Why the moderation toggle is deliberately annoying

There's a toggle in the Wayground content moderation tool that turns on keyword blocking for an entire district. Flip it, and the keywords an admin has entered start hiding content from every teacher and student under them. We made that toggle harder to use than it needed to be. On purpose.

That's an odd thing to admit. Most of the time my job is removing friction — fewer taps, clearer copy, get out of the user's way. Here I put friction back in and defended it in review. This is why.

What the toggle actually does

Step back to the stakes. This tool exists because conservative districts were blocking us at the firewall over "Parents Bill of Rights" laws, and Sales wanted to hand admins a kill switch for the whole content library. I fought that and we built granular keyword blocking instead — admins hide specific content, teachers can still see what's hidden and request access. It preserves most of the library's value while giving admins real control.

But "real control" is the scary part. When an admin turns keyword blocking on, they are reaching down and pulling resources out of every teacher's hands across the district. Some of those teachers will land on a blocked lesson mid-plan and have no idea why. This is not a preference toggle like turning on dark mode. It's an action with a blast radius measured in classrooms.

The failure mode I was designing against wasn't the malicious admin. It was the casual one. The admin who's clicking through settings, sees a moderation toggle, flips it to see what it does, and moves on — never registering that they just changed what thousands of teachers and students can access. If the control feels as light as any other setting, people will treat it as lightly as any other setting.

Friction as a signal, not an obstacle

So we made turning it on feel like a decision rather than a flick. The toggle is a gate, not a switch you brush past — enabling district-wide blocking has to read as a deliberate act, not something you bump on your way through the settings page. The intent was that an admin can't turn it on without, for a beat, being made to understand what they're about to do.

That weight is doing a specific job: it's converting an idle click into a considered one. The extra deliberation is a place where the admin either confirms "yes, I actually want to hide content from my whole district" or backs out because that wasn't what they meant. The people who genuinely need the feature — the ones under political pressure, who came to this screen on purpose — walk through it without a second thought. It costs them a few seconds. The people it stops are the ones who would have flipped it by accident and caused a mess they didn't intend and couldn't explain.

That's the trade I was making. Frictionless design optimizes for the median user doing the common thing. But the common thing here — idly toggling — is the thing I most wanted to prevent. When the cheap accident is expensive for other people, friction stops being a tax and starts being a guardrail.

Friction is a material

I've come to think about friction the way I think about color or spacing — a material you place deliberately, not just a cost you minimize. You spend it where the consequences are high and you strip it out everywhere else. The moderation toggle got friction precisely because it's one of the few controls in the product where a careless click harms people who never touched it.

You can see the same instinct in how we handled keywords. Admins told us they'd paste in 1,000+ keywords if we let them upload a CSV — and that volume would have degraded our search and recommendations, because the system has to account for every blocked term. So we didn't build CSV upload. We made them copy-paste or type keywords in, which is more annoying and caps how much damage an over-eager admin can do to search quality. Same philosophy as the toggle: the easy path was the dangerous one, so we didn't pave it. (That decision has its own essay.)

The weighty toggle also paid off downstream. Because turning on blocking was a deliberate act, an admin who did it had, at minimum, been shown what they were signing up for. That made the whole feature more defensible — when a teacher asked why their content vanished, the answer traced back to a decision an admin made on purpose, not a switch someone bumped.

What I took from it

Removing friction is the default, not the goal. The goal is matching the weight of the interaction to the weight of the consequence. Most of the time those point the same way — low stakes, get out of the way. But for the handful of controls where one person's careless click lands on everyone downstream, adding friction is the user-centered move. You're protecting the people who aren't in the room.

The tell that you're looking at one of those controls: ask who pays if this gets triggered by accident. If the answer is "the person who clicked it," minimize the friction — let them undo and move on. If the answer is "a thousand people who've never seen this screen," slow the click down. The moderation toggle was firmly the second kind, so I made it annoying, and I'd make it annoying again.

This decision comes from the Wayground content moderation work.