[wp-meta] [Making WordPress.org] #8166: Plugin Search Algorithm Creates Barriers for New Plugin Discoverability

[Making WordPress.org]

#8166: Plugin Search Algorithm Creates Barriers for New Plugin Discoverability
--------------------------+------------------------------------------------
 Reporter:  webstepper    |      Owner:  (none)
     Type:  enhancement   |     Status:  new
 Priority:  normal        |  Milestone:
Component:  Plugin        |   Keywords:  search discoverability new-plugins
  Directory               |
--------------------------+------------------------------------------------
 As a plugin author, I've been looking into how the plugin search works,
 and I think there's room for improvement in helping users discover new
 plugins.

 == The Problem ==

 The current search algorithm makes it really hard for new plugins to get
 found unless they copy the naming conventions of already-dominant plugins.
 This leads to a situation where searching for something like "discount
 rules" returns dozens of plugins all named variations of "Discount Rules
 for WooCommerce" - because that's what the algorithm rewards.

 I dug into the search code (`class-plugin-search.php`) and found some
 things worth discussing.

 == What I Found ==

 === Title matching gets a 5x boost ===

 {{{
 $should_match[] = [
     'multi_match' => [
         'query'  => $search_phrase,
         'fields' => $this->localise_es_fields( 'title', 'slug_text' ),
         'type'   => 'most_fields',
         'boost'  => 5,
     ],
 ];
 }}}

 This is the biggest boost in the algorithm. It pushes authors to name
 their plugins exactly what users search for, which usually means copying
 the top plugin's name.

 === Active installs penalty is steep ===

 {{{
 'exp' => [
     'active_installs' => [
         'origin' => 1000000,
         'offset' => 0,
         'scale'  => 900000,
         'decay'  => 0.75,
     ],
 ],
 }}}

 The algorithm uses 1 million installs as the baseline and penalizes
 everything below that exponentially. A plugin with 1,000 installs
 basically scores nothing compared to one with a million. Popular plugins
 should rank higher, sure, but this seems excessive for new plugins trying
 to get started.

 === No support history means penalty ===

 {{{
 'field_value_factor' => [
     'field'    => 'support_threads_resolved',
     'missing'  => 0.5,
 ],
 }}}

 New plugins don't have support threads yet, so they get penalized by
 default. The same goes for ratings.

 == Why This Matters ==

 * '''Plugin names all look the same''' - Search for any popular feature,
 and you'll see dozens of nearly identical names. The authors learned this
 is the only way to rank.

 * '''Blog recommendations become confusing''' - "I recommend the Discount
 Rules plugin" - which one? There are 50 with that name.

 * '''Innovation gets buried''' - Build something new with a unique name?
 Nobody finds it. You're forced to copy what's already popular.

 * '''Established plugins keep winning''' - They rank first, get more
 installs, and rank even higher. New plugins can't break in.

 == Suggestions ==

 I don't have all the answers, but here are some ideas:

 === 1. A "Discover" or "New & Notable" section ===

 Modern app stores and social platforms have discovery features to surface
 new content. A dedicated section for quality new plugins (recently
 reviewed, good code, responsive authors, growth rate) would help new
 developers get seen.

 === 2. Search result diversity ===

 Include 1-2 "rising" plugins on each search results page, clearly labeled
 as new/rising.

 === 3. Reduce penalties for new plugins ===

 * Cap the minimum score for low-install plugins
 * Don't penalize plugins under 6 months old for having no support history
 * Lower the title boost so unique names have a chance

 === 4. Add sorting options ===

 Let users sort by "Newest" or "Rising" in addition to relevance.

 == Why This Matters for WordPress ==

 WordPress has always been about democratizing publishing. But the plugin
 directory is consolidating around whoever got there first with a
 particular naming pattern.

 If new developers see that success means copying what exists, they'll
 either give up or make clones. WordPress loses the innovation that made it
 great.

 Platforms that don't help new creators get discovered eventually lose
 them. I'd hate to see that happen here.

 Happy to discuss further or provide more data if helpful.

-- 
Ticket URL: <https://meta.trac.wordpress.org/ticket/8166>
Making WordPress.org <https://meta.trac.wordpress.org/>
Making WordPress.org