WordPress Plugin Directory Redesign: Why So Many People Feel Their Feedback Was Ignored

Earlier this week, the WordPress plugin directory relaunched with a new design and an improved search algorithm. Feedback has been rolling in from Tavern readers and members of the Advanced WordPress Facebook group.

The improvements to search has received a number of positive comments in addition to people requesting the ability to sort results. The responsive design and aesthetics of the page have also received praise.

There are a number of items that plugin developers and users have reported as downsides to the new design. Some of the most notable include:

  • Plugin stats are only available to the plugin author by logging into the admin.
  • Tabbed layout was replaced with a series of Read more links.
  • Plugin banner images are distorted.
  • The Installation tab which provided instructions on how to install the plugin is gone.
  • Screenshots do not open in a lightbox, but open on a new page instead.
  • Links to download older versions of plugins is missing.

Many of the issues reported above are known and have been known for months. While it may seem like a lot of people are suddenly complaining and didn’t participate in the feedback loop, it’s important to look back to see how the plugin directory redesign reached this point.

From Prototypes to Open Beta

In May 2016, the WordPress Meta team published prototypes of a new design for the WordPress plugin directory. The announcement generated abundant feedback with a strong push towards adding data instead of removing it.

Many users reported that moving the plugin author, last updated, and active install information from the plugin’s listing page to the plugin’s details page was a downside. Samuel Sidler, Apollo Team Lead at Automattic, responded to the feedback explaining why he didn’t think showing the information to users was useful.

Author, as you said, is only really useful for insiders. The latter two, meanwhile, are already taken into account in the search results. If a plugin doesn’t have a recent compatible version, it will move down the list. If it’s too old, it won’t get shown at all (which is the case today).

Active installs is more interesting, but we account for it weighting search results as-is. I actually find it refreshing to not show the active installs as it allows for less-popular plugins to get more downloads. Users will be less likely to click the popular plugins (outside of familiar names) and more likely to find the plugin they actually need.

There was also a lot of discussion on how to improve and display search results.

In June 2016, at WordCamp Europe, Konstantin Obenland, WordPress core contributor, announced that the WordPress plugin directory redesign was in open beta. The team received feedback from Matt Mullenweg, co-creator of the WordPress project, on changing the direction and design of the page. “We’re really just at the beginning of design iterations,” Obenland said. “He thinks we can do better, which he’s right about. We can and we should.”

Kevan commented on the open beta and pointed out many of the same issues that were reported with the prototypes. In July 2016, Obenland announced version three of the open beta. The use of Read more links in favor of the tabbed interface was again brought up in the comments by John Blackbourn.

“I’m really concerned about the liberal use of ‘Read more’ links on individual plugin pages,” Blackbourn said. “They’re being used in order to fix the information overload problem caused by placing all the information onto one page, instead of using the tabbed interface of the current directory.”

In response to Blackbourn, Joy stated that the issue of Read more links had already been discussed and that the feedback was ignored. The feeling of being ignored was also shared by Jon Brown.

“It’s hard to see feedback has been heard and in this case a lot of it doesn’t seem to have been taken into account (bring back tabs, no read more…).” Brown said. “I don’t see everything, or even much, so I could have certainly missed it.”

Mika Epstein responded that the feedback was not ignored and that Obenland was in the process of collecting and collating information.

Four months ago, Kenshino created a ticket on WordPress Meta that outlined usability concerns with using Read more links instead of tabs.

“Clicking on Read more – say on the change log requires me to scroll to the bottom before I’m able to reach the next section,” he said. “Essentially for me to go through all the sections, provided that they are long enough, I’d have to click, scroll a few pages, click, scroll a few pages etc until I get through all the sections.”

The ticket was closed as a duplicate of Greg Ross’ ticket. In it, he suggests a Jump section be added so users can navigate to specific sections of the Read Me without having to scroll through the entire page. The ticket has received minor support and recently had its milestone changed from version three to a future version.

WordPress Contributors Feel Ignored

In a meeting held earlier this week in the WordPress Meta Slack channel, Matt Cromwell, Head of Support and Community Outreach at WordImpress, suggested that the Meta team describe the process of collecting and acting on feedback in as much detail as possible in the announcement post.

“I’ve chatted with a lot of developers and many feel like they provided actionable feedback about the direction of the plugin directory and all advice was ignored,” Cromwell said. “Whether or not that’s an accurate assessment, it’s a real feeling that is shared by many. I’ve been here most weeks and know well that feedback was received and acted on, but there’s still a strong perception.”

Despite the concerns of using Read more links brought up during every phase of the project, they don’t appear to have been addressed. Other gripes expressed by members of the community at the outset of the redesign have largely remained. It’s no wonder that so many people feel like they’ve been ignored.

Mullenweg commented in the meeting saying he, too, felt ignored, “For what it’s worth, I feel like my feedback was ignored as well,” he said. “I hope we can do another major iteration on the directory, because I’m not really a fan of the new one.

“Perhaps the WordPress backend will make it easier to make incremental improvements in the future, as being on bbPress before was often cited as the reason things were slow to iterate previously.”

Although Mullenweg’s feedback in June 2016, was not public, his recent comments indicate not much has changed between then and now.

Alex Shiels, a member of Team Apollo at Automattic, says user feedback wasn’t ignored. “It’s not that anyone’s feedback has been ignored; just that we’ve had limited resources and a big back-end component to the project, and a lot of conflicting requests” he said. “I’ll make sure to give a clear invitation to provide feedback in the announcement post, and include something about future iteration.

