Plugin names have too little "weight" in the search algorithm

[littlebizzy]

Alternative summary: Plugin ratings (reviews) etc have too much "weight" in the search algorithm

Dozens if not hundreds of very good plugins have all but died in recent years due to lack of being seen, and thus never really had a fighting chance to "catch on" in the WordPress ecosystem... this arguably results in a sort of monopolizing effect where older plugins (with lots of reviews, etc) blot out any new competition...

In our case, we are able to leverage our 200,000+ active installs to promote our various plugins, and we also promote our plugins around the web. But for many other plugin newbies, it does not seem fair, or accurate.

Here's 2 examples from our own plugins:

1. Disable Gutenberg -- does not show up anywhere in the 18 pages of ​search results for "disable gutenberg" despite being only one of two plugins with this exact match name.

2. Custom Functions -- shows up on page 8 of ​search results despite being the only plugin in the entire directory with this exact match name (there is another plugin called My Custom Functions, however).

While the ratings average is a useful feature (among other items that are used in the current search algorithm), it would seem that plugin names -- and meta descriptions -- do not have enough weight in the calculation. Perhaps the goal is to force plugin authors to earn their reputation, which is fine, but it has the negative effect of preventing otherwise "good" plugins a chance at ever being seen, which is probably increasing the rate of plugin abandonment (not to mention extremely irrelevant/bizarre search results based largely on rating averages).

Keep in mind that commercial "app stores" have multiple methods of highlighting new apps such as Trending, Popular This Month, Recommended For You, and other intelligent cross-promotions. But if WordPress.org is going to maintain its singular search algorithm, plugin names and meta descriptions should be given more weight.

"Well they just need to promote themselves better."

Don't forget that the majority of plugins are not profit-oriented. Ultimately, this issue is punishing the most generous contributors to the WordPress project, who are not focused on self-promotion per se.

"But literally all they need to do is get a few reviews to be seen better."

See above question/answer. Also, do we need more spam in the WP Admin Dashboard? In many cases we've had to wait 1-2 years to get our first reviews of a plugin, not to mention several thousand active installs...

[littlebizzy]

search results for "custom functions"

[tellyworth]

Names are highly rated. disable-gutenberg is number one in the search results for me:

​https://wordpress.org/plugins/search/disable+gutenberg/

It's a relatively new plugin, that may have had something to do with it. AFAIK nothing has changed there, we didn't bump it up in response to your ticket.

custom-functions-littlebizzy is on the first page, 4th from the top. Again it's a new plugin, it might simply have taken some time to bubble to the top.

Ratings are given relatively little weight in the search algorithm, and reviews don't make a difference. Text (names and descriptions) are the most important factors. Things like the active install count, tested-up-to, recency, and support thread responses are more important.

Have the results improved for you? You are searching on the English site, right?

[littlebizzy]

We could provide dozens more examples but names and meta descriptions are definitely not highly weighted currently, if anything it seems the overall readme.txt contents count higher than both of them. Previously we overloaded our readme.txt with tons of "keywords" not realizing it was disallowed, and we ranked extremely well for dozens of phrases even when they were not related to our plugin names or meta descriptions.

The common denominator = we had a high ratings score.

Ratings are given relatively little weight in the search algorithm, and reviews don't make a difference. Text (names and descriptions) are the most important factors. Things like the active install count, tested-up-to, recency, and support thread responses are more important.

To be clear, by meta description we mean the 150 character description, not the entire readme contents, and by "reviews" we simply mean the star ratings average (when it's high, you can rank for any phrase).

Not sure where you are seeing our plugins:

​https://archive.fo/Gupcs

​https://archive.fo/JCoex

Our plugins definitely do not show up in either of these searches. These 2 examples are evidence enough that names and meta descriptions don't have nearly enough weight... unless we've been targeted.

Here's another example, Go Slider, which does not appear:

​https://archive.fo/5uvZ0

And another one (3+ months old) called WP Advance Comment Form Validation does not show up either:

​https://archive.fo/A45pa

This is not only killing new projects before they even begin, but creating a class of "plugin elitists" who control the entire plugin directory by ranking for any phrase they want, while silencing all newcomers.

[SergeyBiryukov]

Related: #2642, #3327.

[littlebizzy]

Related: #2642, #3327.

Thanks @SergeyBiryukov had not found those tickets on first research (both have interesting ideas).

One thing that none of these ideas take into account is that "changing things up" sometimes is one of the best solutions to users constantly seeing the same search results all the time.

Not sure if you want to merge (or close) this ticket, but seems these are easiest options:

• rotate algorithm a few times each year (e.g. Google)
• reduce the weight of readme (this would also reduce stuffing)
• increase the weight of plugin names and 150-word meta descriptions (e.g. force conciseness)

Or more extensively:

• introduce more serp "tabs" that might sort varied results (e.g. video tubes)
  • newest (oldest)
  • highest rated
  • most active installs (month/year/all time)
  • most downloads (month/year/all time)
  • trending
• introduce a "similar plugins" feature at bottom/sidebar of plugin profile pages (like mobile app stores)
• introduce plugin "categories" (like mobile app stores)
• rotate "featured" plugins e.g. Editor's Choice

This would blow up the plugin directory, and tempt back a lot more developers.

Keep in mind the next 10+ years, WordPress is competing with slick app stores like Shopify.

More technical ideas simply will not work:

• stemming word play
• requiring "quotes" or slug-names (99% of users are clueless)
• telling authors to get more reviews while simultaneously telling them not to spam the WP Admin
• telling authors to not push updates too often while simultaneously ranking plugins based on how often they push updates and how updated their Tested Up To headers are, etc

[Otto42]

We're happy with the weights at this time.