Over the next few weeks, Salesforce’s killer collab app, Slack, is rolling out Huddles. Huddles are voice conversations in a private message or channel.

But are Huddles really better than Slack calls?

My introduction to Huddles was rough. A collaborator clicked the nameless toggle and I was greeted with a message that he wanted to have a Huddle with me.

1. What the heck is a Huddle?
2. Wait, where did it go?

On his end, he exited as quickly as he entered – confused. Then what did I do? Clicked the same toggle he did and experienced the same sense of confusing dread.

Rough.

So, I went looking for a setting that would allow me to turn this confusing and potentially harmful feature off. Guess what? No such setting existed. 

Now, this isn’t the first time I’ve run into this, and I am going to guess it’s the same for you.

I have a lovely new car. A 2021 Outback. I love this car – it’s beautiful but it came equipped with one of my least favorite modern car features – Auto-Stop/Start. It’s the thing that cycles down your engine when you are at stop lights. It’s supposed to save on fuel but after 4 months, I haven’t even saved a quarter of a gallon. But it has succeed in irking my nerves every single time I drive.

This problem is so big that on my infotainment system, the Auto-Stop/Start on/off switch gets a prime location. But that preference doesn’t persist, so each time I get in the car, I have to turn it off.

What both of these experiences have in common is that they could be solved with global user settings. As product people, we develop features based on customer needs and as a market response. Both of these features reek of market response motivation. 

I think of the product folks at Slack and Subaru looking at the usage stats for these features and thinking they’ve got a right hit on their hands. Adoption will be off the charts. But what’s missing is the chance to opt out. User generated feature opt-outs have so much to teach us about what our customers are actually using our product for. 

A policy for data-backed feature enhancement is fantastic, but only when the data you are basing your assumptions on is good. In the case of both of these features, the adoption data is going to be very dirty. 

When rolling out a new feature, don’t only consider enablement. Take the time and have the empathy to give your users a nice way to say, ‘no’ as well.