Before I write a line of code, I decide where enquiries go and how the site will get found. I skip heavy CMS platforms for speed and lower running cost, wire every form straight to WhatsApp or a CRM so leads get answered while the person is still interested, and plan pages around the specific questions customers already ask.
When someone tells me they want a website for their business, I don't ask what it should look like. I ask what happens the moment someone fills in a form on it. Most people planning a new site start with colours, logos and which template to pick. I start with the form, because that decision affects everything that comes after it.
Where does the enquiry actually go
I build landing pages whose forms send each lead straight to WhatsApp or a CRM. That sounds like a small technical detail. It isn't. A form that lands in an inbox nobody checks is a form that loses leads, no matter how good the page looks above it. When the enquiry arrives where the team already works, someone answers it while the visitor is still thinking about the business, still has the tab open, still remembers what they were looking for.
A beautiful site with a contact form that emails a shared inbox checked once a week is not a lead-generation tool. It's a brochure. Before I spend time on anything a visitor sees, I want to know where a form submission lands, who owns responding to it, and how fast that happens. Everything else on the site is decoration around that one decision.
The CMS choice comes second, not first
People planning a site often assume a dashboard-driven CMS is the default, something to pick before anything else. I've stopped treating it that way. I build small-business sites without a heavy CMS for two practical reasons, not out of any ideology: speed and running cost. A lean site loads faster, which helps visitors and helps search ranking. It also has less to host and fewer things to keep paying for every month.
The trade is real. Editing a site like this needs a developer, or an AI that can edit the code directly, rather than someone logging into a dashboard and changing a headline themselves. For a business whose site changes a handful of times a year, an opening hours update here, a new service listed there, that trade works in the owner's favour. A business that publishes new content every week, or has non-technical staff who need to edit pages constantly, is a different case. The mistake I see is choosing the familiar, heavier option by default, without asking how often the site will actually change.
What the pages should actually be about
For BabyJunctions, what moved search traffic wasn't one clever trick. It was long-tail articles that each answered one specific question a parent actually asks, paired with structured data telling search engines what each page was about. I treated SEO and AEO as one job rather than two, and that mattered more than I expected. The same article that ranks for a narrow question is the kind of page AI answer tools like to quote, and BabyJunctions has been cited in AI answers as a result.
That changes how I think about a site's pages before a single one gets written. Instead of one page called Services, I want pages that each answer a question a real customer would actually type, worded the way they'd type it. Say a plumbing business is planning its site: rather than one general services page, it makes more sense to have a page answering how much notice an emergency callout needs, another answering whether someone should check boiler pressure before calling anyone out, and so on. Each page answers one thing completely, and each one is something search engines and AI tools can point to on its own.
Where does a form submission go, and who is responsible for answering it. How often will this site actually need to change, and does that justify a heavier CMS. Are the pages built around real questions customers ask, worded the way they ask them.
None of this is complicated, which is probably why it gets skipped. It's easier to argue over a homepage banner than to decide who answers a WhatsApp lead on a Friday evening. But the banner doesn't lose the customer. The unanswered form does.
