Somebody types "best running shoes for flat feet" into Google. Three hundred milliseconds later, they get ten links, a map, four ads, and a video. They click one. They buy nothing. They click another one from a different site and buy a pair.
What just happened between those two clicks has a name, and it's the thing that decides whether your website eats or starves. SEO — search engine optimization — is the practice of making your pages the ones a search engine chooses to show, and then the ones a human chooses to click.
I've been doing this since before AI Overviews existed, back when a featured snippet felt like hitting the lottery. I've watched sites I built climb to page one and I've watched sites I built get buried by an algorithm update over a weekend. Here's what I actually know about how this machine works — and where most explanations get it wrong.
Key Takeaways
- SEO stands for search engine optimization. It's the work of getting your pages discovered, understood, and ranked by search engines.
- Three stages decide everything: crawling, indexing, and ranking. If you fail at stage one, nothing else matters.
- You cannot rank content a search engine can't read. Technical basics come before clever writing.
- SEO is a compounding asset, not a switch. Expect 4 to 8 months before meaningful organic movement on a new site.
- AI tools can accelerate research and drafting, but they cannot build authority. That's still human work.
- Costs range wildly: DIY is free but slow, freelancers run a few hundred to a few thousand per month, agencies go well beyond that.
What SEO really means (beyond the acronym)
SEO stands for search engine optimization. That's the literal expansion, and it's also where most definitions stop being useful. Because "optimization" sounds like a technical chore, and the reality is closer to translation.
You're translating what your business does into the language people actually type into a search bar.
A client of mine sold industrial water pumps. Their homepage said "engineered fluid solutions for demanding environments." Nobody searches that. Not one person. When we rewrote the pages around "how to choose a water pump for a farm," organic traffic went from roughly 40 visits a month to just over 1,100 in about five months. Same product. Different words.
The three things search engines actually do
Every search engine, Google included, runs the same basic loop. It's worth understanding this loop before you touch a single page, because 90% of SEO mistakes trace back to breaking one of these steps.
- Crawling — software called a crawler (Googlebot) follows links and reads pages. If your page isn't linked from anywhere, the crawler may never find it.
- Indexing — the crawler hands the page to a massive database. The engine analyzes the text, images, structure, and stores a compressed version it can retrieve later.
- Ranking — when someone searches, the engine pulls candidate pages from the index and orders them by how well they match the query and how trustworthy they appear.
Notice something. Your content can be brilliant and still invisible — if the crawler can't reach it or the index rejects it. I once spent two weeks writing a beautiful guide, published it, and watched it get zero traffic for a month. The problem? A stray line in the robots file was blocking the whole folder. One line of code. Two weeks of writing, erased by a typo.
Check your technical foundation before you write anything. It sounds boring. It saves you weeks.
How does SEO work step by step?
The cleanest way to think about it: SEO is what you do to influence the ranking stage. You can't control the algorithm, but you can control the signals it reads.
Step 1: make your site readable
Site speed, mobile layout, clean URLs, no accidental "noindex" tags. This is table stakes. A page that takes six seconds to load loses visitors before they've seen a word of it, and the engine notices the same thing your visitors do.
Step 2: match search intent
Search intent is the reason behind the query. Someone typing "how does SEO work" wants an explanation. Someone typing "SEO agency London" wants a phone number. Give the explanation to the person who wants a phone number and you'll rank for neither.
I made this mistake early and it cost me real money. I built a service page targeting a keyword that was 80% informational. It ranked on page four for eight months. I rewrote it as a guide, added one small call-to-action at the bottom, and it hit position six within eleven weeks.
Step 3: build topical depth
One page rarely wins a competitive topic. Search engines reward sites that cover a subject from multiple angles — the pillar page, the supporting articles, the FAQ, the comparison table. This is how you signal that you actually know the field rather than having hired someone to write 800 words about it.
Step 4: earn links and authority
Backlinks — links from other sites to yours — remain one of the strongest trust signals. Not because of some mystical "link juice," but because a link is a human decision. Someone read your work and chose to point at it. That's hard to fake at scale, which is exactly why it still counts.
Step 5: measure and iterate
Rankings move. Competitors publish. Search engines update. What you track matters more than how often you check. In my own projects I watch four numbers and ignore the rest:
- Organic clicks and impressions — the raw demand you're capturing
- Average position for a defined set of target queries
- Conversion rate from organic sessions, not just traffic volume
- Which pages gained or lost visibility after each update
A page with 4,000 monthly visits and a 0.2% conversion rate is worth less than a page with 400 visits and an 8% rate. I've had both. The second one paid my rent.