The Inability to Measure the Impacts of the Redesign

Kevin Hoffman, who participated in discussions on trac about how to display screenshots, asked if there was any user data and feedback that the Meta team could share that led to some of the decisions and changes that were made. For example, changes to the UI or how often the Read more links are clicked. Shiels responded that the team does not have that data because they don’t have the tools to do it.

“There is Google Analytics tracking, but access is very limited and from what I’ve seen of it (I don’t have direct access) it’s messy and hard to draw objective conclusions from,” Shiels said. “I absolutely agree that analytics and A/B testing would be great, but we just don’t have the tools for it right now. I think that needs to be considered a next step, but a separate project by itself really.”

Hoffman stepped back from contributing to the redesign after receiving this response seven months ago from Sidler on Trac where he says much of the feedback received was from plugin developers and not representative of users.

First, we’re building the site for users, not developers. Certainly some of our users are developers, but not the vast majority. One of the WordPress core philosophies is building first for the 80%, not the 20%, and that applies in this case as well.

Second, almost all of the ‘overwhelming community feedback’ we’ve received so far is from plugin authors who are (typically) developers. It’s all great feedback and we’re obviously listening closely. But it’s not necessarily representative of the users who visit the plugin directory regularly. We’ll certainly run user tests before launching, and we can probably run a more obvious beta, with a link from the current directory to the new one. But the directory isn’t yet ready for that.

Third, this ‘self-imposted limitation’ as you call it is not arbitrary. Rather, it’s the result of research and in-person discussion with a number of designers. Will it be the final design? I dunno. But since we haven’t tested it with real users (see the paragraph above), it’s hard to say right now. I think pursuing this direction is worthy for a number of reasons, which aren’t really worth rehashing here. First and foremost though, until we build out a product that can run through user tests, we won’t know if it’s the right direction.

How can the team know that the redesign is working or is the right direction if WordPress.org doesn’t have the tools necessary to measure its impacts? There’s also the question of what data was obtained or user testing done that supports the decisions that were made? Who are these users and how does the team know what the best user experience is for them?

Iteration Is the WordPress Way

Now that the new directory is publicly available, the Meta team is reviewing feedback, gathering bug reports, and organizing tickets on Trac.

“If there are tickets that were closed or postponed during the project that you think need revisiting, then this would be a great time to start reviewing and re-opening them,” Shiels said. “We already have a bunch of tickets against the future milestone.

“The new directory has been built with future maintainability and iterative enhancement in mind. We’re looking forward to hearing feedback from the whole WordPress community, and making regular improvements and additions.”

To report a bug or enhancement with the plugin directory, you’re encouraged to create a ticket on Trac. If you have any questions concerning the directory or would like to get more involved, visit #meta on Slack.

The Vocal Minority

There are a lot of people voicing their complaints about the WordPress Plugin Directory redesign and to summarize them as a vocal minority is unfair.

“Let’s also not forget that one of WordPress’s philosophies is The Vocal Minority,” Aaron Jorbin, WordPress core contributor, said in a conversation on the Meta Slack channel concerning feedback. “Many people being loud on the internet isn’t a reason to do anything.”

As I’ve documented above, the issues people are complaining about today are roughly the same as those reported during the prototype stage more than seven months ago.

If months have passed and the issues brought up by members of the community who are part of the so-called vocal minority were not addressed before shipping to the public, can we blame them for complaining and feeling ignored? What about those who think that getting involved to be part of the solution was a waste of time? While the plugin directory design will undoubtedly improve with time, the complaints and concerns expressed by people this week are justified.


90 responses to “WordPress Plugin Directory Redesign: Why So Many People Feel Their Feedback Was Ignored”

  1. First, we’re building the site for users, not developers.


    I actually find it refreshing to not show the active installs as it allows for less-popular plugins to get more downloads. Users will be less likely to click the popular plugins (outside of familiar names) and more likely to find the plugin they actually need.

    It is clearly against the best interest of users, as anyone wants to get the best plugin.

    • To clarify: search tool should provide me with the results and with additional information that would help me evaluate those results in order to make a decision.

      I do not need search tool to make a decision for me, as active installs is one of the ways to tell about the quality of the plugin, as people would not use constantly a product, that they don’t like.

    • Number of installations gives me an indication of ‘quality’ – you don’t get 100k installations if the thing doesn’t work and bugs are not fixed. And, more users find those bugs faster, plus more people will report those bugs…. In that sense it might not be ‘fair’ to new or smaller plugins, but numbers do speak…

      In search, now WP decides which plugins to display at the top. Is that a prelude to paid listings (like the regular search engines)?

      • In that sense it might not be ‘fair’ to new or smaller plugins, but numbers do speak…

        The search experience could be changed to show, recently added plugins that match the search, and the popular ones listed at the bottom.

        Sort of like paid listings at the top of Google search results, maybe even turning that newly added section into a horizontal scroll area to show the plugins, or add a view all link to recently added plugins (that match the search query).

        Then you get both the newer plugins which may solve your need, or you can see the tried and true “popular” plugins.

        While I don’t know if this is feasible for the search results, it may be an option worth pursuing.

  2. I tried to make my feedback positive, but I know there were a number of things that I thought needed work. I’m glad to see that some changes were made and the current iteration is better than the one I tested.

    At this point, I guess my only thought is, if they weren’t happy with who they had providing feedback, was anything done to get feedback from non-developers?

