EST. 1998  ///  BUILD YOUR OWN CORNER OF THE WEB  ///  HAND-TESTED TOOLS, NO HYPE  ///  GEOCITIES IS GONE. YOUR OWN SITE DOESN'T HAVE TO BE.  ///  EST. 1998  ///  BUILD YOUR OWN CORNER OF THE WEB  ///  HAND-TESTED TOOLS, NO HYPE  ///
EST. 1998

Field notes

The useful website is making a comeback

For fifteen years the winning move on the web was volume: more posts, more keywords, more thin pages wrapped around the same recycled answer. AI ended that, twice. Once by writing the filler faster than any content farm could, and once by answering the generic questions directly, so the traffic those pages lived on never arrives. Everyone covering the web calls this the death of the website. I think it's the best thing to happen to websites since 1998.

What dies and what survives

What dies is the site that existed to intercept a question on its way to an answer, the "what time zone is Chicago in" page with four ads and a newsletter popup. No one will miss it. What survives, and increasingly what gets cited by the AI assistants themselves, is the site that does a job: the calculator that embarrasses the pricing pages, the reference table someone bookmarks, the review written by a person who actually paid for the thing, the niche calendar kept accurate by someone who cares. Usefulness was always the moat. The web just spent fifteen years pretending otherwise because the arbitrage paid better.

The old web knew this

The sites I built in 1999 were useful or they were nothing; there was no algorithm to game yet. A fan site was the place with every episode airdate. A coupon site saved you actual money or you never came back. That spirit, one person building the genuinely handy thing about what they know, is precisely what works again, except now the building takes a weekend instead of a winter, because the tools got absurdly good.

The test for your own site

Before building a page, I ask: would anyone bookmark this? Not read it, bookmark it. A blog post rarely passes. A true-cost calculator passes. A "best builder for musicians" page written by someone who runs music sites passes. If nothing on your planned site passes the bookmark test, you don't have a site yet; you have a brochure, and the AI assistants already wrote a better one.

Build the tool, not the takes

Practical version: whatever you know professionally or obsessively, there's a small tool or dead-honest reference page inside it that doesn't exist yet, because the people with your knowledge don't build websites and the people building websites don't have your knowledge. That gap is the whole opportunity. It costs a domain name and a weekend to find out; here's the weekend, hour by hour.

The web changed. The need for good judgment didn't.

Build something useful
How we make money: some links on WebMogul are affiliate links. If you sign up for a tool through one, the company pays us a commission at no extra cost to you. We only recommend tools we have actually used to build real sites, and we say so when one falls short. More on how this site works.