I’m turning our most popular plugin into a chat plugin

For about a year I’ve watched our most popular plugin decline. Social Icons Widget, the one with the most installs in the WPZOOM catalog, went from roughly 135,000 estimated active installs in June 2025 to just under 100,000 today. A quarter of the base, gone in fifteen months. No bad release to blame, no bug, no wave of bad reviews. It just keeps going down.

Some history first. The plugin started in 2015, eleven years ago, as a small companion widget for our themes: a way to put social icons in a sidebar without hard-coding them. It was never meant to be a product. It turned out to be useful on any site, not just ours, and that is how a free widget with no plan behind it got to 135,000 installs.

Why it is falling now is not a mystery. WordPress has shipped a native Social Icons block since 2020. Every new site gets it for free, it’s in the block inserter, and it’s good enough. So why would anyone install a plugin for social icons? The people still running ours are mostly sites that installed it years ago and never had a reason to remove it. Nobody new is arriving.

The Pro version

There has been a Pro version, Social Icons Widget PRO, since June 2021, at $39 to $149 a year. In four years it made about $8,900 in total from 223 sales: $1,700 last year, $1,366 so far this year.

The free plugin is already generous: 400+ icons from five icon sets, custom SVG uploads, icons in menus, sharing buttons for 12+ platforms, all of it working in the block editor. The Pro adds the things a marketing team would want: a floating share bar, share counts on the buttons, a share analytics dashboard, a Pinterest “Pin it” hover button with custom pin images, a like button, and lately AI share buttons for ChatGPT, Claude and Perplexity. Good features. But read that list again: every one of them assumes the site has enough traffic to care about share counts and analytics. Most sites don’t.

That is 223 sales from a free plugin that at its peak had 135,000 active installs. Not because the Pro was bad, but because nobody needs to pay for social icons. Most of those 135,000 sites are blogs, and blogs don’t spend money on plugins for things WordPress already does. That’s the lesson, and it took me longer to accept than it should have.

For contrast: Recipe Card Blocks, our best-selling plugin, shows 10,000+ active installs on WordPress.org. The Pro version is a separate plugin that does not need the free one (it deactivates it when you activate Pro), so paying customers are not even counted in that number. Even so, month to month, Recipe Card brings in about fifteen to twenty times what Social Icons Widget PRO does, with a ninth of the installs. Installs and revenue are just not the same number.

So the install base was big, but it never made much money. Once you see that, losing part of it stops being scary.

Changing direction instead of rewriting

The obvious move would be to build a new plugin. But a new plugin on WordPress.org starts at zero: zero installs, zero reviews, zero ranking. The install base is still the asset, not as revenue but as distribution. So is the 4.9 rating from 144 reviews. So instead of letting it fade, I decided to change its direction: slowly re-adapt it into a chat plugin.

I did this in small steps so it wouldn’t look, or feel, like a radical change:

1. Click to Chat floating buttons. A few months ago I added floating icons for WhatsApp, Telegram, Messenger and Viber. A natural extension of what the plugin already did: a floating icon that opens a conversation instead of a profile.
2. Chat keywords in the name. The plugin name is the strongest ranking signal in the WordPress.org search. I added the chat-related terms there.
3. AI Chat, and a new name. The most recent step: Yamidoo now sits inside the same launcher, an AI chat that answers from your own pages and hands off to a human when needed, and the plugin is now called WPZOOM Connect.

So what I’m really doing is a trade. I’m trading the “social icons” keywords, a term I can’t win against WordPress core, for “chat”, “click to chat” and “AI chat”, where demand is growing and where a plugin can actually be the best answer.

Small steps

The sites already running the plugin installed it for icons and sharing buttons, and that is what they keep getting. Every step kept everything that already worked and added exactly one thing. Someone who only wants icons in the footer still has icons in the footer; nothing moved on them. The rename came last, not first. By the time the name changed, the plugin had already been doing the new job for months.

The pushback

One user emailed me, left a one-star review, and uninstalled. In short: if I wanted AI, I would install it myself; putting it into an unrelated plugin without asking is not okay. The review was later removed by the WordPress.org moderators over its language, not over its argument, which is a fair one.

I understand the reaction. Nobody wants to see “AI” show up in a plugin they installed for something else, and right now the word alone is enough for some people. What I can say is that the AI chat is opt-in: it is off until you click “Turn on AI Chat” in the settings, and the icons and buttons work exactly as they did before. But I would rather write this down than pretend the reaction didn’t happen. With ninety thousand installs, a change like this will always cost some people. What matters is whether it brings in more than it costs.

Keywords

This is the part I find most interesting. The plugin still ranks #2 for “social icons widget”, “social icons” and “social media icons”, and #7 for “share buttons”, because those features are still in the plugin. Nothing was removed, so nothing was lost.

Meanwhile it is already #5 for “click to chat” (up four spots recently) and #6 for “ai chat”. Those positions were earned before the Yamidoo launch, by the Click to Chat step and the name change. So by the time the AI chat shipped, the plugin was already being found by the people looking for exactly that. That was the point of doing it in this order.

Competitors

Two very different pictures. On “click to chat” the plugins above us do only that, and have 200,000 to 700,000 installs each. That is a fight for the long term. On “ai chat” the results are still a mess of MCP servers, form builders and messaging plugins: nobody owns that search yet, and with 90,000 installs WPZOOM Connect is already one of the largest plugins on that page. That is where the room is. And on “social icons” we are still #2, so the trade so far has cost nothing.

The plugin’s new job

There is a bigger idea behind this than one plugin. A WordPress plugin with a Pro version sells a licence once a year and then spends the year supporting it. Yamidoo is a hosted service with a monthly plan, and the plugin is the free front door to it. What I am really trying to do is turn a plugin business into a SaaS business, using the distribution we already have instead of buying it. Social Icons Widget is the first test because it had the most installs and the least to lose.

The only real moat in the plugin is Yamidoo: the AI chat, the one thing WordPress core will not ship and a block cannot replace. So the plugin’s job is now to be a path to it.

So far it has brought about a hundred free Yamidoo users. Some of them will convert. More importantly, they are helping us learn who actually pays for this: businesses with visitors asking questions (stores, agencies, hotels, SaaS), not blogs. There is also a standalone Yamidoo plugin on WordPress.org, the same connector without the icons. It doesn’t matter which one people install. It only matters that the sites that need a chat find one.

So yes, I’m fine continuing to lose installs, as long as the ones that stay, or arrive, are the ones that matter.

What I expect

One early signal, with a caveat: since the rename, daily downloads are up, about 4,700 a day over the last 30 days and 63% more than the month before, per wp.md. But downloads include updates, and I ship often, so I don’t read much into it yet.

Installs will keep going down for a while. The people leaving are leaving for the social-icons reason, and the people who want chat haven’t found the plugin yet. If the bet is right, the curve turns at some point in the next year. If it’s wrong, I’ll have learned something about how much a plugin’s category can be changed from the inside.

Either way, I’ll post the chart again in six months.