WordPress vs WordPress

There’s been a lot of drama in the WordPress community this week, as people are questioning the relationship between wordpress.org and wordpress.com in Google’s search results.

Its a long-standing issue that people don’t really understand the difference between the two sites and their propositions, but one that we’ve broadly learned to tolerate as immutable. Now the spotlight is being shone on some potentially harmful implications of that relationship. 

The story so far

WordPress.com (which sells hosted solutions for WordPress) recently replicated the plugin directory from wordpress.org (which facilitates self-hosted WordPress). That has some interesting SEO implications, as dot com may outrank dot org when people search for plugins. That may also happen for more general searches for the kinds of phrases that might return plugins; eg, “wordpress security” or similar. It may also affect the indexing and ranking of both sites in unpredictable ways, as they ‘compete’ with each other.

When people have raised questions to Matt Mullenweg (the co-founder of WordPress, and CEO of Automattic – the company that runs wordpress.com) about this, his response has been that it’s not a problem – that more real estate drives more discovery, more downloads, more distribution, and more usage; all of which are good things. He also argues that because WordPress plugins are released under GPL licensing, it’s perfectly fine for anyone to create their own marketplace for them (and that this is a good and harmless thing).

That might be generally true, except that the existence of this particular marketplace comes with significant commercial implications – and may also represent an existential risk to WordPress.

Confusing users?

WordPress.org lets you freely download and install plugins (on your self-hosted site), whereas wordpress.com only lets you use plugins (on your dot com hosted site) if you pay for their ‘business’ hosting tier. There’s potential user confusion and misaligned incentives here; wordpress.com makes money from winning in these kinds of searches in a way that might negatively affect the perception, positioning, adoption, and preference of WordPress as a solution.

Most people don’t understand the difference between wordpress.org and wordpress.com, and it’s easy to imagine how folks might end up feeling like they’ve ‘been tricked’ into spending money on a solution when they might not otherwise have had to (or perhaps worse, that this happens without their awareness and they perceive it as the norm). It’s also easy to imagine that folks might seek non-wordpress solutions if/when they discover that they ‘have to’ (or rather, didn’t have to) pay for plugins, or feel like they’ve been conned. 

The SEO consideration

A lot of the arguments around whether this is good or bad hinge on the user experience for logged-in, solution-aware searchers on wordpress.com. Matt’s stance is that most users hitting wordpress.com plugin pages are logged in and understand what they’re looking at.

But that’s not the case for Google. 

Google sees a logged-out version of a plugin page on a strong(er) domain, with no signals or strategy to help them to understand the desired outcome (i.e, which site should ‘win’).

Google judges sites based on what it sees, and uses that to determine if and when it presents those pages to its audience of searchers. The experience of logged-in users doesn’t matter (and isn’t accessible) to search engines. The dot com repository is a (more or less) direct, competing plugin directory.

That means that without that active management, we create a scenario where dot com ‘accidentally’ ends up winning more often than not. If that’s a deliberate strategy, it’s smart (although ethically questionable). If it’s an accidental oversight, it’s problematic. WordPress.org’s SEO leaves a lot to be desired; despite its size and ‘clout’, its technical setup, content, strategy, performance, accessibility, resourcing and other factors are woefully below par. There’ll be no shortage of cases where it loses to its cleaner, sleeker sibling.

So, what should we do? 

WordPress.com plugin pages could use canonical URLs to reference their wordpress.org equivalents. This would send a clear signal that the duplication is intentional and that the wordpress.org pages should win out. But if we consider Google’s recent updates to how they prefer “syndicated” content to be handled, then this isn’t a viable solution. Not to mention that canonical URLs are a ‘hint’ rather than a directive and might be ignored when other signals (such as the poor state of wordpress.org pages) hint that the user might benefit from a different interpretation. There’s no guarantee that canonical URLs would reliably solve the problem.

Instead, the best option might be for dot com pages to be noindex’d. If the target audience for these pages is indeed mostly logged-in users, then the small opportunity costs associated with preventing them from being indexed by search engines should be inconsequential.

That’s a quick, trivially simple fix to implement, and should be non-contentious – unless the potential revenue from misrepresenting what WordPress is or isn’t to unknowing searches is more important than the integrity and health of the project.

Ultimately, we need to understand that how Google perceives and interacts with these two distinct websites and brands has the potential to radically alter the balance of power and money in the ecosystem. If we’re not actively managing that (or if we’re intentionally distorting it), we’ll end up harming ourselves and our users.

guest

5 Comments
Inline Feedbacks
View all comments
Javi

It’s far from being accidental. A quick glance at Calypso repo gets you a bunch of PRs about improving the SEO of the .com plugin pages:

https://github.com/Automattic/wp-calypso/pull/76496

Matt

Thank you very much for this feedback, will definitely give it some thought.

Slava Abakumov

It’s sad that one year has passed, and on wordpress.com/plugins/[any-plugin]/:
1. robots meta tag is still missing and not set to “noindex”
2. X‑Robots-Tag is still missing and not set to “noindex”
3. canonical is missing and not pointing to .org

But at least schema is saying “sameAs => https://wordpress.org/plugins/%5Bany-plugin%5D/”. But I’m not an expert to understand whether that’s enough. Maybe Jono can revisit the situation? 🙂

What I see is that for some plugins google returns wordpress.com/plugins/* result higher than wordpress.org/plugins/*, and for other plugins – positions are reversed.

Apparently, confusing google is fine as long as CTA on .com leads to where it needs to lead.

Web Generation

Thanks for the post. This seems indicative of a substantial conflict of interest that is unlikely to be good for the community in the long run. Hopefully if enough of the community start to voice concerns about this, they will listen to reason. 

It’s a real shame that we should have to waste our energy on this issue.