The complete guide · SEO

SEO that compounds: the complete guide

Everything we actually do to grow organic search, written out in full and kept current for how search works in 2026. How ranking and AI Overviews work, the technical foundation, structured data, Core Web Vitals, content and topical authority, link building and digital PR, measurement, and the engagement that ties it together. No fluff, no secrets held back.

A working reference, not a sales brochure. When you want it done, start with a free audit.

This guide is long on purpose. SEO is sold as a mystery so it can be billed as one, and most of what passes for strategy is either too vague to act on or too tactical to matter. What follows is the opposite: the actual mechanics, the order we work them in, the diagrams we use to explain them, and the numbers we hold ourselves to. Read it end to end and you will understand our engagement before we ever speak.

How search actually works

Three things have to happen before a page can earn a click. Google has to crawl it (a bot fetches the page), index it (the content is parsed and stored), and rank it (the page is selected and ordered for a given query). Most SEO problems are a failure at one of those three stages, and the fix is different at each one, so the first job is always to find out which stage is failing.

The path from a published page to a click
CRAWL INDEX RANK CLICK Blocked by robots,orphan pages Thin content, straynoindex, bad canonical Weak relevanceor authority Weak title ormissing rich result
A page that is not crawled cannot rank. A page that is crawled but excluded from the index is invisible. A page that is indexed but ranks low is a relevance or authority problem. A page that ranks high but earns no clicks is a snippet problem. Diagnosis comes before treatment, every time.

Ranking itself is decided by hundreds of signals, but they collapse into a handful you can actually influence: how relevant the page is to the query, how authoritative the site is, how good the experience is, whether the page demonstrates real expertise (what Google calls E-E-A-T), and how well it matches what the searcher wants to do. The rest is noise you cannot move, and chasing it is how budgets get wasted.

The two levers that move rankings

Strip away the jargon and there are two levers. The first is earned authority: whether other credible sites reference you, because that is still how Google reads trust. The second is the technical and content foundation: whether your own site is built so the trust you earn can turn into traffic.

Authority decides whether you can rank. The foundation decides whether ranking becomes clicks. Pull one lever without the other and you leave most of the value on the table. A site with great links and broken titles ranks and gets no clicks. A flawless site with no authority never ranks high enough to matter. The whole engagement is about working both at once, in the right order.

The one line version. Earn authority through link building and digital PR. Build the foundation through technical SEO and content. Measure everything in Search Console. Repeat monthly until the curve bends.

The free audit and what it measures

Every engagement opens with an audit, and we give it away because it is the only honest way to scope the work. We pull your real data and answer a specific question: where is recoverable traffic leaking right now, and what is each leak worth in clicks per year?

The audit covers four layers:

  • Search Console reality. Sixteen months of queries crossed with landing pages, so we can see which pages rank well but under earn, which queries sit in striking distance, and where click-through rate is far below the benchmark for the position.
  • Crawl and indexation. A full crawl to surface 404s on indexed pages, orphan pages with no internal links, duplicate and parameter URLs eating crawl budget, and pages excluded from the index that should not be.
  • Authority benchmark. Your backlink profile against the sites already winning your queries: referring domains, average authority, the count of genuinely strong sources, and the topical gap.
  • Technical health. Schema coverage and validity, Core Web Vitals across mobile and desktop, title and meta quality, and canonical and hreflang correctness.

The output is a ranked list. Every item has the data behind it, an estimate of the click recovery it is worth, and the effort to fix it. You see it before any money changes hands, and you keep it whether or not you engage us.

How searchers actually click

Ranking is not the goal. Clicks are. And clicks are distributed brutally unevenly across the results, which is why a small ranking move near the top is worth far more than a large move further down, and why a great snippet at position three can out earn a dull one at position one.

