Brayden Wisniewski

Slack · UX Research

Making Slack’s hidden state visible

As research lead on a team of three, I studied where everyday Slack tasks break down. Two of them failed for most people, and two small changes to language and system state took success from 3/8 and 2/8 to 8/8.

Role
Research Lead
Team
3 researchers
Timeline
10 weeks, 2026
Methods
Heuristic review, moderated & unmoderated testing

Overview

Two everyday tasks were quietly failing.

We reviewed 20 Slack workflows against Nielsen’s heuristics, then watched 8 people try the everyday ones. Getting around was fine. Two tasks fell apart, and most people thought they’d finished.

3/8
found a saved message
2/8
managed their notifications

Saved messages

The concept wasn’t the problem. The language was.

Saved, Pinned, and Later all sounded alike, and none said who could see what. We renamed the actions around ownership: Save for Me, Pin for Everyone.

“There is usually a saved button here. This is terribly confusing.”Participant
design
SC

Sarah Chen9:41 AM

Final specs for the onboarding flow are in Figma. Can everyone review before Thursday?

MW

Marcus Williams9:44 AM

On it, I’ll leave comments by end of day.

Saved

No Saved for Me messages yetClick the bookmark on the message to add it here

Try it: save the message, or pin it for the channel.

3/88/8in validation testing

Notifications

Showing people what Slack will notify them about.

Notification controls were spread across menus, with no sign of what was already on. We added a persistent status and put every control in one place.

“I have no idea where to even start with this.”Participant
design Mentions
JK

Jordan Kim10:02 AM

@Brayden can you take a look at the new nav before standup?

You’ll only hear from #design when someone mentions you.

System NotificationsControl how and when you’re notified

0 on All 0 on Mentions 0 muted

By ChannelChoose which messages trigger alerts for each channel.

#general
#engineering
#design
#random

Try it: change a channel’s alerts and watch the status and overview update.

2/88/8among people who finished the task

What I took away

Most of the fix was wording and visibility, not new features.

We validated both changes with 8 new people on UserTesting, with no researcher in the room. The same failures didn’t come back.

View more