Hacker News Reader

Comment by unixhero | Hacker News Reader

I attempted to work with WordPress and its plugins in a commercial scenario. Whenever we hit a tiny corner case with a theme, something we believed should have functioned well - The whole project failed.

So yeah I am not taking on any Wordpress projects ever again. I just tell people to use Wix and or Shopify and don't try to run an ecommerce site, it is too difficult for them.

View on HN

> Whenever we hit a tiny corner case

These are the nightmare of WordPress sites. Someone builds a house of cards out of a pile of plugins, and then I'm asked "Why might it cost $$$ to upgrade my WordPress plugins?!?". Well sire, because plugins can throw in breaking changes, and an upgrade may take 2 minutes an work perfectly, or it might introduce a corner case that breaks something. And you don't have any automated tests to detect this breakage, so someone has to check by hand. Or forget a corner-case, it may completely break part of your site because a (commercial) plugin developer decided to rewrite and get rid of a feature you depended on [true story].

I try not to touch anything beyond a WordPress landing/marketing site, but every so often one comes bundled with another project or I give in to a sob story...

This is arguably what's so dangerous about a plugin ecosystem in general. On the one hand, you can quickly implement a bunch of useful new functionality at the drop of a hat, and that's obviously incredibly useful for non technical folks.

But at the same time, you then often end up dependency hell, and realise that any changes going forward will break something or another that your site depends on, leading to many, many sleepless nights. God help you if you're one of those folks who installed hundreds of plugins, because an upgrade often meant 'revert everything to default, then hope at least some of your plugins had equivalents for the next version'.

It's why I'd recommend anyone who can do so to only use their own code for a WordPress (or other CMS based) site, and keep the third party stuff to popular plugins you know will get regular updates.

Between Squarespace and Shopify it's really hard to argue in favour of Wordpress these days, yeah. In theory WP can do everything, but it needs a full-time PHP dev (and these days, also a full-time react dev for Gutenberg) to baby sit it and keep up with vulnerabilities, incompatibilities and all the other stupidity of the Wordpress ecosystem.

You just limit yourself. I have built database-style wordpress sites with custom taxonomies/types using just my own glue code, theme templates and Advanced Custom Fields (and ACF is just there for the backend experience mostly). It's also obviously not the best choice for everything, but some of these sites I built 10 years ago and they're still going without major issues.

Take a lesson from the big boys, here: WordPress for landing/marketing pages only.

Why not a SSG, like Jekyll, for that? Isn't WP overkill?

You probably need a headless CMS of some kind so the client can manage it. I've used decap/Netlify CMS, and it's fine, but it's clearly not anyone's priority anymore.

Unless you have a good solution in place that abstracts away the CMS, and triggers automatic (near-instant) redeploys, that's usually not a good idea.

Your marketing team(/person) will have to do many dozen changes a day on your landing pages, and you don't want them having to wait for changes to appear on the website.

A base WP setup that is managed via version-control + WPs normal live CMS is still one of the best CMSs out there IMHO.

> Your marketing team(/person) will have to do many dozen changes a day on your landing pages

I would really like to see examples of this.

It seems that either the marketing team is hugely incompetent and are making rudimentary typo-style edits, or the page is far from landing.

Maybe "landing page" wasn't the correct term here.

> I would really like to see examples of this.

I almost struggle to come up with a non-example for this for any profitable tech company.

Look at the sitemap of e.g. Solarwinds[0] or Hashicorp[1]. Both have thousands of sub-pages with different marketing copy as content to capture as much organic traffic as possible and to convert paid traffic as good as possible. That marketing copy doesn't write and update itself (or at least it hasn't until recently).

If you look on most job boards you will also find many positions that contain "SEO" in their title. While some of them are more rosy "SEO management" positions, if you squint a little bit, you'll see that in practice those people spend a good chunk of their day in their CMS (likely WP) to create new pages and updating existing pages based on conversion tracking data. That also usually doesn't have a lot to do with "incompetence", as that process is rarely something where you can one-shot the right solution. More realistically you'll just have to try out a lot and see what sticks (both with search engines and people).

[0]: https://web.archive.org/web/sitemap/https://www.solarwinds.c...

[1]: https://web.archive.org/web/sitemap/https://www.hashicorp.co...

Reply on news.ycombinator.com