Organic click-through rate by position (typical desktop)
28% 15% 11% 8% 7% 6% 5% 4% 3.2% 2.5% 12345678910 SERP position
Figures are typical benchmarks and vary by query and device. The shape is what matters: roughly a third of clicks go to position one, and the curve falls off a cliff. This is why striking distance work (moving a page from position seven to position three) and snippet work (lifting click-through at a position you already hold) are the two cheapest forms of traffic there are.

One consequence drives a lot of our technical work: a rich result, such as review stars or an FAQ accordion, can lift click-through by a third or more at the same position. A 35 percent lift at position three can pull as much traffic as ranking first without it. You do not always have to rank higher. Sometimes you just have to earn a better looking result.

AI Overviews and answer engines

The biggest shift since the last edition of this guide is the AI generated answer that now sits at the top of many results, and the rise of answer engines like ChatGPT, Perplexity, and Gemini as places people start their research. Both change what visibility means, and both are widely misunderstood.

The fear is that AI Overviews kill organic traffic. The reality is more specific. When an AI Overview appears, organic click-through on the classic links below it drops materially, by something in the range of a third to well over half depending on the query. But the answer is assembled from pages that already rank, and the brands cited inside the overview earn more clicks than the uncited competitors below it, not fewer. Visibility moved; it did not disappear.

What AI Overviews do to clicks
100 ~45 135 Classic result,no AI Overview Below an AIOverview, not cited Cited inside theAI Overview
Indexed to 100 for the no-overview case. Sitting below an overview without being cited is where traffic is lost. Being cited inside it is where traffic is won. The work that earns a citation is the same work that earns a classic ranking.

So the strategy does not change as much as the headlines suggest. Answer engines pull from authoritative, well structured pages, so the levers are the same: authority, clean structure, and content that answers the question directly. A few practices matter more than they used to, and we bake them in:

  • Direct answers near the top. A clear forty to sixty word answer to the page's core question, high on the page, is what gets extracted and cited. Bury the answer and a competitor gets quoted instead.
  • Clean heading structure. A logical H2 and H3 hierarchy with scannable lists lets a model find and lift the right passage. This is also just good for humans.
  • Claims with evidence. Specific numbers with a source attached are quoted far more often than vague assertions.
  • Freshness. Recently updated pages are cited more often, so the content refresh program below is now doing double duty.
  • Entity clarity and schema. Organization, author, and product markup help a model understand who you are and trust what you say.

What this means for you. Do not panic and do not chase a separate AI SEO product. The thin sites that never ranked are the ones losing here. Authority, structure, and direct answers win both the classic result and the citation. We optimise for both at once.

Technical SEO foundations

Technical SEO is unglamorous and it is where the fastest wins usually hide, because the traffic is already earned and something is blocking it from converting to clicks. We run it as a fixed scope of work, not an open retainer.

Titles and meta descriptions

You can rank in the top three and still lose most of the traffic, because the title and description decide whether anyone clicks. We routinely find pages at position two with a click-through rate a fraction of what the position should earn. The ranking is fine; the snippet is the problem. We rewrite titles to stay under roughly sixty characters so they do not truncate, lead with the query and a proof point, and give the description one concrete reason to click. Where intent is ambiguous, we ship two variants and let the data choose.

Crawlability and indexation

We make sure the pages you want ranked are reachable, crawlable, and in the index, and that the pages you do not want indexed are cleanly excluded. That means auditing robots rules, canonical tags, the XML sitemap, and the internal link graph. A page with no internal links pointing to it is an orphan; Google crawls it rarely and ranks it poorly, and most sites have hundreds of them.

Site architecture and internal linking

Internal links are how authority moves around your site and how Google understands what matters. A flat, well linked architecture pushes equity to the pages that convert. We map the current graph, find the pages starved of internal links, and lay out a linking plan that points authority at your priority pages with descriptive anchors.

Canonicalization and URL bloat

Duplicate content and parameter URLs (filters, sorts, session ids, tracking) waste crawl budget and occasionally surface the wrong version in search. We decide, per pattern, whether to canonicalize, noindex, or block, and we document the exact rule so your developers can apply it without guessing.

