WordPress's Matt in debate with Netlify's Matt
theregister.com
theregister.com
Wordpress is not going away. As long as you don't have to code to use Wordpress, it will be easier and far more accessible for people that don't have a lot of money for a website to hire a web designer (lower skill, lower pay than developer).
Gatsby or another Jamstack client could build themes that don't require coding, similar to Wordpress, but I don't see it happening. Webflow is sort of doing that now by allowing designers to build websites without coding.
Ultimately, the debate is silly because Wordpress and Jamstack have two different use cases. Wordpress will probably become less popular for big companies, but it will still be very popular for your average small to mid-sized business.
I haven't regularly done client WP work in several years but I'm planning on doing some over the next year and I'm curious - how good is Gutenberg now? Before it would be adding fields to the admin with a plugin like Advanced Custom Fields - were they able to make it truly WYSIWYG for editing content?
WordPress core team is still working on the "full site editing" project (that is, being able to edit headers, footers and sidebars in Gutenberg). It's coming probably later this year or beginning of 2021. But you can still edit a lot of things with Gutenberg as-is.
I want to get our team using it for simpler projects, and I haven't really hit any blockers with Gridsome, whereas I've hit plenty with Wordpress when we need a specific url structure.
But even for me, who doesn't need a backend GUI stuff like a contact form was a pain. Luckily I host with Netlify and they have an easy form solution but imagine that across many other situations even for a somewhat simple site. Now you're having to set up accounts with all these other services when before it was all contained in one place. All for what? The benefit doesn't seem substantial enough to me to give all of that up.
Even if you use WordPress as the backend, you've still got to deal with integrating it's API which is more work than something generating all the markup you need etc...
Honestly...what other stack could get you that far in 3-4 HOURS?
For clients though? If they want WordPress (or give a feature list that sounds up WordPress' alley without needing a bespoke Laravel build) they get it. We'd lose all our customers if I tried to get them editing Hugo markdown files!
Netlify could do with watching their PR at this point I feel. The honeymoon period is over. I've personally switched my site to a fiver DigitalOcean box and Cloudflare because I already had the DO box for another project. The deploy is infinitely faster and I don't need to worry about hitting limits bumping me up to a 50 quid a month account
As a former user who had one of those typical personal Wordpress blogs on a shared hosting setup, I experienced the nightmare of dealing with spam, hacking, having to constantly upgrade, etc. So for my own use I would never consider anything other than a static site generator; but I can see the appeal for many use cases, especially if the backend is managed (services like WP Engine) or semi-managed (the guy who set it up is getting a small monthly fee to keep an eye on things).
My sense is that WP Engine targets busy professionals who have first-hand experience of how much hassle a WP installation can be to maintain, people for whom downtime or being hacked has potentially a severe cost in terms of lost business, reputational damage, etc. The other way to look at it is that $300/mo sounds like a lot for "just a WP installation", but it's peanuts compared to hiring another person for your IT department.
You're also paying for decent support tech's and support availability. I think to this day you can call and a human will pick up the phone. Tech's are well versed in the WordPress ecosystem and many are developers who can help you troubleshoot things at the code level, whereas with some budget webhost, they'll tell you to pound sand and hang up (albeit more politely).
That said, I think these "premium" WordPress hosts are purely for non-devs or devs with no time to provide their own support. If you're a developer, there is no reason to pay that premium to keep your blog online. If you want a more managed solution, use something like ServerPilot or SpinUp WP, or any number of other service that will provide a thin maintenance layer on top of any garden variety VPS provider.
The API will have an official plugin, or someone will have written one because their client needed it and the dev decided to support and sell it. Complicated forms can be turned out with plugins like Gravity forms.
This is why lots and lots of sites are still built on wordpress because the ecosystem around it lets people do a lot of stuff quickly without touching a line of code.
Very simple and more then did the job. It costed basically nothing to use and pretty much anyone could maintain the site.
I have semi-recently converted a couple of Wordpress and also Docuwiki sites to static sites. First with Hugo, then converted again from Hugo -> Jekyll.
I could not be happier with the results (...with Jekyll anyway, less so with Hugo). And there is no npm.
