Writing for people, not search engines

Technical SEO is table stakes. The businesses that win in search are the ones writing content that actually connects with real people. Modern algorithms are good at telling the difference.

15 Mar 2025 · 4 min read · Growth

A business invests heavily in technical SEO. Site speed is excellent. Schema markup is in place. The keyword research is thorough. Six months later, rankings have improved but enquiries haven’t. The content reads as if it was written for a crawler, not a person.

That’s the technical SEO trap: treating optimisation as the goal rather than the means.

What the algorithms actually reward

Google has spent years improving its ability to detect whether content genuinely helps the person who found it. The BERT update gave the algorithm a much deeper understanding of language context. E-E-A-T (Experience, Expertise, Authoritativeness, Trust) formally incorporated signals of real-world credibility. The Helpful Content system created a site-wide quality signal, meaning low-quality content anywhere on your site can drag down pages that would otherwise rank well.

More recently, AI search has raised the stakes further. AI overviews and AI-powered summaries pull from content that answers questions directly and specifically. Vague, padded writing that once survived on keyword density gets bypassed in favour of content that actually gives the searcher what they came for.

The direction has been consistent for years: algorithms are getting better at identifying content that serves real people, and rewarding it accordingly.

What human-first content looks like

The distinction isn’t about tone or personality. It’s about whether the content is structured around what the reader needs, or around what the writer wanted to say.

It answers the actual question: Not a related question, not a broader topic, not a version of the question that's easier to answer. If someone arrived asking "how much does a WordPress redesign cost?", the page needs to address cost, not just explain what a redesign involves.
It uses the reader's language: Industry terminology creates distance. If your clients call it a "website refresh" rather than a "site migration" or "CMS overhaul", write in the language they use. It also helps with long-tail search visibility.
It's structured for scanning: Most people don't read web content linearly. They scan for the section that's relevant to them. Subheadings, short paragraphs, and clear formatting help readers find their answer without reading everything, and signal to search engines that the page is well-organised.
It addresses the surrounding concern: People searching for information often have an underlying worry or goal beyond the question itself. A business owner asking about GDPR compliance isn't just looking for a checklist. They want reassurance that they won't get fined. Content that acknowledges this layer builds more trust than content that doesn't.

Building this into your content

The most common failure mode is starting with a keyword and working backwards: identifying what needs to rank, then producing something that covers the topic. The result is often technically correct but somehow empty. It checks the right boxes without genuinely helping anyone.

Working the other way produces better results. Start with the question your readers actually ask, either from real conversations or from Google’s own “People Also Ask” results and search console queries. Write the clearest, most honest answer you can. Then consider how to structure it for the web: headings, formatting, length.

This isn’t a longer or harder process than keyword-first writing. It’s a different frame.

For existing content, the same principle applies. Rather than a full rewrite, identify the primary question each page is supposed to answer and ask honestly whether the page answers it. If the answer is buried on page two, move it to the top. If it’s missing entirely, add it.

One client rewrote a handful of service pages this way, shifting from capability descriptions to direct answers to the questions their enquiries were actually asking. Traffic doubled within three months and conversion rate tripled. The pages hadn’t changed in terms of the services offered; they’d changed in terms of who they were speaking to.

The engagement signals that matter

Beyond rankings, human-first content creates measurable differences in the signals that feed back into search performance over time.

Pages that answer questions well see longer average time on page, more clicks through to related content, and lower exit rates from visitors who hadn’t found what they came for. Those engagement signals feed back into how search engines evaluate the page’s usefulness.

Useful content also earns links and shares more naturally than content produced for algorithms. Other sites link to resources that genuinely helped them. People share content that solved a problem. Neither of these things happens reliably when content is optimised to rank rather than to help.

If you’re looking at metrics, focus on whether visitors are taking next steps: contacting you, reading more, returning. These are stronger indicators of content quality than ranking position on its own.

Want visibility and tracking set up properly? Explore SEO foundations →

Frequently asked questions

Does technical SEO still matter if I focus on writing for people?

Yes. Technical SEO is the foundation. A well-written page that loads slowly, has broken internal links, or can't be crawled won't rank regardless of quality. Think of it as two layers. The technical layer gets your content into consideration. The human layer determines whether it actually ranks and converts. You need both, but most sites have the technical basics covered and are being held back by the content layer.

What does writing for people actually mean in practice?

It means leading with the answer rather than building up to it, using language your readers actually use rather than industry terminology they might not, structuring content so someone can scan it and find what they came for, and being specific rather than vague. A page that genuinely helps someone accomplish something is what "writing for people" looks like in practice.

How do I know if my content sounds too algorithmic?

Read it out loud. If it feels stilted, repeats the same phrases in a way real people wouldn't, or takes three paragraphs to say something that could be said in one sentence, it was probably written with keyword targets in mind rather than a real reader. You can also ask someone unfamiliar with the topic to read it and tell you if they found what they were looking for.

How has AI search changed what writing for people means?

AI overviews and search summaries pull directly from content that answers questions clearly and specifically. Vague, padded content that ranks via keyword density gets bypassed entirely. Being cited in an AI overview requires the kind of direct, authoritative writing that helps a real person, which is exactly what has always worked for human readers. The bar has gone up, not changed direction.