Redirects and 404s

Every indexed page that returns a 404 is a click you paid for with SEO budget and threw away. We export the list, map each dead URL to its correct destination, and hand over a redirect map ready to import, then verify each redirect returns the right status after deploy.

Who writes the code. We write the specifications, the copy, and the validation. Your developers apply the changes on your platform, and we review the work before it ships. That keeps the work fast and keeps you in control of your own codebase.

Structured data and rich results

Structured data is how you earn the richer search results: the breadcrumb trail, review stars, FAQ accordions, the knowledge panel, the sitelinks search box. It is invisible markup that tells Google exactly what a page is, and it is one of the most reliable ways to stand out in a crowded result without ranking any higher. It also helps answer engines understand and trust your pages, so it does double duty in 2026.

Which schema type maps to which page
EVERY PAGE Organization BreadcrumbList WebSite and Search SPECIFIC PAGES Product and Review FAQPage and Article LocalBusiness
Structured data is not one tag. It is a small map of types, each matched to the page where Google expects it. We write each as a reusable template and validate it before and after deploy, because invalid schema is worse than none.

The practical payoff is real: pages with correct structured data tend to earn meaningfully higher click-through, and product schema in particular lifts click-through on commercial results. We map each schema type to the page where it belongs, validate it with Google's own tools, and watch the enhancement reports until the rich results actually appear.

Core Web Vitals and performance

Page speed is a ranking factor and, more to the point, a conversion factor. A slow page loses people before they read a word. Google measures three things you can actually move, and in 2026 the responsiveness metric is Interaction to Next Paint (INP), which replaced First Input Delay and is stricter because it measures the full time until the page visually responds, not just the initial delay.

The three Core Web Vitals and their thresholds
LCPINPCLS LoadingResponsivenessVisual stability GoodNeeds workPoor 2.5s4.0s GoodNeeds workPoor 200ms500ms GoodNeeds workPoor 0.10.25
The targets to clear: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1, measured on real users at the seventy fifth percentile. Desktop usually hides problems that mobile exposes, so we check both.

The culprits are predictable: oversized images that are not in a modern format, render blocking scripts, fonts that load late and shift the text, third party tags doing too much, and long JavaScript tasks that block the main thread and wreck INP. We audit the pages that carry your traffic, order the fixes by impact against effort, and hand your developers a prioritised list rather than a vague plea to be faster.

Content and topical authority

Content is how you decide what you rank for. Google rewards sites that demonstrate depth on a subject, not a single page that mentions it once. That is topical authority: covering a topic and its neighbours thoroughly enough that you become the obvious answer, and it carries real weight in how pages are ranked.

A topic cluster: one pillar, many supporting pages
PILLARpage Sub-topic Sub-topic Sub-topic Sub-topic Sub-topic Sub-topic
A pillar page covers the broad topic; supporting pages go deep on the specific questions around it; internal links tie the cluster together and signal that this is an area you own. One strong cluster on a topic your buyers care about beats fifty disconnected posts chasing random keywords.

Existing content matters as much as new content. Pages decay: a guide that ranked two years ago slips as competitors update theirs and the query evolves. We find the pages losing ground, refresh them with current information and better internal links, and often recover more traffic from a refresh than a new page would have earned. Freshness also feeds the answer engines, which cite recently updated pages more often.

Search intent and the query map

Every query carries an intent: the searcher wants to learn something, compare options, or buy. Rank a commercial page for an informational query and you will convert no one; rank a thin blog post for a buying query and you will lose the sale to a competitor. Matching the page to the intent is half the battle.

We map your query landscape: the terms that drive impressions, the intent behind each, and the page that should own it. Then we look for two specific, high value patterns:

  • Striking distance queries. Terms ranking in positions five to fifteen with real volume. They already rank; they just need a push. Bringing the best of them into the top three is the cheapest traffic you will ever buy, because you are not building a new page, just improving one that already works.
  • Keyword cannibalisation. When several of your pages compete for the same query, Google does not know which to show and the clicks dilute across all of them. We consolidate, canonicalise, and relink so one page wins cleanly.