The biggest problem for me was mentioned in the article - security. It felt like I was frequently having to babysit updates in Dokuwiki and Wordpress. Both of the update procedures felt precarious and dangerous - more than once they'd fail to update and hose something on the server and I'd need to go in with a SFTP client or SSH and try to rectify whatever the hell happened. I grew to fear the updates and started to procrastinate doing them which was a warning sign. Perhaps they are perfect and flawless these days?
And even when I had updated there was always this lingering feeling that there was some unknown/unpatched exploit or I was 12 hours too late patching and so someone was "inside" my wordpress running a spam/phishing operation without me knowing (happened to me, but perhaps 15 years ago now and was due to the hosting provider so not really relevant... I only found out about that one when a victim of the phishing operation did a whois on the domain and sent me threatening messages. Nice.)
These sites that were on Wordpress/dokuwiki didn't even need to be "dynamic" anyway - they just needed to serve content and provide a simple way for people to edit the content. Github pages does a great job for this - non-tech users just use the in-browser editor and if they accidentally delete something there is the git changelog. The time "lost" to do a one-time compilation of the site is totally paid back by the near instantaneous loading time compared to the bloated feel of the previous sites waiting for PHP+MySQL to do its thing on some VPS or shared host (I did not have a CDN).
That way you get Wordpress content authoring and all the advantages of a static site you mention
it's brittle, and the updates aren't semvar-- a minor verions upgrade could ereally hose you.
But with WP, the WP 5.5 thing where they killed the automatic jQuery migrate inclusion was about the only time in the last year or so when I was caused serious problems by an update (for context, I manage around 500 WP installs for gov/large businesses and a univeristy).
There was a pretty bad security issue with a popular file manager plugin, but we council folks to stay away from those solutions anyhow.
I feel like in general the ecosystem has gotten a lot better about the update system.
E.g. just getting a simple child page listing at "domain.com/something/" to automatically list all child pages took hours and hours and I still couldn't get it to work. It literally had a blank page.
It felt over-engineered and they'd crow about speed improvements or some niche new whizz-bang feature or whatever for their new releases when I was there still just trying (and failing) to get what I feel is a totally fundamental basic feature working.
I also found creating a custom theme from scratch in Hugo quite complicated with archetypes or whatever they were called.
It felt like the project was more interested in the engineering challenge of the site generator rather than making something that is usable (which is totally 100% fine for them to choose that path) - after all directory listings seems like a basic feature to me yet it was apparently impossible without considerable fiddling around with _index files or other custom nonsense. Everything just felt a little bit more complex and laborious than it should have been to get a simple 4 or 5 layer site up and running exactly the way I wanted it.
I am sure that if I kept at it and kept trawling the support forums and posting questions and hanging out on IRC someone could have helped me ... but life is too short for that and I've got things to do so it just wasn't worth fighting with.
Jekyll on the other hand is super-simple and I was able to port entire sites - complete with a from-scratch custom theme + CSS - in under a day, and have them look and act exactly how I wanted.
I only picked Hugo over Jekyll originally as at the time I was a fan of golang, so I was not really that wedded to Hugo for another reason.
Netlify is a company working hard to get lots of people locked into their companies ecosystem. This is a very bad thing for the open web.
The principles behind JAM are platform-agnostic though.
As I see it, Netlify is just dressing serverless in a new hat.
What are you implying?
That has a specific meaning whereby they prevent customers from leaving, somehow, but forcing everyone onto their platform.
As far as I can tell, the basis for this claim is that somehow the "jamstack" label is so valuable that people will be "forced" to work with Netlify so they can enjoy the privilege of using that trademarked term, but you'll forgive me if I find that to be a pretty weak basis upon which to levy such a strong claim.
Ultimately, again, this is just an architectural pattern and a bunch of open standards. Netlify is absolutely guilty of hyping the holy heck out of this thing as though it's the silver bullet that'll solve all of the world's problems. And sure, maybe they're trying to build a brand around it.
But I fail to see how any of this is so nefarious as to qualify as attempted "lock-in".
Even their serverless functions api and DB can be connected to any serverless provider.
