A roofing contractor I worked with was spending several thousand dollars a month on shared leads while his own website sat on page four. Six months later he was first organically across his service area and had cancelled the lead platforms entirely. Nothing exotic happened. We did the ordinary things in the right order.

The order matters more than the tactics

Most SEO advice is a list of tactics with no sequence attached, which is why so much of it fails. The tactics are largely correct. Applied in the wrong order they do nothing.

You cannot link build a page into relevance if the page does not exist. You cannot rank a page that a crawler cannot reach. You cannot fix rankings with content when the actual problem is a six second mobile load time. The sequence below is the one I follow on every project, and the early steps are the boring ones.

Step one: understand what the query actually wants

Before anything else, look at what is currently ranking for the term you want. Not to copy it, but to find out what Google has already decided the query means.

If the first page is all product pages, Google has classified it as transactional and your blog post will never rank there no matter how good it is. If the first page is all guides, a product page will not rank either. This mismatch is the single most common reason a page that is objectively better than its competitors fails to move.

Search your target term and categorise the top ten. Are they buying pages, comparison pages, definitions, or local listings? Then build the type that is already winning, and make it better. Fighting the classification is a fight you lose.

Step two: make the site legible

This is the least interesting part and the part that most often turns out to be the actual problem. The checklist:

  • Every page you want ranked is reachable by following links from the homepage in three clicks or fewer.
  • Content is present in the HTML source, not injected by JavaScript afterwards. Google can render JavaScript, but rendering is a second pass with a queue in front of it, and other systems are far worse at it.
  • One H1 per page that describes that page, with a real heading hierarchy underneath. Missing H1s are startlingly common on modern template sites.
  • Load time under about two and a half seconds on a mid range phone on a normal connection. Test on a real device, not a desktop emulator.
  • Images in modern formats, sized correctly, lazy loaded below the fold, with alt text that describes the image.
  • No redirect chains, no orphan pages, one canonical version of every URL, and a sitemap that lists everything you want indexed and nothing you do not.

On the roofing project this stage alone took PageSpeed from 41 to 94. It was not the whole reason for the ranking, but nothing else would have worked without it.

Step three: one page per thing

The most common structural failure I find is a business with six services and one page listing all six. That page competes for six different queries and wins none of them, because for any given search it is two thirds irrelevant.

One page per service. One page per location where location genuinely changes the offer. One page per meaningful question. Each with its own title, its own H1 and its own reason to exist.

The trap on the other side is generating hundreds of near identical pages with a place name swapped in. Google has recognized and discounted that pattern for years. A location page earns its place when it says something true about that location: the actual neighborhoods, the jobs done there, the conditions that differ. If you cannot write two genuinely local paragraphs, do not make the page.

Step four: be the better answer, not the longer one

Word count is not a ranking factor. It correlates with rankings because thorough answers tend to be longer, and people confuse the correlation for the cause and then pad.

What actually works: answer the question in the first sixty words, cover the sub questions the top ranking pages missed, use specifics rather than generalities, and stop when you are finished. A 900 word page that fully answers a question beats a 2,500 word page that answers it in paragraph nine and pads the rest.

Include the things your competitors leave out because they are awkward. Prices, timelines, what you do not do, when someone should choose a different provider. Those sections get read, get quoted and build the trust that turns a visit into an enquiry.

Ignore everything about outreach at scale. As a small business, the links available to you are these, and they are enough for most local and mid competition terms:

  • Industry associations and trade bodies you belong to, most of which list members and most of which you have never checked.
  • Suppliers and manufacturers whose products you install, many of whom run installer or partner directories.
  • Local business directories with actual editorial standards, chambers of commerce, and business improvement districts.
  • Local press for anything genuinely newsworthy: a hire, an expansion, sponsorship, a community project.
  • Genuine guest articles in trade publications, which are far easier to place than people assume because trade publications are perpetually short of contributors who know the subject.

Do not buy links. It works until it does not, and the recovery costs more than the rankings ever earned. I have declined work over this and would again.

What the timeline honestly looks like

  • Weeks 1 to 4. Audit, technical fixes, page architecture. Rankings usually do not move. This is normal and it is where most clients get nervous.
  • Weeks 4 to 12. New pages published and indexed. First movement, typically on long tail and low competition terms.
  • Months 3 to 6. Main service terms start moving. Local terms can reach page one in this window. This is where the roofing project landed.
  • Months 6 to 12. Competitive national terms become realistic, assuming the credibility work has been running alongside.

Add roughly three months to all of this for a brand new domain. Anyone giving you a fixed date on a competitive keyword is guessing, and the guess is a reason to be careful with them.

What to ignore

Meta keywords, which have been ignored since 2009. Keyword density targets, which were never real. Submitting your site to hundreds of directories, which is a footprint rather than a strategy. Any tool score presented as a ranking factor. Domain authority is a third party estimate, not something Google calculates, and optimizing for it is optimizing for a metric that has no seat at the table.

Spend the time on the five steps above instead. They are slower to sell and they are what actually moves the ranking.