at clouds 😉

I’m currently doing an OU module in cloud computing, and we have to use NetBean 12.2 for the IDE. It’s bloody awful, it doesn’t support modern monitors so the text is tiny, like 6 point tiny, but the university can’t get their examples to work with modern software (and they are teaching us!) so we have to put up with it because it’s easier for them. We can’t just walk away, but if it was a website, one that wouldn’t allow you to zoom in without breaking something else, well, bye.

The point of a website isn’t what’s best for you, it’s what’s best for the visitors you want to attract.

If you choose a slow host, a poor theme, the wrong design, who is losing out more? The people that walk away rather that put up with it, or you?

Sometimes though, the colour does matter. I was redeveloping a site for Pool of Life and their pink ladies (it’s a cancer support thing) and the shade of pink had to be right. I had such fun with that. smiley Huh?. I used to be a lab tech with a paint supplier and could grade 30 shades of silver (dull, shiny, smooth, gritty, depended on the aluminium content) but pink’s pink, right?

Pink ? “No, that’s more of a salmon colour.”
Light pink then ? *hard stare*
Deep ? “No!”
*Stares at the photo I have to work from*
Hot ? “Absolutely not!”
Fuchsia, right ? “Still no. That’s too bright.”
But… “No.”

Did you know there are over 100 shades of pink, complete with (mostly unofficial) names and settings for HEX, RGB, and CMYK. One of them is called Jazzberry pink! Half of them just look like shades of brown to me, and I’m not colour blind.

Eventually settled on #fe68c3

Just keep telling yourself, “the customer is always right”. Especially when it’s your wife smiley whistling

Then there’s the other side (you feel me here devs, right):

“We heard page speed really matter and we want to be top in Google, can you help?”

Sure, what’s the problem?

“Well, it takes so long to load it won’t even load for some of our visitors and staff.”

Ewww, OK, I’ll put something together.

{ Later that week, with 12s shaved of the load time }

How’s this?

“Hey? Where’s our Twitter and Facebook feeds? Where’s the sliding images thingy? We want pizzazz!”

I took all that crap out.

“What, no, we want our whistles and bells. And the chairperson wants trumpets as well. Just leave it as it was, but make it load in 2 seconds instead of 20, that’s all we want.”

{ Mouth opens and closes }

OK, Good luck, have fun with that, bye.


Problems left to fix

So far, so good, as far as Ackadia goes. ~Ish!

I still have a few issues to fix (or put up with). The first is lazy-loading screwing up LCP for above the fold images. It comes down to how the server won’t know what device and setting everyone has, so it just defers the lot. However, you don’t want the first image seen to be last in the queue, so tests like Googles LightSpeed test are like “Ewwwww” Smiley gives it the thumbs down.

I suppose there’s hacks and ways around this. Certainly it would explain the growth of sites with massive, screen-filling headers, something themes like Astra can do well. If the above the fold header is all ‘white space’ with a small logo, it evades the issue. There is supposed to be a fix coming in WordPress, according to Web.dev, and other improvements to speed up load and server times, according to WordPress. (WordPress 6.1 is currently at Release Candidate 2).

97 percent for mobile pagespeed insight
{ Ackadia mobile PageSpeed Insights, October 2022 }

Increasing the size of the header and padding out the empty white space with the page title and and tag in text pushed the score up to 98%. The LCP (Largest Contentful Paint) was reduced from 2.8s to just 1.9s. Remember, even 0.1s can make a different! (For desktop, with a score of 100%, the LCP is just 0.5s, with only the DOM size flagged orange).

So, at the moment, Ackadia appears OK. Not perfect for mobile, yet, but better than most. Still improvements to be made, but getting there!


Big bad DOM

The other issue is a wretchedly high DOM count. This is the sum total of all the nested html tags, links, and most of that ugly count is down to WordPress (or other CMS) and the theme. Also, the popular block editors are notorious for bloated code. Some are better than others, I suppose, but the garbage I had to rip out of a site that had used Divi… Smiley heaving!

I had nearly 1,600 on my DOM count and it wasn’t all my code, I write tidy code with proper markup, thank you. Sure, I can break my longer articles into smaller pages, but really that makes more bloat and more work in the load run.

No, it was garbage like this:

pagespeed-insights-avoid-excessive-dom-size
{ PageSpeed Insights: Avoid excessive DOM size }

It was around 1,580, to drop it down by 300 all I had to do was remove the menu! Bit drastic, but needs tidying up anyway. The fact that JUST the menu accounts for 20% to 25% of the total DOM size is ridiculous. But anything over 1,500 nodes gets red-flagged, so I ripped it out for now.

On the plus side, my score is already better Google’s (which is currently 95% on mobile) and – unlike Google – I am not flagged for “Image elements do not have explicit width and height” (That is just lazy, Google. Shame on you!)

With a certain amount of glee, I also note that Google flags themselves for “Reduce unused JavaScript” – and it’s unused Google scripts that Google are calling.

I improved my mobile page loads by 1,500ms just by removing Google analytics and Adsense code! That’s 1.5s shaved for a mobile test, removing Google’s bloated calls. They were sloppy in the early days too, but since they removed their "Don’t be evil" slogan… There’s a lot to be said (not good) about a company whose corporate philosophy runs akin to “If you don’t want us to spy on your illicit activities, don’t do them; we are watching EVERYONE.

Links

See also: Google (2017): The need for speed: Evaluating the perception and reality of speed on the mobile web

Also worth a read: Portent (2022): Site Speed is (Still) Impacting Your Conversion Rate

Interestingly, if only ‘cos it surprised me when I read it, I’ve been using WordPress for nearly 20 years now! Apparently, I started using it in November 2003, just after it was released! Ackadia: 20 years of web design

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll to Top
%d bloggers like this: