Ten Seconds of Nothing: The Slow Web Experiment That Made My Readers Snap

Somewhere between my third coffee and my fourth abandoned side project, I listened to a podcast episode about the Slow Web movement. The argument goes like this: the internet trained us to expect everything instantly, and in doing so, it killed the kind of reading that actually sticks. What if content arrived slower? What if you had to choose to wait for it, the way you choose to wait for a kettle to boil?

Most people would nod along and close the tab. I opened my editor and wrote a middleware function that adds a flat 10-second delay to every page load on josephkhan.me.

The Setup, or: How to Sabotage Yourself on Purpose

The implementation was embarrassingly simple. A tiny Express middleware, a sleep function, and a comment that just said // sorry. I deliberately did not add a loading spinner, a countdown timer, or any explanation. If a reader hit the site, they got a blank white page with my header at the top and absolutely nothing else for ten full seconds. Then, all at once, the article would appear.

The rules I set for myself:

• Two weeks, no exceptions. If I chickened out on day three, the whole experiment was worthless.
• No announcements. No blog post explaining the delay, no banner, no apology tweet. The whole point was to observe natural behavior.
• Track everything. Bounce rate, scroll depth, time on page, return visits, and — my personal favorite metric — angry emails.

I deployed it on a Tuesday evening, poured a bourbon, and watched my real-time analytics like it was a horror movie.

The First 48 Hours Were a Bloodbath

My bounce rate, which usually lounges around a comfortable 55%, shot up to 81% overnight. People were landing on my site, staring at a blank page for a few seconds, and leaving. Of course they were. I'd have done the same thing. Probably did do the same thing on a dozen other sites without knowing they were slow-web philosophers.

But here's the interesting part: not everyone bounced. A stubborn minority — roughly one in five visitors — sat through the full ten seconds. And those people? They read differently. My average scroll depth among the patient crowd jumped from 44% to 78%. They weren't skimming. They'd paid a toll to get in, and by god they were going to get their money's worth.

It reminded me of something my old boss at a Chicago agency used to say about trade shows: nobody remembers the brochure they grabbed for free, but everyone remembers the booth that made them wait in line.

The Angry Emails Were My Favorite Part

By day four, my contact form had collected five emails, which is roughly five more than a typical month. They ranged from genuinely confused ("is your site broken? love the writing though") to one beautifully hostile note from a developer in Austin who reverse-engineered my delay, found the sleep call, and emailed me the line number with the subject line "found it. why."

Honestly? That's the most engaged anyone has ever been with my infrastructure.

One email stopped me cold, though. A reader — someone who'd subscribed to my newsletter for over a year — wrote to say she'd been reading my posts on her commute, but ten seconds of dead air on a spotty subway connection meant the page simply never loaded at all. She wasn't angry. She just quietly couldn't read my work anymore.

That's the thing about friction experiments: they never hit the hypothetical power users hardest. They hit the person on the CTA Blue Line with two bars of signal.

The Numbers, Summarized for the Impatient

After fourteen days, I pulled the plug and crunched everything:

• Bounce rate: 55% → 81% during the experiment, back to 57% within two days of removal
• Scroll depth (survivors): up 34 percentage points
• Return visits: down 22% — some people just didn't come back
• Newsletter signups: actually up slightly, which I attribute to the survivors being a self-selected cult of patient readers
• Angry emails: five, of which I'd classify two as fan mail

What I Actually Learned

The slow web, as a philosophy, is lovely. As a user experience, it's a tax — and like all taxes, who pays it matters more than how much it costs. My RSS subscribers never noticed a thing. My direct traffic tolerated it with grumpy grace. My first-time visitors, the ones I claim to care about most, fled in droves.

But I can't say it taught me nothing. Removing the delay felt like taking off heavy boots — everything felt instant in a way I'd stopped noticing. And I now think about every millisecond of my site's performance as a promise I'm making to a stranger.

The delay is gone. The middleware comment // sorry lives on in a branch somewhere, a little monument to the fortnight I made my readers wait for prose they didn't ask to wait for.

Would I do it again? Ask me after my next podcast episode.