What Is Google SEO? Five Surfaces to Earn Visibility
Google SEO is the practice of making a page eligible to appear on the surfaces Google controls: the results page, the map pack, image results, the Discover feed and AI-generated overviews. That is the whole definition, and it is wider than most people expect. A page can hold a top blue link and still be absent from Maps, invisible in Images and never quoted in an answer summary. Once you see the five surfaces as layers of one stack, the question turns into a checklist: which layer is missing, and what asset does it need?
What is Google SEO in one sentence?
Google SEO is the work of earning unpaid visibility inside the properties Google owns - Search results, Maps, Images, Discover and AI overviews - by publishing pages, profiles and media that each property can crawl, understand and trust enough to show to the right searcher.
Google is five surfaces, not one results page
For twenty years the mental model was a list of ten links. That model is now wrong in a measurable way. A single query can return an answer summary, a map pack, three image tiles, a short video shelf and a handful of links, and each of those blocks is filled by a different system with different inputs. Google does not rank a website as a whole. It ranks individual assets - a page, a photo, a business profile, a product feed - inside the surface where that asset can appear.
This is why two sites with near-identical service copy can have completely different outcomes. One has a crawlable page with a clear topic and clean internal links; the other keeps the same information inside a script, a login or a PDF. One has a verified profile with recent reviews; the other has a pin dropped somewhere in the city. The stack metaphor is useful because it forces you to name the layer before you argue about tactics. If you want the plain-language definition of the discipline first, read Google SEO Services and then come back to the layers below.
The order matters. The base layer is a page Google can fetch, render and parse. Every upper layer borrows from it: a Discover card still lands on that page, and an answer summary still cites a passage from it. When the base is thin, the upper layers do not explain themselves - they simply do not appear, and the traffic report looks like a mystery.
So a practical audit reads bottom-up. Crawl and index first. Then intent match, then structure, then asset coverage, then measurement. Teams that start at the top - chasing an AI citation or a Discover placement - usually spend a quarter on decorative work while the base layer stays broken.
What each layer of the stack asks for
Each surface reads different evidence. The table below is the version we use in audits: it maps the position to the asset that unlocks it, and to the signal Google weighs when deciding who appears there.
| Layer | Where the searcher sees it | Asset you need | What gets weighed |
|---|---|---|---|
| Search results | Blue links, snippets, answer citations | Crawlable HTML page with one clear topic, a descriptive title and structured data | Relevance, internal links, page experience, how cleanly a passage answers the query |
| Maps and local pack | Three-pack, local finder, directions | Verified business profile, consistent name and address, service pages, a review flow | Proximity, category fit, prominence and review volume |
| Images | Image tab, visual search, shopping tiles | Original photos, descriptive file names, alt text, captions and supporting page copy | Image context, the topic of the page around it, technical quality |
| Discover | Mobile feed cards under the search box | Fresh, image-led articles with a specific audience and a strong first screen | Content quality and originality, reading behaviour, page experience |
| AI overviews | Generated summary above the links | Short quotable passages, clear entities, facts stated the same way across the site | Passage relevance, agreement across sources, clarity of the cited source |
Read the rows as dependencies, not as a menu. A factory site that sells to distributors may only ever need the base layer and the image layer; a local installer needs the map layer before anything else; a publisher lives or dies on Discover. The right question is never "are we doing all five" but "which two move revenue this quarter".
Five habits that flatten the stack
1. Treating SEO as one channel with one dashboard
If the only number you watch is organic sessions, you cannot tell a base-layer failure from a Discover dip. Split reporting by surface and the diagnosis takes minutes instead of a quarter. Search Console already separates Search from Discover; the local layer lives in the business profile; image performance has its own report.
2. Writing to a keyword density target
Repeating a phrase does not move a page up a layer. What moves it is answering the query in the first screen, naming the thing precisely, and giving the reader a reason to stay. Pages built around a phrase count read like forms; pages built around a question get quoted.
3. Shipping pages Google cannot read
A page that renders its copy only after three script calls is a page that may be indexed as an empty shell. Test with the rendered HTML, not the browser tab. If the main copy, the price table or the spec sheet is missing from the rendered version, the base layer is not finished.
4. Skipping the map layer because the website looks fine
For any business with a service radius, half of the buying intent happens in the map pack, where the website barely appears. Categories, service areas, opening hours and review responses decide that layer, and none of them are fixed by another blog post.
5. Publishing images with no context
A photo named IMG_4821.jpg with empty alt text is invisible to the image layer even when it is the best photo in the category. Rename it, describe it, caption it, and place it next to copy that says what it shows. That is a ten-minute fix with a long tail.
Where to start on Monday
Run the layers bottom-up. Each step makes the next one cheaper, and each one produces a document you can hand to a developer or a writer without translation.
Steps one to three are usually a two-week sprint on a mid-sized site. Step four is where most budgets get misallocated, because it is the point where someone asks for a new landing page when the real gap is a verified profile or six original photos. Step five is not a report for its own sake: it is how you notice that a template change quietly removed the schema from 400 product pages.
What we do when we take over a stack
Our Whats Is SEO? engagement opens with a layer audit, not a content calendar. We map what exists, what is missing per surface, and what the site can realistically support in the next ninety days.
Layer audit
Crawl, index, intent, structure and asset coverage across all five surfaces, written as a prioritised list.
Template fixes
Titles, headings, internal links, schema and page speed applied at template level so every URL benefits.
Asset production
Briefs written to answer one query, photo and alt-text guidance, and local profile work where it applies.
You get the audit document whether or not you continue, and every recommendation is tied to a layer, so your team can see why it is on the list. If the base layer is already healthy, we say so and start higher up.
Want the five-layer audit run on your site?
Send the domain. We will tell you which layer is leaking before you spend on content.
Questions we hear before the first call
Is it the same thing as SEO?
Practically, yes, because Google still delivers the overwhelming majority of search traffic. The distinction is useful when you work on surfaces Google does not own, such as Bing, YouTube or marketplaces. For most B2B and export sites, the two terms describe the same program.
Do I need all five layers?
No. The base layer is non-negotiable. The map layer matters if buyers can visit you or need a local supplier. The image layer matters if your product is visual. Discover and answer summaries reward publishers and brands with a point of view. Choosing two layers and doing them properly beats five half-built ones.
How long before a new layer shows results?
Indexing fixes can move within days. Template and content work usually shows in four to eight weeks. Local profiles and image libraries build gradually over a quarter. Answer summaries are the least predictable, because they depend on how consistently the wider web describes your topic.
Does this work for a factory or an exporter?
Yes, with a different centre of gravity. Export sites usually need the base layer, a strong image and spec layer, and content that answers technical sourcing questions. The map layer rarely applies, and Discover rarely converts a distributor inquiry.
What do you need from us to start?
Read-only access to Search Console and analytics, a list of the pages that generate inquiries, and one person who can approve changes. If you sell locally, add access to the business profile. That is enough for the first audit pass.
How do you report on AI overviews?
We sample a fixed set of queries each month, record whether a summary appears, whether the site is cited, and which passage is used. It is a sample, not a census, but it shows direction and it stops the team from guessing.