SEO isn't a marketing job — it's an engineering job. But the most important marketing thing isn't SEO either

Toward the end of every proposal, the same question lands:

"Will the site rank on Google after we launch?"

My standard answer: Ranking isn't hard, but you're asking the wrong question.

"Will Google find me" is engineering hygiene, not a feature; the real question is "is anyone searching for what I do?" — and that's a marketing question, not an engineering one. This post pulls apart where SEO actually sits in the engineering / marketing stack, and what the most important marketing thing actually is.

The engineering side of SEO: 80% is infrastructure, do it once and move on

In engineering terms, SEO isn't "optimisation" — it's "don't get this wrong."

If any of the following are missing, no amount of great content will get indexed:

Mandatory (skip = doom)

1. Every page has a unique <title> and <meta description> — and they actually differ per page
2. canonical URL — tells Google which URL is the source of truth, prevents duplicate content penalties
3. hreflang tags — for multilingual sites; without these, your zh and en versions cannibalise each other (Google genuinely doesn't know which to serve to whom)
4. sitemap.xml + submitted to Google Search Console
5. robots.txt not broken (don't ship Disallow: / to production)
6. Server-side rendering — pure client-side React apps (no SSR/SSG) get crawled as empty shells
7. Mobile Lighthouse Performance ≥ 60 — Core Web Vitals are a ranking factor
8. Correct H1 / H2 / H3 hierarchy — exactly one H1 per page

These 8 items: a competent agency treats them as the floor (every project I ship has them); an incompetent one waits for you to ask, then bills extra.

Advanced (additive but ROI declines fast)

• Structured data (JSON-LD): Article / Product / FAQPage / Breadcrumb
• Image alt text + image sitemap
• Internal linking structure (which pages route traffic to "key pages")
• Continuous performance optimisation (LCP < 2.5s, CLS < 0.1)

But once these are done, that's the ceiling. Going further into SEO optimisation (keyword density tweaking, backlink campaigns, anchor text strategies) hits diminishing returns fast — and most of that energy should redirect to the next thing.

The most important marketing thing: not SEO, it's "positioning + content + cadence"

Direct version: SEO ranks third at best in marketing priority. The two ahead of it:

1 · Positioning — who you serve and what problem you solve

The biggest pit I see owners fall into: they ship a beautiful site whose homepage doesn't say "what they do."

Open the homepage and the copy says "professional and reliable services," "exceptional quality," "customised solutions" — identical to every competitor. Google has no signal for ranking you against them; visitors leave with no idea how you differ.

Right positioning looks like:

• ❌ "We provide professional website development services."
• ✅ "I build custom sites for SMB owners. No templates, no monthly lock-in, source code is yours after launch."

The second version answers in one read: Who? What's different? Will I get locked in? — these are the questions in a customer's head as they're Googling you.

Wrong positioning = all SEO is wasted, because you don't even know who's supposed to find you.

2 · Content — answer the specific questions customers actually Google

Google's most important algo update post-2024 is E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). Plain version: articles that show real experience and a point of view rank far ahead of keyword-stuffed content.

What's a "real-experience" article? Examples:

• ❌ "Comprehensive Guide to Website Pricing" (keywords stuffed, content empty)
• ✅ "I've reviewed 30 client quotes — here are the 5 variables that drive the spread" (specific experience, has a POV)

The blog post you're reading right now is the second style. Every post includes real cases (anonymised), specific numbers, conclusions you can verify. Readers come away thinking "this person actually did this," and Google judges the same way.

Content that ranks reads like an expert's notes from the field, not an SEO textbook for robots.

3 · Cadence — Google likes "live" sites

A brutal SEO truth: a site with 100 articles will outrank a site with 5 great articles, every time.

That's not to say quality doesn't matter — it's that frequency itself is a ranking signal. Google reads "publishes consistently" as "active source in this field."

How often?

• New site: 1 post / week, sustained 6-12 months, accumulating 30-50 baseline articles
• Mature site: 2-4 posts / month, long-term sustained
• Lazy mode: one article every 2-3 months when the mood strikes — expect zero SEO traffic from this

Most owners can't sustain this part. SEO isn't a launch-day task; it's a 12+ month accumulation game. Sites willing to accumulate end up with 10-50× the traffic of sites that don't, three years later.

Should you buy SEO tools (Ahrefs, SEMrush)?

99% of SMBs don't need them. These tools (starting at $99/month) help you:

• See what keywords competitors rank for
• Get search volume data
• Monitor your ranking changes

But before you have 30 articles and 1,000+ monthly visitors, none of that data is useful — what you should be doing is clear: write articles, write the right ones. Buying tools just leads to analysis paralysis: lots of dashboards, nothing shipped.

My recommendations:

1. Free Google Search Console (Google's own, complete) — required
2. Free Bing Webmaster Tools — 5 minutes to set up
3. Free Google Trends — see topic interest over time
4. Once monthly traffic crosses ~5,000 and you want advanced optimisation, consider paid tools

Action checklist for owners (in priority order)

A realistic execution order — most people do SEO wrong by doing it in the wrong order:

Month 1

• Engineering basics: title / meta / canonical / hreflang / sitemap / robots.txt all verified
• Submit to Google Search Console + Bing Webmaster
• Rewrite the homepage — make positioning crystal clear (who, what problem, how you differ)
• Write 3 articles answering the questions clients ask most (one per language if multilingual)

Months 2-3

• One article per week — if you can't sustain this, don't do SEO
• Build internal linking: posts cross-link, posts link to service pages
• Add works / case study pages to sitemap
• Every image: alt text, webp format, dimension-capped

Months 4-6

• After 15-20 articles, check Google Search Console for which keywords start showing impressions
• For pages with "impressions but low CTR," rewrite title / description (Google has placed you in results — it's the click-through losing)
• Identify the few articles getting organic traffic and replicate their style on similar topics

Months 6-12

• At 30+ articles, traffic typically hits a "second curve" (a few articles suddenly take off)
• Now consider: schema.org structured data, brand-keyword ranking strengthening
• Referral / social traffic kicks in — SEO and social compound each other

Honest timeline

Following this flow, the first noticeable traffic wave usually arrives at month 6-9. Clients willing to walk this path see 5,000-20,000 monthly organic visits at the 18-month mark; clients who chase quick-fix SEO consultants are usually still at 0-500 visits at month 18, then conclude SEO is a scam.

Google's algorithm has become more like a person

The 10-year arc of SEO: from "trick the algorithm" to "serve the real reader."

A decade ago, SEO engineering meant keyword stuffing, link buying, and clever exploits. Today, 95% of those tactics not only don't work — they get you penalised.

What's left is:

1. Get engineering basics right (engineer's job)
2. Write positioning clearly (owner + marketing's job)
3. Publish substantive content consistently (writer + owner's job)

Notice the engineer only owns 1 of these in the funnel — and it's the "don't get this wrong" lane, not the "win the race" lane.

If you're thinking through your site's SEO strategy, book 15 minutes. I'll tell you straight:

• Which engineering basics you're missing
• Whether your positioning is clear enough to even justify SEO investment
• A reasonable content cadence design for you (not "write every week" — but a plan)

SEO isn't an engineering project that ships once; it's a living thing you have to keep feeding. Decide whether you have the 12-month appetite to feed it before you decide to start.