  3. Unfortunately, I have yet to see the 80%-20% rule used with anywhere near concrete evidence. Its virtually always anecdotal. Even when I, not that I’m perfect, cite the philosophy its usually anecdotal (although I try to at least mention that its anecdotal).

    I’m going to try to not try invoke the powers of 80%-20% or vocal minority philosophies without concrete data in the future. I hope others will do the same.

    • Matt it is my great pleasure to agree with you on this. The reason we are not getting a visual code editor in WordPress 4.8 or any time soon is that apparently syntax highlighting is not needed by 80% of the people, even though 20% (if not 35% of us) would use it and value it multiple times/day every day.

      Andrea Fercia told us repeatedly:

      I’d strongly encourage to prove with data the real need for a syntax highlighting feature which, by the way, it’s not the only feature CodeMirror has. I’m open to change my mind when someone can prove CodeMirror would be appreciated and actually used by 80% or more of end users.

      I have money and developers to contribute to integration of a modern visual code editor based on CodeMirror or Ace project but we can’t even it get off the ground seven years later and 116 Trac comments later.

    • True. More often it seems to be used as a fallback argument when all else is failing,. Honestly, the rule often doesn’t logically apply, especially when a small percentage of affected WordPress users might represent hundreds of thousands.

    • While I agree the 80/20 rule is not used correctly, the underlying issue isn’t the rule but our inability to use it due to lack of and inability to collect data and information about the end-user. IMO any replacement of the 80/20 rule needs to include a clear direction toward user-centered design and focus on data gathering, testing, and accessible low-barrier feedback methods.

  4. Exactly right. What we see is a slew of excuses meant to allow those doing the redesign to ignore input and ignore feedback once the redesign comes out. But this has been going on since at least 2015 with an earlier Plugin redesign rollout, and of course well before then. The same attitude, the same level of poor quality interface redesign, and the same desire to simply push out the new design and *get it done*. Legitimate concerns are rejected for a slew of badly thought-out reasons, including: 80/20, not for developers, you don’t have the data, that’s an edge case, we do listen to users (just not you), etc.

  5. Considering the scale and reach of WP and WP.org and the impact even minute changes have on thousands or even millions of users, it is time we start proving redesigns with quantitative and qualitative user testing and telemetry before those changes are implemented. This is no longer a project for and by the people who build it. We are building solutions for millions of people who do not have a voice when we make decisions. With great popularity comes great responsibility.

    • The Plugin Directory seems like a logical place to start with telemetry. It’s a controlled, centralized environment that doesn’t require pulling in data from thousands of individual WordPress installations. It would be worthwhile to gather this data to not only measure design decisions of the Plugin Directory itself, but also to gain experience before potentially implementing telemetry across WordPress as a whole. This would be a good proof of concept so to speak.

      • Matt didn’t say the 80/20 rule is invalid; he said it’s not used as intended, which I’ve experienced firsthand. Telemetry is exactly the solution that allows us to use the 80/20 rule as intended. If we can identify features with which 80% of users interact, then we can more easily settle debates regarding what features are worth keeping, removing, or improving.

        • I was kinda sarcastic, it is totally obvious the guideline is not used anymore for some time. There is very little overlap between what users and developers ask for and what is actually done in core. Admitting that something is not followed is better than pretending it is.

          Having telemetry will not make a difference if no one cares. Can you see oemded removed now that even the tavern stopped using it?

          • It’s difficult to make people care when it’s opinion vs. opinion. When it’s opinion backed by data and validated through testing, then that opinion becomes harder to dismiss. So while you say telemetry isn’t worth it because no one cares, I’d argue that telemetry is precisely what will make people care.

            Your evaluation of oEmbed based on the Tavern’s use of it is exactly the kind of anecdotal evidence and conjecture that is driving decisions today.
            We need to make decisions based on the experience of the many, not the individual.

            • Yes. If you come into the argument jaded it’s likely the outcome will be, predictably, as expected. But I’ve learned the hard way the community can get pretty small if one is always the naysayer. Much more instructive and constructive to present well researched data and reasonable alternatives.

              There are quite a number of improvements to the new plugin directory. I’m sure these issues will get ironed out.

  6. To be perfectly honest, I don’t care about the UI. I’m not a UI guy, and hey, the designer types can do whatever they like there and I’m fine with it.

    The thing to me is that nearly all the complaints are about the look of the thing, and I care about function. And I’m so freaking thrilled to finally have this launched. I can finally build everything way more easily. Me and others on the plugins team have been talking about process and interface on the backend, and how to bring on open review systems.

    This could have been running the “Classic” theme and I’d be fine with it. There is a disconnect between people looking at it and those doing the work.

    And that’s fine. Feel free to debate the design and adjust it and change things to fit your needs. But the design has been public and visible since June and only after launch we get this much heated complaint? Maybe some of the feedback was ignored, maybe not. I dunno. I am just thrilled by what we​ can do going forward.

    • Otto,

      You have just explained the whole issue. You think function is separate from design. And apparently the “designer types,” as you call them, think design is separate from function. You’re both wrong.

      Good design flows from function. You can’t have one without the other.

      So those complaining about the abandonment of tabs, for example, are not just talking about a major design flaw. They are talking about a loss of functionality too, because important information is now no longer (easily) accessible.

      It’s good that you can now do back-end stuff that you have wanted to do before but now can’t. But if those using the repo can’t find the information they need, you’ll just be fiddling while Rome burns.

