1. We priced by the hour, and it bought us the wrong clients
Hourly billing feels fair until you notice what it rewards: more hours. It puts the studio and the client on opposite sides of the clock instead of the same side of the outcome. Every project is scoped and priced up front now. The number a client signs is the number they pay, and the incentive lines up with theirs: finish it well, finish it properly.
2. We built what was asked for, not what was needed
Early on, a brief was a checklist. If a client asked for five pages and a contact form, that's what they got, built well, delivered on time, and quietly useless. Nobody had asked what the site was actually supposed to do for the business, so nothing forced a decision about it. Now the first conversation is about the one action the site has to drive. Everything else gets built around that, or it doesn't get built.
3. We treated speed as a launch-day number
A site can pass every performance check on the day it ships and be sluggish within a year: a marketing script added here, an embed there, an image nobody optimized. Speed isn't a box to tick once, it's a number that moves in one direction unless somebody's watching it. That's the whole case for staying on after launch rather than handing over a finished file and disappearing.
4. We waited to be found
For a long time, "SEO" meant a few meta tags and hoping. That was already thin advice for search engines. It's worse advice now that a real share of buying decisions get made inside an AI answer that never shows a results page at all. Structuring a site to be legible to both, cleanly, with the right schema and a straight answer where one is called for, isn't an add-on anymore. It's built in from the first day of every project, which is the subject of the next set of notes.
None of these were dramatic failures. They were quiet ones, the kind that cost you a client eighteen months later and you never quite know why.
What stayed the same
The part that never changed: the people who scope a project are the people who build it. That's not a mistake we made and fixed, it's the one thing we got right from the start, and the reason the other four were fixable at all.