But I admit that the hype will eventually die up and someone else would be able to trademark their model of simple stupid website publishing, hey I guess I should do just that. SSWP here we go!
But Biilmann is also right. The Javascript world (React, Vue, the npm ecosystem, ES6/7+), modularized services like Stripe, Algolia, etc, plus eg Git, have passed the Wordpress development ecosystem by.
Wordpress needs a bridge to modern development, so people can get the best of both worlds. That's how we see Gatsby, so we built a source plugin to import content, built Gatsby Cloud to do live content previews, created incremental builds to match the instant content go-live experience.
There's a great quote from Jason Bahl, the maintainer of WPGraphQL, on this: "I don’t see Jamstack competing against WordPress. This isn’t a zero sum game. If the Gatsby + WordPress experience can allow users to use the best CMS in the world while using modern dev tooling, it’s a win all around.”
https://wptavern.com/matt-mullenweg-and-jamstack-community-s...
In particular, it will never find an audience among non-technical users which is a very different audience from technical users who often think only of their own likes and needs.
If you don't like WordPress, there are more modern CMS apps with modern coding practices (often written in PHP). They aren't part of the Jamstack and but are often much simpler to deploy.
Netlify CMS has been great in testing, but I haven't deployed anything to a client with it. I know that, as a dev, I'll never get stuck in a situation where I can't accomplish something with Jamstack (Gridsome is the one I prefer). I just fear the day a client needs to use a CMS and they can't edit/demo something the way the expect.
The vast majority of users that WP caters to are better served by Facebook Pages, Wix, Squarespace and Webflow.
Large and high traffic sites aren’t a great fit for WP either as the codebase was never meant for scalability and still maintains a painful developer expierence.
There are plenty of options, “JAMStack” included, that are still a much better fit.
Most sites I work on these days are either simple enough to be solved with no-code solutions, or complex enough that Wordpress is going to hold you back.
Granted, in most of the world, that's still a huge swath of businesses, but most of them realistically don't need a heavily customized online presence. Professional services are a great example of potentially higher revenue businesses that don't need a fully custom web presence. In competitive markets, I've seen plumbing firms/electricians/pest control all compete for SEO, thus spending more on web dev services, but they're the minority.
I work for a company that specializes in E-Commerce, with a narrower focus on high performance WooCommerce and custom Shopify development. We've seen some enterprise level customers start on WP and Woo for an MVP and then spend half a decade working to get off of it to their own in-house stack, or to transition to a more managed platform like Shopify. Each one has it's ups and downs, but by no means is any one a silver bullet. It largely depends on what your business requirements are and the expectations of your customers.
Selling a dozen individual SKUs with limited customization? Shopify is fantastic and has an extremely optimized checkout flow.
Need more customizations to the checkout or you want to create extremely customized user experiences for interactions like Add-to-cart? WP might be a better fit since you can control the entire stack. But WP doesn't scale as easily as Shopify (though it can scale, that's our job) so you'll pay a premium to develop the platform to handle that scale.
Outside E-commerce, I see a lot of potential for WP in the higher ed space still. Community colleges and trade schools especially as you can work towards a unified branding effort, while still providing customization and flexibility for teachers. I know a few colleges that keep a small staff of less than 5 people, who manage web development and maintenance for decent sized institutions.
All that I guess to say, is that WP still has a place, but that it's shrinking in scope, despite the increase in actual usage numbers. Only time will tell if the trend stays positive for another decade, or if we start to see a slip as other offerings continue to improve and mature.
Hosting it yourself is a mistake. Not because it's inherently insecure (the base install isn't) but because no one can control themselves with plugins.
There's an entire cottage industry revolving around wpscan and securing/popping the millions of poorly secured wordpress instances floating around.
We switched our WP hosting to WP Engine a couple years ago and have been happy with that.
Also, millions of people host WordPress themselves and it works fine considering how much value it delivers.
The moment marketing wants to be able to login to make dynamic updates and load the instance up with plugins it's time to outsource the hosting. My team already has too much going on to spend time on this foolishness.
This got especially worse with 5.x since you can't 403 the REST API off the corporate network anymore (aka the REST API that shipped vulnerable on day 1). It's used for customer facing 'stuff' now.