There is a version of link building that buys cheap links in bulk and hopes Google does not notice. It does notice, and the cleanup costs more than the links ever returned. We do the slower, durable version: editorial placements on publications that have their own audience and their own reason to exist.

Referring domains: a typical challenger gap
Leader A719 Leader B690 Leader C399 Your site117
Illustrative figures from a real audit. The gap is rarely about one missing link. It is a structural shortfall in how many credible sites reference you at all, which is exactly why link building is an ongoing program rather than a one off project.

Relevance beats volume. One link from a publication your buyers actually read is worth more than fifty from a link farm. We choose targets each month on authority, topical fit, and track record, write the content ourselves so the placement reads like editorial rather than an ad, and point the links where they do the most good: your homepage and priority pages first, then the specific pages you need to push.

It only works as a rhythm, so we run it monthly and you see into it. At the start of each month you get the plan: which publications, which anchor text, which target URLs. Nothing goes live before you have approved where it is going. At month end you get a report of what published, the links acquired, and how the authority metrics moved.

Anchor text and risk

The anchor text mix has to look natural, because an over optimised profile (every link using your exact target keyword) is its own kind of risk. We keep a healthy spread of branded, partial match, and natural language anchors, and we never chase a single keyword so hard that the profile looks engineered.

Owned media and brand authority

Earned links from other people are one route to authority. The other is authority you own. For brands with enough catalogue or expertise to sustain it, an editorial property on its own domain, publishing consistently, can rank for informational and commercial queries and feed qualified referral traffic back to your product. It runs on separate infrastructure so it never competes with or slows your main site. It is a longer play, and we will tell you plainly whether your situation justifies it.

Brand search is the quiet compounding asset underneath all of this. As more people encounter you through placements, content, and product, branded queries rise, and branded queries convert far better than anything else. Good SEO lifts brand search as a side effect, and we track it as a leading indicator. It is also a signal answer engines lean on when deciding whose answer to trust.

International and local SEO

If you serve multiple countries or languages, hreflang annotations tell Google which version of a page to show which audience. Done wrong, and it usually is, you get the wrong language ranking in the wrong market, or duplicate content dilution across versions. We audit the hreflang graph, fix the return tag errors, and make sure each market sees its own version.

If you have a physical location, local SEO unlocks the map panel and the local pack for brand and geo queries. That means LocalBusiness schema, a consistent name, address, and phone footprint, and an optimised Google Business Profile. For a single storefront it is a quick, high return addition; for multi location brands it is a program of its own.

For the tactical detail behind that, our local SEO strategy guide walks through Google Business Profile, citation tools and review management, and how to feed the resulting location data back into a CRM.

Measurement and reporting

We hold ourselves to numbers you can check yourself, not vanity metrics. The source of truth is Google Search Console, because it reports what Google actually did, not what a third party tool estimates. The metrics that matter:

MetricWhat it tells you
Clicks and impressionsReal demand reaching your pages, and how much of it you capture
Click-through rateWhether your snippet earns the click at the position you hold
Average positionWhere you rank, and the direction of travel
Rich result coverageHow many enhancements you have live and valid
Referring domainsAuthority accruing from link building
Branded query volumeThe compounding demand that the whole program builds

Reporting is weekly during active work: the numbers, the deltas against last week and last month, and the queries and pages that won or lost ground, with a note on anomalies (a traffic crash, a schema enhancement dropping out of eligible, a spike in impressions without clicks). One caveat we are honest about: Search Console data lags real life by a few days, and the click-through-rate window is a rolling average, so a snippet you rewrote today can take weeks to show its true effect. We set a watch window before re touching a surface, rather than reacting to noise.

The engagement model

The authority side is ongoing; the technical side is a focused, finite project. We run the technical work as three sprints across roughly eight weeks, with cleanup in parallel, then a long monitoring tail.

