96.55% of all content published online receives zero organic traffic from Google. Not because Google is unpredictable, but because the vast majority of pages are created without any structured process. Content is published, people wait, and when traffic does not arrive, the conclusion is that “SEO does not work.”
SEO works. What does not work is doing it backwards: optimising without auditing first, publishing without keyword research, building links on a broken architecture. This guide is a 7-step operational process you can apply to any project, in the order in which the steps should be executed, with the minimum tools needed at each stage.
The difference between this guide and learning SEO from scratch is one of angle: that resource explains how to acquire SEO skills; this one explains how to execute a real SEO project. They are complementary — but if you already have the theoretical foundation and want to know what to do on your site this week, you are in the right place.
Step 0 — Starting-Point Audit: Know Where You Are Before Optimising
A proper SEO process should start with step 0: measuring the current state. Most people skip it.
The initial audit is not a luxury or a step reserved for large organisations. It is the diagnostic that tells you what problems are holding your site back before you invest time in optimisations that may be entirely irrelevant. Optimising product page titles when you have 3,000 pages blocked in robots.txt is like repainting the façade of a structurally compromised building.
The minimum stack for an initial audit consists of three free tools: Google Search Console, PageSpeed Insights and Screaming Frog (the free version crawls up to 500 URLs, which covers most projects at this stage).
In Google Search Console, the first report to examine is the Indexing report (previously called Coverage). Look at the ratio between valid indexed pages and excluded pages. If more than 20% of your URLs sit in any exclusion category, you have indexation problems to resolve before doing anything else.
Document your findings in a spreadsheet with four columns: problem identified, affected URLs, estimated impact (high/medium/low) and specific action. That prioritised list becomes your work plan for the coming weeks.
According to the Semrush SEO Checklist 2025, the most common problems found on sites that have “never been audited” are: duplicate meta descriptions (present in 47% of sites), pages with load times exceeding 3 seconds (41%), and unmanaged 4xx crawl errors (35%). All three have direct, fast solutions once identified.
Step 1 — Keyword Research: Finding Your Customer’s Vocabulary
Keyword research is not about finding the terms with the highest monthly search volume. It is almost an ethnographic exercise: what exact words does your potential customer use when they have the problem you solve?
Think of it as a needs translator. Your customer does not search “integrated enterprise document management solutions.” They search “how to organise invoices in the cloud” or “software for archiving contracts.” SEO means speaking the language of real search queries, not the language of your product catalogue.
The practical process has four phases. First, list the main topics of your business: the 5–10 thematic areas you cover. Second, for each area, generate search variants using Google Suggest (the autocomplete suggestions appearing in the search bar), the related searches at the bottom of results, and the questions in the “People also ask” box. Third, validate volume and competition using a tool — Ahrefs or Semrush for complete data, or Search Console’s Performance report to see which terms your site is already appearing for. Fourth, group keywords by search intent: informational (they want to learn), navigational (they are looking for a specific brand or site), commercial (they are comparing options) and transactional (ready to buy or hire).
A keyword is not just a term; it is an architectural decision. Each group of keywords sharing the same intent becomes one page. Two keywords with different intents, even if they appear similar, require separate pages. Confusing this produces keyword cannibalisation: two pages competing against each other for the same terms, splitting authority instead of concentrating it.
For a complete reference of SEO vocabulary you will need throughout this process, the SEO terms dictionary saves you time searching for definitions at every step.
Step 2 — Site Architecture: URL Structures Google Understands
Site architecture is the map you ask Google to follow. A poorly designed architecture means important pages receive insufficient crawl budget, authority is diluted across irrelevant URLs, and users abandon because they cannot find what they are looking for.
The guiding principle is hierarchical simplicity: every URL should exist at the shallowest depth possible from the root. The standard followed by well-architected projects is that any important page should be reachable within no more than three clicks from the homepage.
URL structure follows three practical rules. First: semantic, descriptive URLs — not parameter strings or numeric IDs. /services/seo-consultancy/ rather than /page?id=47. Second: hierarchy reflected in the URL where appropriate. /blog/technical-seo/core-web-vitals/ signals to Google that this page belongs to a sub-topic within a broader topic. Third: trailing slash consistency — choose between /page/ and /page and apply the same convention throughout the site, with 301 redirects from whichever pattern you do not use to the one you do.