      • I understand the tabs thing, but at the same time, I understand the reasons for a single page design. Tabs make it easy to put lots of information there. But information overload is also a problem. I liked tabs, but I want to see what plugin authors do with limited options as well. A single page design limits what you can post, and limitations are not a bad thing in and of themselves. They can create focus as well.

        We needed to launch to continue developing things. Change is hard, but also continuous. This is not an end, it’s a beginning.

      • Otto, please see the comments of the v3 Open Beta announcement that Gary linked above. These are very much the same concerns that are being voiced this week.

        You mentioned here that you don’t care for design, and you mentioned in AWP that many voices don’t mean any more to you than a single voice. With that perspective, I can understand why many people voicing concerns of over the same design issue may not make an impact on you.

        I happen to believe that 1) Design does matter 2) It is not separate from functionality and 3) Many people voicing the same concern should raise its priority.

        • That’s fine, but when you make it appear that the community is only now voicing concerns that are clearly on the record en masse from 9 months ago, it only perpetuates the perception that the feedback is ignored.

        • Kevin, Gary, allow me to rephrase.

          If you have design concerns, then they are valid and can be addressed in Trac tickets. Places where things actually get done. I spent the last two days doing just that for all sorts of issues.

          I never saw any feedback for the beta. Not one bit. But, I didn’t help create any of it either, I’m just dealing with it now. We can change things. We can discuss and argue and create together. That really wasn’t possible prior to launch. But we’re listening and writing code and implementing patches right now. The whole thing is open source. Help me help you. Happy to help. Really. ?

          Every comment goes to my inbox. So do most plugin commits because that’s all I do all day long. Hey, I’m crazy. But never feel that I’m not listening. I am. I just disagree sometimes. That’s not personal.

        • Kevin, I understand the frustration, but you’re never being ignored. I hear everything and pay attention. If there is a real issue, you can always count on me to try to address it. Just let me know.

          There were decisions made here that I know I’ll eventually need to undo. Mostly minor backend trivial stuff, but problematic nonetheless. The truth is, until it’s live, you never know what will fall out from it..

        • If you have design concerns, then they are valid and can be addressed in Trac tickets. Places where things actually get done.

          I agree 100%; if you’ve got the skills, contribute on trac.

          But how do we obtain input & feedback from non-devs in the wider WP community who lack the technical ability to contribute on trac?

          We can send out tweets about Slack chats, Make blogs, and surveys, but that only reaches a relatively small core group of devs & power users. Then that feedback gets cited as being unrepresentative as a result.

          Suggestions from Morten & Kevin about telemetry and user testing get a big thumbs up from me.

        • If there is a real issue, you can always count on me to try to address it. Just let me know.

          An attitude far too common among certain WP folks. Its often a real issue for those that have a problem or suggestion. Its a bit arrogant.

        • I’m aware of this, but I’m not a designer. I only ever looked at the function. I’m a terrible designer. I know my limits.

          Otto – I was active and vocal months ago about how unusable I felt the “read more” links were. That discussion was mostly on make, before there were really trac threads, but then did move into trac.

          The summary was as beautiful as the new design looked, many over the course of months voiced frustration over the usability/layout issues with the biggest by far being the switch from tabs to “read more” links.

          I banged my head in chorus with others against the .org wall as long as a could before I gave up and went away. From Mark it sounded like maybe their might be a design iteration with tabs, that never came that I saw.

          Let me say however that I never once saw your voice in those discussions, and wouldn’t really expect to hear your voice
          in design discussions. Perhaps people are confusing you with Sam Sidler who was in that discussion and was the one talking “80/20” and dismissing the opinions expressed on make and trac as being from “developers not users”.

          So, Yay! There is a new backend and search that is a huge accomplishment, props on that.

          But, Boo! Design feedback from the community was summarily ignored. The best we got in response was “you were heard we just decided not to make any changes”. You can _not_ look through all those old threads and not see that 99% of the comments were against the “Read more” links, yet, that’s what we ended up with with no alternatives ever presented.

    • Otto, if we all had your level of self-reflection and spoke as directly as you the world would be a better place.

      I understand your position. It is that of a developer, and I have no doubt the developer-centric parts of this revamp introduce much needed improvements and enhancements. The problem is in the design, more specifically how design feedback is handled. As others have pointed out, while filling a Trac ticket will make an issue obvious and addressable, that very modality adds an unreasonable level.of complexity to the feedback process. The user most impacted by this update (and any update to .org and core) is the average user whose only interaction with the application and the site is the application and the site itself. This user is likely not aware of Trac, nor should they be: Trac is a development tool, not a user feedback system. What is missing in our current approach to redesign is outreach and systems for gathering feedback from the actual user. Without this, we are effectively doing work whole blindfolded, or more likely doing work for ourselves rather than our audience. That to me is the beginning and end of this issue: What little data is available is written off as the “vocal minority” yet we know nothing of what the actual person at the user end of our interface wants, needs, thinks, and does.

  7. It bothers me not to see active installs, because that’s one of the ways I can tell whether an older plugin still works. (Or at least, that it will probably work.) Very few people actually update the works/doesn’t work info.

    I also find the long scrolling arrangement awkward, but since I was used to the tabbed layout, the problem may be primarily that it’s different and I need to get used to it. Tabbed layouts were invented for a reason, though I understand they can be problematic for accessibility.

    No matter what changes you make, some people will complain. The only way to know whether the changes are a problem or just unfamiliar is to give the new directory some real-world use.

