what i learned building websites in 2023, the year everything got harder and easier at once

tech · 2024-01-05 · 6 min

2023 was the year i stopped expecting building a website to be the hard part.

that sounds like a small realization. it is not. for years i had been treating the build as the project and everything after the build as cleanup. content, traffic, maintenance, updates, fixes, revisions, the slow work of making a site actually useful — those were second-order problems. the site was the thing. the rest was what happened after.

2023 taught me that the build is often the smallest part of the work, and the easiest part to fake.

what got harder

the first thing that got harder was deciding what a site was for. that sounds obvious in retrospect and was much less obvious while i was doing it. a site can be a brochure, a storefront, a portfolio, a blog, a landing page, a collection of useful pages, a home for a person, a machine for collecting emails, a place to publish notes. any of those is legitimate. none of those is interchangeable. the more i built, the more i noticed that the sites that worked were the ones with a clear job, and the sites that drifted were the ones that tried to be a little bit of everything and ended up being a little bit of nothing.

the second thing that got harder was quality, when quality meant something a visitor could feel. not fancy. not trendy. Just solid. Fast enough, clear enough, trustworthy enough, usable enough that a stranger did not bounce before they understood what was being offered. That bar is higher than it looks from the inside. From the inside, a page usually feels finished sooner than it really is. From the outside, the same page usually fails earlier than its maker expects.

the third thing that got harder was maintenance becoming real. Not the theoretical maintenance of “you should update things.” The practical maintenance of a site that was actually in use, with a few real pages, a few real visitors, a few real breakage points. A site that no one maintains is one thing. A site that needs maintenance and has no owner for it is another. That distinction took me a while to respect.

the fourth thing that got harder was judging success honestly. A site can look like it is doing well because it exists, because it looks decent, because someone shared a link, because analytics shows any number at all. Those signals are weak. The stronger question is boring: did the site do the job it was built to do, for the people it was built for, often enough to matter?

what got easier

the first thing that got easier was the actual build. The tools kept improving. Not in some magical sense. In a practical sense. Faster setup, less glue, more of the boring parts handled by something that did not require me to maintain it by hand. The range between “starts quickly” and “ships something real” got shorter for a lot of kinds of sites.

the second thing that got easier was getting a site live. That is not nothing. The old pattern — hosting wrangling, deployment anxiety, the feeling that launching was itself a technical milestone — had quietly disappeared for a large class of projects. A site could be published without a ceremony. That freed up attention for the parts that actually mattered.

the third thing that got easier was learning from a live site instead of guessing in advance. Once a site was real, even a small one, it started telling the truth. Where traffic came from. Where people left. Which pages did the job and which ones did not. A site in use is a better teacher than a site in planning. That lesson only lands after you have one in use.

the fourth thing that got easier was understanding that a website is not a project with an ending. It is a small owned property that needs tending. That sounds like more work, and in one sense it is. In another sense it is easier, because it replaces the fantasy of a perfect launch with the reality of ongoing care. Ongoing care is manageable. The fantasy of one big finish is not.

the things i unlearned

i unlearned the idea that a great site is mostly a design achievement. Design matters. Design is not the whole job. A visually agreeable site that does not answer the visitor's real question is still failing.

i unlearned the habit of treating content as decoration. Content is the work. A page full of “our solutions,” “our expertise,” and “our commitment” is usually a page that has not done the work of deciding what it is actually saying. Visitors are not impressed by confidence. They are persuaded by specificity.

i unlearned the assumption that traffic is the default measure of success. Traffic is one measure, and sometimes a useful one, but it is not the only one and not always the right one. A site can get very few visitors and still do its job well. A site can get a lot of visitors and still be a poor fit for its owner. The measure has to follow the purpose.

i unlearned the idea that complexity signals seriousness. It often signals fragility. The site that needs a dozen moving parts to do a simple job usually has a simpler job than its builder thinks it does.

what i would tell the 2023 version of myself

start from the job, not from the tool. A site is a tool for a purpose. If the purpose is vague, the site will be vague in expensive ways.

build the smallest thing that can do the job honestly. Not the smallest thing that looks finished. The smallest thing that can actually answer the question it exists to answer.

write for a stranger who is in a hurry and mildly skeptical. That is the audience most sites actually have. Not the audience the owner imagines on a good day. The stranger.

make the mundane parts good. Contact paths, loading behavior, mobile behavior, clarity of the first screen. These are not glamorous, and they are often where the site wins or loses.

treat updates as part of the work, not as a future problem. The future arrives carrying the same problems you ignored.

measure against the job, not against vanity. If the site is supposed to bring in inquiries, did it? If it is supposed to explain a service, did it? If it is supposed to help someone decide, did it? Those are better questions than “does it look impressive?”

and finally, do not confuse finishing the build with finishing the site. The build ends. The site keeps working, or not, and the difference is usually made by the unglamorous stuff that happens after the launch.

the quieter lesson

the quieter lesson of 2023 was that building got easier exactly where it should have, and that this did not automatically make sites better. It just made bad sites cheaper to produce and good sites easier to start. The gap between “exists” and “works” stayed the same.

That gap is where the real work lives.

for the year-lessons companion: my first real product. for the practical post-launch routine that makes a site earn its keep: after publishing.