There are plenty of commercial hosts at all price points that manage this stuff.
Shopify has probably eaten all the WP ecommerce sites.
Blogging has probably moved to Medium, Substack, static sites, etc.
Shopify is great for a simple business model with little hands on requirements and pushes towards low friction checkouts.
Woo is less optimized to scale out of the box (this can be tweaked to scale easily enough), but you have entire stack available to you. It's just PHP and MySQL at the end of the day, so anything is possible. This allows you to build more custom experiences depending on what type of products you're selling.
Think of a local t-shirt store. They might start on Shopify, selling their shirts for the local sports teams and tourist attractions, but as they grow, they move into selling custom branded clothes using a printer in their back office. They might transition to WordPress to build a more customized checkout flow, where a user can upload various design files (in a variety of formats: svg, ai, pdf, etc), tweak them on a model and then start a back and forth conversation with the shop owner if there are problems. All that can happen in their account and customizable. That sort of custom experience just isn't possible on Shopify without jumping through some absolutely ridiculous hoops (hosting a separate API and then running an SPA in your shopify theme, etc).
All that to say, it depends :)
Where are you getting this information? WP is an order of magnitude larger than all of them combined for blogging. Matt Mullenweg's 50% market share goal isn't crazy hyperbole.
I'm working on a platform for content creators (including blogging) so whenever I visit a blog I check the source code to see what they're using. It's not so often that I come up with a WP site to be honest, although I'm obviously biased in my interests.
> Matt Mullenweg's 50% market share goal isn't crazy hyperbole.
I agree, although it seems (again anecdotal impression) WP is a lot more common as a general purpose CMS.
For rest of the world, and other Small Business, They use Wordpress. And has probably never heard of those solution before.
The closest thing is Wix.com
- Ecommerce store via WooCommerce
- Online Course via Learndash LMS
- Lead Generation/Capture via WPForms, or other plugins.
Is it? I honestly thought it was they other way around.
> Medium/Substack is basically graveyard at this point
This is definitely not true. Just look at how many Medium articles are shared to HN daily:
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
We have quite a few NGOs and large enterprises which we do that kind of work for.... it's a very popular and easy to use system for non-technical users who need to manage content.
Oh I remember that. But that was before the era of SSD and Multicore CPU. And Echoing all the other comments, the two system ( at its current form ) aren't even targeting the same market.
Choose your poison.
I did... as an experiment for my blog: https://www.louiechristie.com/blog/
The secret sauce is https://github.com/henrikwirth/gatsby-starter-wordpress-twen...
I'm doing what I can to persuade clients that WP isn't the only solution, and that CMS's like Statamic and frameworks like Laravel can be a viable and cost effective solution.
Ghost, maybe?
I've stopped recommending the self-hosted (wordpress.org) version to non technical people (even if it's a 1 click hosting platform), because there are just too many things that can go wrong, particularly with plugins.
As a developer, my experience with Statamic and Laravel has been joyful, compared to WP.
I did this for a few friends, hosted them on my server, set up Wordpress. Worked for some but most found even Wordpres too complicated. I sent to them to Squarespace and they're happy with that, albeit paying a hefty enough monthly fee after the initial trial period. Not my problem though! washing my hands
I last used it in 1., and picked it up 2 weeks ago when I found myself needing a headless CMS.
I got a sense from the new features since 1. that it's moving towards a membership system which might not be relevant for many websites.
In terms of an editor and a way to expose content, is good enough; but like I say with more developers, this could bring more hands to implement stronger plugin systems and improve the core editing options.
I really wanted to like Gutenberg. I learned how to build blocks (and the f- me, they totally changed a ton of stuff about how that works).
At this point, I have some pretty deep knowledge about WP and how to make it do all kinds of funky stuff it shouldn't do... but holy hell if I could get the sales folks to start selling about any other CMS I'd do that in a heart beat.
But I'd still have 600 or so legacy WP sites to support. I'm grateful to have work, and WP can be a real shit show when you look at that many sites.
Those would give you a modest plugin ecosystem to get stuff like form submissions, but also give developers an escape hatch for when something more custom is required.