  8. This is interesting as I have seen this conversation and debates in several places. As I mentioned in a slack conversation, myself, it works fine for me in how I use the plugin directory. Basically to search and get a link.

    I can say, as I have said on other conversations, that the typical user, from my experience of working with them the last 7 years, very seldom goes to look for plugins there. They either search or find it via their dashboard, install and activate. Typically that is a result of someone having told them to install it or some other research they have done. As a result, they could care less what the page looks like there as long as they can snag their plugin of choice. Just saying…

    • the typical user, from my experience of working with them the last 7 years, very seldom goes to look for plugins there [plugin repo]. They either search or find it via their dashboard, install and activate.

      My experience training people is that if they already know which plugin(s) they want to use, they’ll search and install them from inside wp-admin.

      As Bob outlines, typically they will have already heard of a plugin, or received a recommendation in that case.

      However, if they’re trying to find and compare multiple plugins for a use case themselves, like a membership plugin for example, they’ll search via Google or WP.org first – then they read through the plugin repo pages and/or read plugin reviews from other blogs in multiple browser tabs.

      Once they’ve found the plugin they want, then they search for it through wp-admin. This means that they don’t have to manually download & upload the zip file.

      I can imagine that having to endlessly click the ‘read more’ link on every plugin page just to read the description could quickly becoming annoying, particularly when comparing multiple plugins.

      While the install happens via wp-admin, often the initial discovery and decision to use a particular plugin is a result of having visited the plugins repo page at an earlier point.

  9. Holy Smokes! I don’t want to sound condescending or anything, but I just want to point out that this is an amazingly well written and organized article with a fair and balanced approach. Very polished and high quality. Much like what you’d see pushed by a more traditional news source. Kudos!

  10. Hi,

    I’ll start by saying thank you and congratulations to the people who worked on this Plugin Directory project. I can understand how complex this kind of project can be and the result is mostly nice.

    That sincerely said, i felt my feedback was ignored and one of the things the team who worked on the project might improve is giving the results of the survey *before* doing the changes.

    You mostly talked about design, there’s actually something that has more to do with feature that you didn’t mention. The 2 last reviews are shown on the plugin’s page. Although many of you might think it’s a good idea, it’s not and i shared this feedback on the survey. Why, because plugin authors can’t choose what review to display. It’s the last 2 and you need to live with that! It’s like if a “masochist” brand would choose to display the last 2 user feedbacks on his front page. 1 new negative feedback would then means 50% of people dislike the brand.

    The users who like plugins are not giving nice reviews that often. But when someone is disappointed he will give a bad review more frequently. So a negative impression can last long ;)

    I personaly chose to stop sharing my plugins on the official directory because i don’t have the same goals many have. I build plugins as a hobby, so my needs is to have constructive conversations and feedbacks from users. Users do not have this in mind anymore, they are not there to contribute and participate but there are now there to use and complain.

    It seems to me since the Plugins “repository” became the Plugins “directory” it’s losing this contributive dimension. It’s probably something that is unavoidable in order to take the “customers” needs changes in account so i don’t blame anyone. It’s just too bad the road of this space seems to be a market place rather than a contribution one.

  11. My two main problems are:

    1) I desperately miss the tabs. At least those cleanly organized information. Now it kind of looks like a huge var_dump() and everything was smooshed into one big column. It actually makes things a lot harder to find.

    2) I’ve had a lot of problems with the search, particularly when I’m looking for a very, very specific plugin and it’s just nowhere to be found in the results. I understand the need to do away with exact matches when doing a “keyword” search, but we should at least introduce a more specific “Plugin Title” search for those people who aren’t aimlessly looking for a plugin, but are actually looking to quickly install a very specific plugin they already know the name of.

    • I’ve had a lot of problems with the search, particularly when I’m looking for a very, very specific plugin and it’s just nowhere to be found in the results

      Can you please post some specific examples of problematic searches, either in a trac ticket or the #meta channel in Slack (both linked in the announcement post). In general, search is vastly better than the old directory. That said, there are always buggy edge cases in a new release, and your bug reports help us to track down those remaining problems that are still stubbornly hiding away.

      • Here’s a “buggy edge case” for you. Try searching for a plugin using the word “search” (obviously without the quotes).

        What do you get? Well, the first three hits are fine. Then things go downhill fast.

        We get SEO plugins and — what a surprise — Jetpack contaminating the results. I have to go halfway down the second page of hits before I see, for example, Dave’s WordPress Live Search, a plugin which has over 10,000 active installs and an average rating of 4.6.

        WP Google Search — with over 9,000 active installs and an average rating of 4.5 — doesn’t appear until page 4 of the search results.

        Better Search — with over 8,000 active installs and an average rating of 4.3 — doesn’t appear until page 5. That’s the same page that shows a forum plugin.

        • The three search plugins you mention are ranked lower because they don’t appear to be very actively maintained. They are all 1-2 versions of WP behind, haven’t been updated in 7 months to a year, and haven’t had any support threads resolved in the past few months.

        • @Greg Brown,

          You mean the developers haven’t marked any threads as resolved in the past few months.

          So this new algorithm rewards developers who mark threads as resolved even when they aren’t.

          Things are actually even worse than I thought.

  12. It looks OK, maybe one suggestion – screenshot’s area could use fixed height, as I don’t like if whole website jumping bc. somebody is not able to upload the same sized pictures.

