SEO for small sites: the short list

A site with fifteen pages and an ecommerce catalogue with fifty thousand have different problems. The catalogue is fighting crawl efficiency, faceted URL explosion, duplicate parameter handling, and pagination. The fifteen-page site is fighting the fact that nobody has a reason to link to it and half its pages answer a question no one asked.

The audit template does not know the difference. It runs a crawler, produces two hundred findings sorted by a severity score the tool invented, and hands over a PDF. Most of those findings are real in the sense that the string exists in the HTML. Very few of them are the reason the site does not rank.

The list that applies at that size is shorter and less impressive.

Match the intent, not the phrase

The first thing to check is not keyword usage. It is whether the page matches the kind of result the query already returns.

Search "project management software" and Google returns comparison lists and category pages, not vendor homepages. Search "how to import a CSV into Postgres" and it returns documentation and step-by-step answers, not a product page with a signup form. The result set is Google telling you, explicitly, what format it has decided satisfies that query. A page in the wrong format loses regardless of how many times the phrase appears in it.

How to check: open an incognito window, run the query, and look at the top ten results as shapes rather than competitors. Are they listicles? Docs? Forum threads? Product pages? Then look at your page. If eight of ten results are comparison tables and yours is a sales page, you have a format mismatch, and no amount of on-page tuning fixes it. Either change the page's shape or pick a different query.

Keyword density, meanwhile, is not a ranking input in any modern sense, and Google has said as much for years. Write the phrase in the title, the H1, and the first paragraph because that is where a reader looks for confirmation they are in the right place. Then stop counting.

Stop making a page for every variation

A common recommendation is to build one page per keyword variant: one for "web design Sofia", one for "web design agency Sofia", one for "website design in Sofia". Three pages, three hundred words each, near-identical content.

This is worse than having one page, for two reasons. The pages compete with each other for the same query, so the search engine has to guess which one you meant and often picks the weakest. And any links or engagement you earn get divided across three shallow pages instead of accumulating on one good one.

How to check: run site:yourdomain.com "your target phrase" and see how many of your own pages come back. Better, open Search Console, go to the Performance report, filter to a single query, and add the Pages dimension. If one query is being served by three different URLs across a month of data, those pages should be one page with a redirect from the other two.

Internal linking, the free lever

Internal links are the most underrated thing on a small site, mostly because nobody sells a retainer for "add eleven links".

They do two things. They tell a crawler which pages you consider important, since a page linked from the navigation and from four body paragraphs is obviously more central than one reachable only from the sitemap. And they let anchor text describe a page in the words a reader would use, which is a signal you fully control.

On a small site the fix is usually mechanical. Check whether blog posts link out to anything at all, and whether service pages connect to each other only through the top navigation. Often the pages you most want to rank are, in link terms, the loneliest on the domain.

How to check: crawl your own site with any free crawler and sort pages by inbound internal links. If your primary service page has two and your privacy policy has fifteen, the site's internal structure is arguing against you. Aim for every commercially important page to be reachable in one or two clicks from the homepage, with descriptive anchor text rather than "learn more".

The technical checks that matter at this size

Four, and they are checkable in an afternoon.

Indexability. Pages that should rank must be crawlable and indexable. Check robots.txt for over-broad Disallow rules, and view the HTML source for noindex meta tags left over from staging. The Search Console URL Inspection tool is the authority here. A crawler's opinion is a guess; Search Console reports what Google actually did.

Canonical correctness. Every page should declare a canonical URL, and on a small site it should almost always point at itself. The classic bug is a template that hardcodes the homepage canonical across every page, which tells the search engine that your entire site is one document.

Sitemap accuracy. The sitemap should list exactly the URLs you want indexed: no redirects, no 404s, no noindex pages. Submit it in Search Console and compare the submitted count with the indexed count. A large gap is a real finding. A sitemap listing pages that no longer exist is a credibility problem for the whole file.

Core Web Vitals. Google publishes thresholds: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1. Measure with PageSpeed Insights, and prefer the field data section over the lab score when it is available, because it reflects real visitors on real devices.

And the ones that are noise at fifteen pages

Crawl budget. Google's own documentation says crawl budget is not a concern for sites below a few thousand URLs, and yours is nowhere near it. Log file analysis for crawl patterns, similarly. Faceted navigation rules, parameter handling, and pagination markup solve problems a small site does not have. Long lists of missing alt attributes on decorative images are an accessibility matter worth fixing on its own terms, not a ranking lever.

"Four blog posts a month" without a thesis

Content retainers are usually sold as volume. Four posts a month, indefinitely. What tends to get produced is four unrelated articles on whatever terms a keyword tool ranked as low difficulty, which produces a blog with no centre.

A topic thesis is a claim about a domain that the site is trying to become the reference for, and it makes each post a component rather than an isolated attempt. It also gives internal linking something to link. Twelve posts that circle one subject and reference each other read as expertise. Twelve posts on twelve subjects read as filler, and Google's stated preference for content written primarily for people rather than for search engines is aimed squarely at the second pattern.

How to check: write your thesis in one sentence, then list your last ten posts underneath it. Any post that does not support the sentence was written for a keyword tool, not a reader.

---

If you want a second opinion on which of these applies to your site, we do a free audit and tell you the two or three things worth changing, not two hundred.