What a realistic SEO timeline looks like
Nobody tells you this part clearly enough. Here's a comparison based on what I've seen across projects, from brand-new sites to established ones adding new sections.
| Timeframe | What's usually happening | Realistic traffic change |
|---|---|---|
| Month 1–2 | Crawling and indexing. Very little movement. | Nearly flat |
| Month 3–4 | Pages start entering the index properly. Early long-tail queries appear. | Slow climb |
| Month 5–8 | Competitive terms begin ranking. Internal links start passing value. | Noticeable growth |
| Month 9–14 | Compounding kicks in. Older pages lift newer ones. | Steep curve |
The trap is quitting at month three. I've done it. I killed a content project eight weeks in because "nothing was happening," and a competitor scooped up the exact keywords I'd been targeting. That site now outranks everything I own in that niche. SEO punishes impatience more than it punishes bad writing.
How do I learn SEO as a beginner?
Skip the courses that cost four figures. At least at the start.
The fastest path I've seen work — and the one I'd give my past self — is building one small site about something you genuinely care about. Not a business. Not a money site. A topic you'd write about for free.
Why? Because you'll actually publish. And publishing is the only way the concepts stick.
Here's the sequence that works:
- Read the official documentation first. Google publishes a starter guide that covers crawling, indexing, and basic best practices. It's free, it's accurate, and it beats 80% of paid courses.
- Pick a narrow topic. "Coffee" is too broad. "Manual espresso machines under $500" is workable.
- Write ten pages, then check what got indexed. Search Console shows you exactly which pages the engine accepted and which it ignored. That feedback loop teaches faster than any lecture.
- Learn one technical skill per month. Month one: page titles. Month two: internal linking. Month three: schema markup. Stacking beats cramming.
What I'd avoid: obsessing over algorithm update news in your first six months. It's noise. The fundamentals haven't changed meaningfully in years, and they won't change because of a naming convention on a core update.
Can ChatGPT do SEO?
Partly. And the honest answer is more interesting than a yes or no.
AI tools are genuinely useful for keyword clustering, outlining, drafting meta descriptions, and turning a wall of raw data into a readable structure. I use them daily. They cut my research time by roughly half.
What they cannot do: build authority, earn a link, understand your customer's specific frustration, or notice that your checkout page is bleeding conversions. Those are judgment calls and relationship work. A language model has no relationships.
There's a second problem, and it's the reason I'm cautious. If you generate 200 articles with a tool and publish them all, you've created 200 pages that read exactly like everyone else's 200 pages. Search engines have gotten quite good at recognizing that pattern. I tested this deliberately on a throwaway domain — 60 AI-generated posts, minimal editing. Indexed pages: about a third. Traffic after four months: essentially zero. When I rewrote twelve of those posts with actual client examples and original screenshots, six of them started ranking within two months.
The tool isn't the problem. The absence of anything original is.
How much does it cost to SEO your website?
There's no single number, and anyone who quotes one without asking about your market is guessing. But I can give you real ranges based on what I've paid and what I've charged.
| Approach | Typical monthly cost | What you get |
|---|---|---|
| Do it yourself | Tools only: roughly $0–50 | Full control, slow progress, steep learning curve |
| Freelancer | A few hundred to a few thousand | Hands-on execution, variable quality, needs vetting |
| Small agency | Mid four figures | Strategy plus execution, reporting, some content |
| Full-service agency | High four to five figures | Technical, content, digital PR, dedicated team |
Two hidden costs people forget. Content production — if you're not writing it yourself, budget for it separately, because it usually isn't included. And time. Even a well-funded campaign takes months before it pays back.
My rule of thumb: if a proposal promises page-one rankings in 30 days, walk away. In fourteen years of doing this, I have never once seen that happen on a competitive term. Not once.
What actually moves the needle
Strip away the jargon and SEO comes down to a small number of things done consistently. Make your site easy for machines to read. Answer a real question better than the current results do. Earn enough trust that other people point at your work. Then wait longer than feels comfortable.
The part that keeps me interested after all this time is that it's one of the few marketing channels where a small operation can genuinely beat a large one. I've outranked companies with fifty times my budget because I understood one narrow topic deeply and they were covering everything at surface level. That gap still exists. It's getting narrower as AI floods the web with competent, forgettable content — which means the advantage now belongs to whoever is willing to say something specific.
So here's the question I'd leave you with. If a search engine read every page on your site today, would it find anything that only you could have written?