    Also FAQ accordion could react and open also if it’s clicked on the end of right side where arrow icon is. My first visit impressed me, that it does not work, as I intuitively clicked on that arrow, not the text.

  13. Interesting. And disheartening. It’s definitely worse now.

    It’s confusing and harder to find things. That’s a usability downgrade.

    I guess they managed to fit more data on one page by hiding it behind teensy “read more” links. I had thought accessibility was supposed to be a priority these days.

    About the argument that only developers look at some of the stuff, fair enough. But when the plugin description is small enough to be almost meaningless without clicking a teensy “read more”, does someone think that users will like that?

    At least they’re trying to make screenshots better by having a “slider”, but if you click on the image it’s the same goofiness as before.

    Oh well. Maybe they’ll iterate.

  14. Most of the people here are talking about the design. I agree the new design is terrible.

    But the search is pathetic after spending a year and this is what they have done absoultely poor product.

    Jetpack plugin is being ranked for every other keyword in top 4 results. Every word jetpack plugin has in its description the plugin is ranked at top for those words.

    If you search for facebook or twitter or management jetpack plugin is being ranked at top. This is greed and no way WP community should allow this to be continued.

  15. Otto,

    The thrill is only for a limited time, You will soon find out the slowing down the rate of new plugin submissions and most old plugins not being managed any more.

    In new algorithm you’ve literally outranked every plugin with less than 10k active installs which might in future replace the contact form 7, JetPack and other most popular ones. Remember most of plugins which are in most popular ranking did not brought the users from outside it was the directory pushing users and giving them hope.

    The directory is now good for older plugins with high active install but what if I worked on a plugin for 6 months and now I upload it and it is ranked on 4th-5th page. Come on We are developers supporting opensource community we don’t know how to do marketing!!!!
    How are we supposed to bring 40-50k active installs to outrank other plugin even if we know the new plugin is better than those old ones with millions of active installs ????

    It’s not a complaint but a feedback. I am sure in 6 months you will find out the flaws but until then the reputation of plugin directory will be phoof!!.

    All I can see is that you are thrilled because finally you can rank jetpack on 20 different keywords.

    • I have sympathy for devs of newer plugins, but the primary purpose of the repo is to help end users find a plugin that suits their needs.

      If I’m searching for a plugin, I do want see those that have an established track record with high numbers of active installs to generally outrank those that don’t.

      If a plugin has a large, established user base, it generally means that significant numbers of people are happy with it. It also usually means that there’s a community around that plugin.

      Devs shouldn’t rely on only one avenue to build their brand and audience. If they want to compete with larger, more established plugins, then they really need to market themselves outside of the repo too.

      If there are specific examples of search queries that you don’t think provide good results, and have examples of other plugins that should rank higher, you should post those examples on slack or trac (linked in the original article) so that devs can see if the weighting for ranking factors should be adjusted.

      • If I’m searching for a plugin, I do want see those that have an established track record with high numbers of active installs to generally outrank those that don’t.

        You should want the most relevant results, then be able to filter if you want to. Having the base search algorithm bias the results out of the gate (see https://wptavern.com/wordpress-plugin-directory-redesign-why-so-many-people-feel-their-feedback-was-ignored#comment-217148 above) is a terrible idea.

        The fact that they seem to be rigging the results in a decidedly non-transparent manner hurts their credibility. If I search for ‘search’ then the most relevant results should be at the top. If I want to filter by installs, there should be a clear UI that allows me to do so (“Show plugins with more than 1000/5000/10000 installs”). That way I can see if there’s some hot new plugin that does just what I want or, if I want the security of going with something that’s been out there a while, I can narrow the results.

        • The fact that they seem to be rigging the results…

          Who is rigging results, how, and who benefits?

          If the suggestion is that Automattic are intentionally rigging results to make JP rank above other plugins, then I think that’s a baseless conspiracy theory.

          If I search for ‘search’ then the most relevant results should be at the top.

          It’s often difficult to gauge user intent and relevancy for generic one word queries like “search”- is a user looking for a search engine optimzation plugin, a plugin for implementing Google custom search, or are they looking for Omnisearch in JP that allows them to search within the admin?

          If we enter one word into Google, and it doesn’t return results we like, then we add more words to the query to make it more specific. It’s the same for the repo.

          On the issue of active installs being used as a ranking factor, it’s useful again to look at how Google does things. When multiple urls contain similar keywords, Google looks at other signals.

          One of those is the amount of inbound links. If url A has one million links, and url B has a thousand, then that can be used as one sign that people seem to prefer one url over the other, and it should therefore be ranked higher (obviously, this can be gamed, the quality of the links and anchor text are also factors).

          In a similar fashion, it makes sense for the repo to heavily weigh the number of active installs in its ranking algo as a trust signal. If multiple plugin pages contain similar keywords, then other factors have to be taken into account in order to determine which ranks higher.

          JP ranks for a wide range of terms because, for better or worse, it contains a wide range of functionality which is detailed on its plugin page – galleries, carousels, CDN (Photon), related posts, Omnisearch, portfolio & testimonials CPTs, Facebook opengraph & Twitter cards, social sharing buttons, bruteforce security and login protection etc.

          In addition to that, it also performs well when looking at other trust signals – number of installs, last updated, support threads resolved, and user reviews.

          I don’t think these trust factors have been introduced into the algo as part of a conspiracy by Automattic to rig results in their favour; they’re being used to provide users with result pages that contain a list of plugins that are relevant to their query, up to date, are supported, are well reviewed, and have large numbers of sites that appear happy to use them.

