On-page SEO
optimization
in San Francisco

Heading structure gets discussed as a ranking factor. That framing produces pages with a keyword in every H2 and no visible logic. The better test is simpler: read the headings alone, in order, and see whether they summarise the page.

They should. A reader who scans the page gets the argument from the headings and stops where their question is answered. That is what most people do on a long page, and it is close to how a search engine selects the passage it will show.

A few rules carry most of the value. One H1, and it states what the page is about rather than what the company is called. Levels in order, with no jumps from H2 to H4. Headings that make sense out of context, because that is how they will be read. No headings used for styling, and no styled paragraphs pretending to be headings.

Length matters less than specificity. Overview tells the reader nothing. What the audit covers tells them whether to keep reading.

Accessibility and search pull in the same direction here. Screen reader users navigate long pages by jumping between headings, so a broken hierarchy leaves them scrolling through content they did not ask for. Fixing the outline fixes both at once.

Check the rendered page, not the design file. Headings are routinely injected by widgets, related-post blocks and consent banners, and a page that looks clean in the editor can ship with three H1 elements. Pull the outline from the live HTML and read it top to bottom. Most of the defects are obvious within ten seconds.

Do this on templates, not on one page. A heading defect in a template is the same defect on every page that uses it. Fix it once and the whole section is clean.

One more check costs nothing. Copy the outline into a document and hand it to somebody who has not seen the page. Ask what the page is about. If they hesitate, the headings are decorative and the fix is editorial.

San Francisco sites get on-page work from us on a page's title, headings, body text, internal links and markup, with every change documented. A standard business site takes 3–4 weeks from audit to implementation. The on-page fixes tier starts from ~$550; full SEO optimization is priced near $1,100.

Readers check who wrote a page before they trust it. They check when, too. Google's quality rater guidelines ask the same questions. Now picture a service page with no date and no named person. Nobody can tell whether its prices are current. Readers and raters alike get little to go on.

Mechanical dates do the most damage. A theme that prints today's date on every page tells the reader nothing. A date bumped without a change to the content misleads. Google's guidance is to show one visible date and repeat it in structured data as datePublished and dateModified. Those should agree with each other and with the lastmod value in the sitemap.

Our rule is plain. The modified date moves when the substance moves: a new section, a corrected figure, a changed price. A fixed typo does not count.

Old dates are no defect in themselves. A guide from years back that is still accurate keeps its original date. Add when it was reviewed. Do not pretend it is new.

Price blocks need the same care. A price with no date invites doubt. One line saying when the prices were last set answers it.

Bylines matter most on pages that give advice. A named author, one line on what that person does, and a page about them let a reader weigh the text. On a service page the equivalent is the person or team responsible for the service, with a way to reach them.

Reviewed-by lines are a related device. They name a second person who checked the text. Use one only when the check happened.

Contact details belong in the same pass. A page selling a service needs a visible route to a human. An email. A form. A phone number, where one exists. Buried three clicks deep, they read as evasive.

About pages carry the rest of the load. They say who runs the company, where it is registered and how to reach it. Short is fine. Vague is not.

Structured data can describe the organisation and the author. It helps only when the visible page says the same thing. Google's structured data guidelines require markup to describe content a visitor can see, so an author named only in the markup is a mismatch.

What we check per template: whether a date is shown, which field in the template prints it, whether it matches the markup and the sitemap, and whether a name and a contact route appear without scrolling to the footer.