Plugins that don't pay rent
When that chat widget first showed up, it was so shiny and full of promise that you didn’t hesitate to give it sitewide, above the fold real estate.
It stayed around a while, and you got used to it. It collected some email addresses, helped route some visitors to Customer Support. People were using it, so it seemed obvious that it had some value.
Here’s the thing about that charming widget: it captures your visitors’ attention, even when they’re trying to accomplish a task that needs no chat assistance.
It takes up space that could be occupied by an even awesomer widget … or by nothing but relaxing whitespace 😌.
Test “widget vs no widget”
The simplest A/B test in the world is one that simply hides an element. Your web developer can implement it in her sleep. QA is as simple as “Yup, no widget 👍.”
You can run this test as long as you need to in order to measure whatever degree of impact you choose to care about. When it’s done, you’ll know exactly how much the widget impacts conversions on your site. Assuming it impacts them at all.
Watch out; the widget will happily report on interactions, and present them to you as if they were (a) inherently valuable and (b) only possible thanks to the widget.
This is misleading; of course if you put a button in front of your visitors’ faces, some of them will click it. It’s the overall impact of having the button that we’re concerned with.
And if you find out your widget isn’t paying rent, you’re free to evict it.
Freeloader widget in the wild
Here’s a quick example of a widget that I’ll wager has never had to pay rent:
On page load, the chat widget covers the product name, description, and inspirational tagline. And the widget isn’t even usable - chat is offline.
If you’ve let some of these freeloaders creep into your site experience, take them to task with a “show/hide” A/B test. Keep the ones who can show their value, and lose the ones who can’t. Your visitors’ eyeballs will thank you.