          On the issue of filterable results, if you read the conversations among the meta team over the past few months, you’ll see that they too want to be able to allow that, but first they had to work on the infrastructure behind WP.org they before they can introduce that functionality. By all accounts, WP.org appeared to be previously held together by a combination of spider webs, magic, and Otto.

  16. I have written a userscript to bring back tabs, they work doe to the new all on one page they are instant (no page reload)

    (The comments systems shows me my comments are being rejected and I should check fields but they are all filled. Am I banned here? I have another account with a protonmail address and that is getting rejected, they have a feature to add +something to the mails. Tried that but still gets rejected, all protonmail addresses get rejected?)

      • Thank you, yes indeed its so much better this way.

        And without having access to the stats I doubt that Ottos claim that most users where complaining about looks and not functionality is wrong. I think it looks great but the functionality is bad.

        I think the wp.org site has a repo and is “open” but I really have the feeling that this is a place where issues go die rather then that I could help make things better. I coded this up in 30 minutes or something, feel free to send me some CSS to make the buttons look better or look like actual tabs and I include it. I just quickly used the first CSS button generator that I could find. In Firefox the buttons don’t have this big border. It more a proof on concept how I think this thing should be. I mean the all in one page stays the way it is and one with some jQuery knowledge can make this in a very short time Many many users want the tabs back not because they don’t like change but because its just horrible UX.

        There is also the bug that the headers look pixelated, I could fix that in 5 minutes as well, I told them how but they have not bothered, they also have not put out new specs for the headers.

        This makes me think “WTF are this people doing?”. I really don’t get it.

        A Userscript will only help a very few people who actually even know how they work and willing to install the browser extension needed to use them.

        The fact that to me there was no change visible from when it were first introduced as beta and the day they throw this horrible UX beta with bugs out to people.

        They use the same ID twice for the screenshot section, I mean … this thing is still totally beta and they have entire teams working on this and had plenty of time to make it right …

  17. I am just a standard user of the Repo and not a developer, but I rely on the “Active Installs” to help me decide if a plug-in, along with other information about Plug-in, has a track record and is worth trying. It appears that WP has decided that this information is not a valid indicator for users to assess a plug-ins value. I personally would like to have this information back.

    • In short the powers the be think install count prejudices people against plugins with fewer installs that might be a better fit for the users needs.

      The problem is users still aren’t getting enough info on the search page, or plugin page, to really evaluate the differences between plugins. Doing so is near impossible, although review count and rating is probably a better indicator than active installs.

      Only now repo plugins are going to become like iOS apps where they pop up request for 5 star reviews all day long making mass low quality reviews essentially just as useful as install count was.

  18. Hi there, i have moved to other CMS systems some months ago as WP was going in a very annoying direction.
    One of my clients insists having a WP site so i came back and the new plugin repo is a complete disaster to say the least. Luckily enough they added back number of active installs, version number and when it was last updated.

    I really hope they will revert back to the tabs in the layout, as only a no-brainer can think that oversimplifying things will make WP easier to use. If so why not just show a banner and the name of the plugin and nothing else? Newbies can not read, so they do not need anything as complex as tabs or more than 8 characters… on the otherhand developers should just shutup because they are in minority and WP is not intended for them but for the users.

    My only question is: Who does WP Core think will use WP in a year if they continue developing and argumenting like that?

  19. I’ve asked this question before – because some users have asked me the same question – but can’t seem to get an answer: is there documentation on how Featured plugins get, you know, featured?

  20. I guess they must have made a lot of improvements already in the past week, because most of the points in this article are not (or no longer) true:

    Plugin stats are only available to the plugin author by logging into the admin.

    I can see this under Advanced View for any plugin without logging in.

    Tabbed layout was replaced with a series of Read more links.

    This one is true; having everything summarized on the first page along with easy links to get more details for those who need it does not seem like a bad thing to me, though.

    Plugin banner images are distorted.

    In a quick review of a dozen or so plugins, this didn’t appear to be true for any of them.

    The Installation tab which provided instructions on how to install the plugin is gone.

    True… But this tab is pointless for 99% of plugins, which are installed by clicking the “Install” button in the directory. Couldn’t be more simple. Plugins with more complex instructions seem to be adding it under the now-more-prominent FAQ section, which works just as well.

    Screenshots do not open in a lightbox, but open on a new page instead.

    Yep, that’s annoying. This needs to be fixed.

    Links to download older versions of plugins is missing.

    Nope. They’re under Advanced View. I actually found them more easily in the new design than I used to be able to on the old version.

    I’m going to assume these were all legitimate issues at some point, but expecting the brand new directory to be absolutely perfect at launch was hardly realistic, and personally I’m extremely impressed at just how rapidly they’ve improved it in the past 5 days.

    • There was no reason for the new plugin directory to be launched broken except arrogance. Read the other posts, CFC. There’s nothing impressive about ignoring six months of detailed feedback and then launching broken.

      Thank heavens some steps have finally been taken to mitigate this disaster.

    • I am not impressed:
      As these “fixes” should have been implemented BEFORE the relaunch – the feedback timespan was about 9 months or so, at least 6 I guess.

      The “Advanced” view as additional page was just launched about 24 hours ago, a lot of comments here for this post, are almost a week or so old. Just consider this.