Internal linking is the operational layer of architecture: the most important pages on your site should receive more internal links than less important ones. Not artificially, but because that linking reflects the real hierarchy of your content. Pillar pages — those covering the broadest thematic scope — should link to the spoke pages of each sub-topic, and spokes should link back to the pillar. This topic-cluster model does not only organise content; it concentrates thematic authority where it matters most.
One issue that appears consistently in audits: orphan pages — pages with no internal links pointing to them. Google can find them if they are in the sitemap, but without internal links it assigns them lower importance. Screaming Frog can generate an orphan page report in minutes.
Step 3 — On-Page Optimisation: Title, H1 and Meta Description
On-page optimisation is the work of aligning every element of a page with the search intent of the user who will find it. It is not “inserting keywords into text.” It is structuring and communicating content in the way that delivers the most value to the user and signals most clearly to Google what the page is about.
The five elements with the most direct ranking impact are, in order of importance: title tag, H1, heading structure (H2–H6), body content and meta description. The title tag — what appears in the browser tab and as the blue headline in search results — is the strongest signal you send Google about a page’s topic. It must include the primary keyword, be unique across the whole site, and run between 50 and 60 characters to avoid being truncated in results.
The H1 is the visible headline on the page. It should contain the primary keyword or a close semantic variant, and there should be only one H1 per page. H2s structure the main sections; H3s and H4s articulate sub-points within each section. This hierarchy is not just for Google: it is what allows a reader to scan the page and decide whether it is worth reading in depth.
The meta description is not a direct ranking factor, but it does influence the CTR (click-through rate) of your result in Google. A well-written meta description — including the keyword and clearly communicating what the user will find — can increase CTR by 5 to 15%, which in turn does affect rankings. Recommended length: 150 to 160 characters.
Tom Parsons, head of SEO at a London digital agency, documented in 2024 how rewriting the titles and H1s of 92 service pages for a B2B software client (without touching the body content) produced a 27% increase in organic clicks over 10 weeks, moving from 1,840 to 2,337 weekly clicks. The site had solid content; it simply was not communicating what each page covered with sufficient precision.
The body content itself must fully address the search intent. If someone searches “how to write a meta description,” they want step-by-step instructions, not an academic definition. If they search “what is on-page SEO,” they want an explanation with concrete examples. The right length is not the maximum possible, but the length that covers the intent without padding — it might be 500 words or 3,000, depending on topic complexity.
Step 4 — Basic Technical SEO: Speed, Crawling and Indexation Errors
Technical SEO is the category of optimisations affecting how Google can access, crawl, understand and index your site. Unlike on-page work, its effects are not always visible directly in content — but the consequences of getting it wrong are: pages that do not appear in Google, rankings that do not improve despite good content, or unexplained traffic drops.
The four most common technical problems, and their direct solutions:
Crawl and indexation errors. In Google Search Console, the Indexing report shows exactly which pages Google has found but decided not to index, and why. The most frequent reasons are: “Duplicate, submitted URL not selected as canonical” (you have duplicate content and Google chooses a different URL than the one you intend), “Crawled — currently not indexed” (Google has crawled the page but considers it insufficiently valuable to index) and “Blocked by robots.txt” (you have accidentally blocked important pages). Each category has a specific remediation.
Speed and Core Web Vitals. Core Web Vitals are the user experience metrics Google has used as a ranking factor since 2021. The three metrics are LCP (Largest Contentful Paint: how long it takes for the largest visual element to appear), INP (Interaction to Next Paint: response speed to user interactions) and CLS (Cumulative Layout Shift: how much content shifts while loading). PageSpeed Insights provides the diagnosis with real user data for any URL. For deeper coverage, the technical SEO guide covers each aspect with the detail required for diagnosis and implementation.
Canonicalisation problems. Canonical tags tell Google which URL is the preferred version when multiple variants exist (with and without trailing slash, with and without parameters, HTTP and HTTPS versions). A misconfigured canonical can cause Google to index the wrong URL or split authority across multiple versions of the same page.
XML sitemap. The sitemap is the list of URLs you want Google to index. It should include only the pages you want indexed (not pagination pages, filter parameter URLs, admin pages, etc.) and be referenced in the robots.txt file. A well-maintained sitemap accelerates the indexation of new content.
According to the Google Search Central SEO Starter Guide, the technical fundamentals — accessibility, speed, unblocked content, correct canonicals — are the prerequisite for any other SEO optimisation to work. Without them, content and link-building work never reach their potential.
