The SEO Checklist for a Brand-New Page

Pilots use checklists not because flying is intellectually hard but because it’s procedural, and procedures fail in boring ways. Nobody forgets how to fly. They forget to lower the landing gear. Publishing a web page is the same kind of activity. The hard part — writing something worth reading — is done. What remains is a handful of mechanical steps, each trivial, each routinely skipped, and any one of which can quietly sink the page.

I’ll tell you what I check, in the order I check it, and then I’ll tell you why the order matters.

First, the title tag. Not the headline on the page — the <title> element in the head, the thing that becomes the browser tab and, usually, the blue link in search results. It should say what the page is about in plain words a searcher would use, and the specific words should come early, because search results truncate long titles somewhere around sixty characters and because people scan the front of a link, not the end. The most common failure isn’t a bad title. It’s a default one: the CMS filled in the site name, or “Untitled,” or the same title as every other page. A duplicate title tells search engines your pages are interchangeable, which is a strange thing to claim about pages you presumably wrote because they were different.

Second, the meta description. It doesn’t affect ranking — Google has been clear about that for years — but it’s frequently the sentence under your link in the results, which means it affects whether anyone clicks. Think of it as ad copy you get for free. When it’s missing, Google assembles a snippet from whatever page text seems relevant, and the result reads like a sentence cut off mid-thought, because it usually is. Writing one takes ninety seconds: one or two sentences, under about a hundred and sixty characters, saying what the reader will actually get. When GazeSite audits a page, “Missing meta description” is among the most common findings it emits, and it’s the purest example of the whole category — a fix that costs almost nothing, skipped almost universally.

Third, exactly one H1, and it should match the title tag in substance if not in wording. The H1 is the top of the page’s outline. Zero H1s means the outline has no root; five H1s means it has five roots. Neither is fatal, but headings are one of the few explicit signals of structure a page offers, both to search engines and to screen readers, and it costs nothing to get them right. The usual culprit is a logo wrapped in an H1 by an old theme, so every page on the site leads with the company name as its main heading.

Fourth, the canonical URL. This one is subtle enough that people either don’t know about it or fear it. A canonical tag says “this URL is the official address of this content.” It matters because the same page tends to be reachable at several addresses — with and without a trailing slash, with tracking parameters bolted on, over http and https — and search engines treat different URLs as different pages unless told otherwise, splitting whatever credit the page earns across the variants. The safe default is a self-referencing canonical: the page points at its own clean URL. The dangerous failure is a copy-paste error where a template stamps one page’s canonical onto another, which politely asks Google to ignore the new page entirely. It will oblige.

Fifth — and this is the one that isn’t mechanical — intent. Before publishing, I try to write down the actual search query I imagine someone typing before they’d want this page, and then I reread the opening paragraph pretending I just typed it. Does the page answer the question, or does it clear its throat for three paragraphs first? A page can pass every technical check and still fail here, because the checks verify that the page is legible to machines, not that it deserves the click. This is why intent goes last on my list but matters most: the mechanical items decide whether you’re eligible; intent decides whether you win.

Notice what’s not on the list. Keyword density. Meta keywords, which search engines have ignored for two decades. Elaborate schemes for internal link sculpting. The industry around SEO has a financial incentive to make the subject seem vast and arcane, because vast and arcane things require consultants. But for a single new page, the surface area is genuinely small. Title, description, H1, canonical, intent. Five items. Five minutes.

The interesting question is why, if it’s this easy, so few pages pass. I think it’s because none of the failures announce themselves. A page with no meta description looks identical in your browser to one with a perfect description. There’s no error, no broken layout, no complaint from anyone. The cost is paid later, diffusely, in searches where your link appeared and someone chose the other one. Invisible costs are the ones nobody fixes, which is precisely why a checklist earns its keep: it converts an invisible omission into a visible unchecked box. That was much of my motivation for building GazeSite — not to be clever, but to be the copilot who reads the list out loud, every time, because the whole point of a checklist is that it doesn’t trust you to remember. And it’s right not to. I built the tool and I still forget the landing gear.