      A lot of main critics from users and plugin authors alike still remain:
      1) stupid, useless “Read more” toggles –> a usability desaster on desktop and mobile alike
      2) the screenshots linking/browsing is a desaster: everyone expects for such slider/carousel appearance the opening in a lightbox, users are “trained” to expect this behaviour; and it was mentioned numerous times in the feedback saison months agon…!
      3) The long content of the one pager overall: so bring back the tabbed layout, make it usable again, and visually appealing at the same time
      4) the issue with plugin banners is this, that a lot of plugin authors still have the former default size of the banners packaged in their /assets/ folder –> this needs to be updated by the plugin authors themselves, it’s not fault of the team, entirely — However, they could have been informed plugin authors weeks or months ago about such an upcoming change. Nothing happened that way, and this is a fault of the team, in my opinion.
      5) The attitude of launching such a half-baked thing with ignoring lots of feedback is an issue on its own. But “praising” this as an improvement, especially when considering the UX issues on mobile, this just left me frustrated…!

      However, to make the directory a bit more useful again for me on the desktop I created a Userstyle! You can use this in any desktop browser with the “Stylish” add-on (works in Firefox, Chrome, Safari, Opera): https://userstyles.org/styles/140944/wordpress-org-plugin-directory

      The problem with this is, that Userstyles don’t work on mobile yet, so here are all the issues laid out above more “prominent” and therefore hurting…!

      • I thought I was harsh but yes. Your so right and I agree with pretty much anything you said. Half bakend thing thrown out to people without new headers specs, or a warning that headers will have no text overlay (mine is specifically made for that), without stupid short default section height, no tabs and even issues I did not notice I read above.

        Otto I am not sure where your claim that the banners only effect hi-dpi screens stems from because its untrue. The banners are pixelated on normal 1:1 dpi screens. They are loaded in the smaller size and then stretched a bit bigger. Its clearly visible because they look blury and it could be fixed at least temporary by serving the high-dpi to everyone.

        My time feeling is had, I did not even realize that this was 6 months of feedback before this beta was thrown at us. Yeah pure arrogance!

  21. Hello,

    I just wrote to WordPress.org:



    I just realized that the “plugin is broken”/”plugin works” feature is missing. This is a huge step backwards for me, since I rely on this feature a lot.

    I know that not many people used it, especially for most minor plugins the results weren’t very meaningful. However for huge plugins like WooCommerce and W3 Total Cache it preserved me from updating to a new version with critical errors. I know about backups and dev testings, which are both very important, but this user feedback was also very useful for me.

    The latest version of W3 Total Cache for example had around 20 people saying “not working” and only 1-2 saying “working”. Now people don’t have this kind of “warning” and many might update, experience the same bugs. This could be avoided with the simple, yet important feature.

    So I kindly suggest to reconsider integrating it again. I say this as a user, not a developer.


    • Nobody used it. Literally, nobody used it at all. That system relied on users voting, and none of them did. They stopped a long time ago, after we added reviews. So, reviews it is now.

      As to your example, 20 ppl saying they had problems out of over a million running it is not a representative sample. Nearly all plugins had “broken” simply because the only people who voted were those with issues.

      • Well to be correct, very few people used it not nobody, and yes its right the people who voted on it are more likely those who had problems but when the plugin actually works for 99.9% of the people and you get no feedback from who every voted “does not work” its pretty disappointing. So no matter now disappointed I am with the current beta of the plugin dir. I agree with this one. I wont miss this and I hope its never comes back.

  22. Can’t comment on the process, while I love this dialogue. It’s easy to adapt to the new spiffy UX, but my concern is with the usability outcome. I’m mostly a site builder and I train content managers and site builders, and researching components is essential at this level of WordPress work. We have extensive review and recommendation resources across the ecosystem, but the repository is the essential index for freemium plugins.

    “Relevant” results are only relevant if one agrees with the algorithm. The relevant use case is typically a committee case, not something relevant to me nor any particular user. So what might be accomplished for a mass spectrum of users?

    Filters and sorts are a common means to add some levers to an algorithmic output. Keyword searches are just a ticket to the fair, taxonomy lets one drill down into a vertical, and folksonomy allows for hyper-jumps. Sorting on dimensions offers more contextual approaches for what might matter in some case.

    Taxonomy, filters, and sorts should be moved up from the backlog. I can dig down to find tags as filters in single listings, but why not place top tag links atop the results list? Boolean operators don’t apply, but that’s an old-school tactic hardly used anymore. I see that installs and recency sometimes trump matches, but I can’t adjust that logic with any filters nor sorts.

    Not saying the algorithm is a problem, just saying that users benefit from some transparency. UX can guide users towards those benefits in simple ways. The themes Feature Filter is an example, though not a very handy one.

  23. Looks better but usability has gone backwards.

    Plugin reviews ( like any reviews ) generally are not credible, especially for anonymous users ( competitors trashing you haha ).

    Overall, not a great a fan but respect the fact people worked on it.

  24. Good news – today I see evidence that they have heard the many requests for tabs or links at the top of the individual plugin pages. Some, but not all, of the old sections now have links at the top so you can jump right to where you want to be.

    On a couple of plugins that I’ve checked, I’m seeing Details, Reviews, Installation, Support, and Development. But I’m missing FAQ, Screenshots, and, importantly, Changelog. Hopefully these will also show up soon!

    The Trac ticket referenced in https://wptavern.com/new-userscript-restores-tabs-to-the-wordpress-plugin-directory was just closed: https://meta.trac.wordpress.org/ticket/2278


Subscribe Via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.