The technical engagement: three sprints across eight weeks
WEEK 1 WEEK 8 SPRINT 1SPRINT 2SPRINT 3 FoundationsSchema completionContent and completeness
  • Sprint 1, foundations. Title and meta rewrites on your top pages, BreadcrumbList and Organization schema, LocalBusiness spec if you have a storefront, and the 404 redirect map.
  • Sprint 2, schema completion. WebSite and SearchAction, Open Graph and Twitter metadata, individual review markup, and resolution of brand cannibalisation and parameter URL bloat.
  • Sprint 3, content and completeness. Product titles and descriptions at scale, FAQ content marked up as FAQPage, direct answer blocks for answer engine visibility, and the highest value striking distance queries pushed.

Every intervention follows the same loop: audit the current state from real data, write a spec precise enough that your developers can apply it without guessing, validate the deploy with Google's tools, then monitor the result at four, eight, and twelve weeks. SEO does not pay off on a two week schedule. The first two months stop the losses and lay foundations; the compounding shows up later, as authority accrues and the fixes mature.

One point of contact. You get one person who owns the relationship, the delivery schedule, and any escalation. Specifications arrive in a standard format with acceptance criteria your developers can check against, and the starting datasets and reports are yours to keep.

What does not work

A short, honest list of the things we will not sell you, because they waste money:

  • Guaranteed rankings. Nobody controls Google's ranking. Anyone who guarantees a position is either naive or lying, and the contract that promises it is the one to walk away from.
  • Bulk link buying. Cheap links from link farms and private blog networks are the fastest way to a penalty. The cleanup costs more than the links ever returned.
  • Publishing for the sake of volume. Fifty thin posts chasing random keywords build nothing. One deep cluster on a topic your buyers care about beats all of them, and AI driven mass content makes this worse, not better.
  • Chasing every keyword. Relevance and intent beat raw volume. Ranking for a high volume term nobody who searches it would ever buy from you is a vanity metric.
  • Set and forget content. Pages decay. Content that is never refreshed slides down as competitors update theirs.
  • Treating AI search as a separate product. The work that wins citations is the work that wins rankings. Anyone selling a standalone AI SEO package is repackaging the basics.

Frequently asked questions

How long before SEO shows results? +
Technical fixes can lift click-through within weeks because they change your search snippets directly. Authority from link building compounds over months, so a realistic horizon for meaningful movement is three to six months, with the steeper gains in the back half of the first year.
Is SEO still worth it now that AI Overviews answer questions directly? +
Yes, but the goal shifts. AI Overviews are assembled from pages that already rank, and brands cited in them earn more clicks, not fewer. The work that wins classic rankings is the same work that wins citations: authority, structured content, and clear direct answers. The sites that lose are the thin ones that never ranked well to begin with.
Do you guarantee rankings? +
No. Google controls the ranking, not us. We commit to the work and to signals you can verify in your own Search Console: click-through rate, rich results, position, and the volume and quality of links acquired.
Who implements the technical changes? +
We write the specifications, the content, and the validation. Your developers apply the code changes on your platform, and we review the work before it ships. That keeps the work fast and keeps you in control of your own codebase.
Is your link building safe? +
Bulk, low quality link buying is what gets sites penalised. We do editorial placements on real publications with real audiences and a natural anchor text mix. That is the approach Google rewards rather than punishes.
What does it cost? +
Link building runs on monthly packages sized to how aggressively you want to close the authority gap, and the technical work is a fixed scope. The exact number comes after the free audit, because it depends on where you are starting from. The fastest way to a real number is a free audit.

That is the whole method. When you want it applied to your site, the next step is a free audit: real findings on your real data, in about a week, with no obligation.

Get a free SEO audit

Ready to put this to work?

A free audit on your real data, the findings quantified, and a fixed scope to fix what matters. Findings in a week.

Get a free SEO audit
No credit card · You keep the audit · 24h reply

Related reading

Related reading

Related reading

Related reading

Get a free SEO audit