Step 5 — Initial Link Building: The First Authority Signals
External links pointing to your site are votes of trust. From Google’s original PageRank to current algorithms, backlinks remain one of the most powerful ranking signals. But there is a critical distinction between building genuine authority and manipulating rankings with links: Google is highly efficient at detecting unnatural link patterns, and manual or algorithmic penalties for link spam are among the hardest to recover from.
The starting point is not to go looking for links, but to create content that deserves to be linked to. Without that, any link-building effort is a distribution system with nothing worth distributing.
Three initial link-building strategies with the best effort-to-result ratio for projects starting out:
Unlinked brand mentions. Use Google Alerts or Semrush’s brand monitoring to track mentions of your brand or business on other sites. When you find an unlinked mention, contact the author and politely request they convert it to a link. Conversion rates are high because the content already exists — only the link is missing.
Quality directories and listings. Industry directories, supplier guides, company listings in specialist media, and Google Business Profile entries are legitimate, relevant link sources. Not all directories have value; the ones that matter are those specific to your industry or those with high general authority (chambers of commerce, professional associations, trusted regional directories).
Linkable assets. Creating resources that others cite naturally: original data from your sector, comprehensive guides, free tools, case studies with concrete results. This kind of content generates what the industry calls “link bait” legitimately — content so useful that people link to it without being asked.
Brian Dean (Backlinko) documented the skyscraper technique: identify the most-linked content in your sector on a specific topic, create a substantially better version, and reach out to those linking to the original. In his original case study, this technique generated a 110% increase in organic traffic in two weeks for an SEO tools list post. The technique remains valid, though the market has become saturated in some niches.
What definitely does not work: bulk link purchasing, mass link exchange networks, and links in blog comments or forums with no genuine context. These practices may produce short-term results, but the exposure to manual or algorithmic penalty risk does not justify the gain.
Step 6 — Monthly Measurement: Search Console, GA4 and Adjustment
SEO without measurement is decoration. The difference between an SEO project that grows sustainably and one that stalls is rarely the quality of the initial optimisations — it is the consistency of the measurement-and-adjustment cycle.
A well-designed monthly measurement cycle answers four questions: which pages are gaining positions? Which pages are losing positions? Which queries are generating impressions but few clicks (optimisation opportunity)? What new technical errors have appeared?
In Google Search Console, the priority sections are: Performance (filter by the last 28 days and compare with the previous period to detect trends), Indexing (new coverage errors or pages that have stopped being indexed) and Core Web Vitals (changes in the percentage of URLs with “poor” performance).
The most useful filter in the Performance section is to order queries by impressions and filter by average position between 8 and 20. Those are the pages sitting on the second page of results with the highest improvement potential from additional on-page optimisations. A page at position 12 that moves to position 6 multiplies its clicks by 3 to 4 times.
In Google Analytics 4, the metrics most relevant to SEO are: organic sessions (absolute trend), percentage of new sessions (reflects whether your content is reaching users who did not previously know you), and the engagement rate of organic entry pages. A page with a high bounce rate and low engagement rate signals that the content is not satisfying the search intent of the users arriving from Google.
The most critical part of the measurement cycle is not collecting the data but deciding the concrete actions that follow from it. Each monthly review should conclude with a list of no more than five prioritised actions: specific content updates, identified technical fixes, new pages to create based on queries generating impressions but without a dedicated page.
To master Google Search Console as the central tool of this cycle, the complete Google Search Console guide covers every section with practical examples and the most common use cases from real projects.
SEO is not a sprint. It is the accumulation of correct decisions executed in the right order and reviewed consistently. The 7 steps in this guide are not a recipe to be executed once: they are an iterative process where each monthly cycle feeds the next.
The difference between sites that achieve sustained organic traffic and those that remain in the 96.55% with zero traffic almost always comes down to two things: having done the initial audit before optimising, and maintaining the monthly measurement cycle instead of abandoning it after the first month.
Start today with step 0: open Google Search Console, go to the Indexing report, and document the first three exclusion categories you find. That — and only that — is the real first step of an SEO process that works.
Resources in this cluster to keep progressing:
- Learn SEO from scratch — the complete learning roadmap
- SEO terms dictionary — the reference glossary for understanding every metric
- What is technical SEO — an in-depth look at the technical pillar
- Google Search Console: complete guide — the central tool of the measurement cycle
- Free online SEO course — structured options to complement hands-on practice