A century of SEO

The universe, as Douglas Adams once observed, is big. Really big. You just won’t believe how vastly, hugely, mind-bogglingly big it is. The internet, by comparison, is only slightly smaller, and considerably more cluttered with bad SEO advice.

I’ve been doing SEO for long enough to have formed opinions about it. Strong opinions, held lightly, because that is roughly tip number one right there: a large proportion of what you read about SEO on the internet is not necessarily factual. Google is cagey about what actually works, which means the entire field runs on experiments, anecdotes, and the digital equivalent of folklore passed down around fires on the savannah, except the savannah is Reddit and the fires are comment sections.

When someone tells you they’ve discovered a sure-fire SEO trick, apply a sensible grain of salt. Google it (the meta-irony is free). Understand what evidence underpins the claim before reorganising your entire content strategy around it.

This post grew out of a framework originally sketched by Dr. Jigsaw on Reddit, which I’ve expanded considerably with my own experience. I’ve managed to assemble 102 tips. Some are obvious. Some will surprise you. One or two might save you a meaningful amount of money. None of them involve crystals or simply believing harder.

 

Part One: Strategy, or Why Most People Are Doing This Wrong Before They’ve Even Started

The first thing you need to make peace with is that SEO takes time. Depending on the competition in your niche, you might wait anywhere between six months and two years before seeing meaningful results. This is not a bug. It’s not a conspiracy. It’s simply how a system built on trust and authority accumulates both of those things, slowly and with the kind of glacial patience that would impress a geologist.

This leads naturally to the observation that SEO is not the right channel for everyone. If you are a newly founded business trying to achieve profitability before the end of the quarter, SEO is probably not your primary lever. You’ll burn through patience and enthusiasm long before the algorithm takes any notice of you. In that situation, look at PPC, outreach, or content marketing that can generate returns on a faster cycle.

Speaking of PPC: it’s actually an excellent validation tool before you invest in SEO. Set up Google Search Ads for your highest-intent keywords. See how well that traffic converts. If it doesn’t convert when you’re paying for it, ranking organically for those keywords won’t help you either, and you will have learned this for the relatively modest price of a small ad budget rather than eighteen months of content production. The best validation tool of all, incidentally, may be Google itself: start typing a phrase into the search box and watch what it suggests. This is free and, in my experience, remarkably illuminating. Google is essentially telling you what people actually want to find.

Once you’ve committed to SEO, you’ll want to know whether it’s working. Google Search Console will tell you. Check monthly to see whether articles are indexed and whether their average position is improving. GSC won’t tell you that you’ve won, but it will tell you whether you’re moving in the right direction, which is rather more useful when you’re in month three and beginning to wonder whether you’ve made a terrible mistake.

 

Content Volume: The Uncomfortable Truth

Let’s talk about content strategy, which is ultimately the most significant element in SEO and the part that most people systematically underprioritise in favour of obsessing over meta tags and schema markup.

Publish a lot of content. A lot. We’re talking a minimum of ten thousand words per month, with twenty to thirty thousand being optimal, particularly if your website is new. If an SEO agency approaches you offering four five-hundred-word articles per week as part of their package, walk away. No one has achieved meaningful SEO results with short, infrequent content. It is an utter waste of time dressed up as a service, which is, in fairness, also a description of a great many things in the world of marketing.

Scaling content production means scaling your writers, and building a writing team properly matters. Once you find a writer who performs well, promote them to editor. Promote your best editor to Head of Content. This creates a structure that can actually manage quality at scale, rather than the alternative, which is one exhausted person trying to edit everything while simultaneously answering emails about whether the Oxford comma is mandatory.

Finding those writers is harder than it sounds. You need to source writers constantly, not just occasionally. Roughly ninety-five to ninety-nine percent of applicants won’t be suitable. Of those that remain, about eighty percent will be weak writers. The remaining fraction are the people you’re looking for. Source continually, filter rigorously, and never rely on UpWork as your primary channel. It is not good. Of course, qualified writers exist there, just as qualified people exist everywhere, but from experience they are few and far between. ProBlogger, the Cult of Copy Job Board, and headhunting on LinkedIn are more reliable alternatives. If you must use UpWork, use it for headhunting rather than posting a job advertisement and waiting to be surprised.

When evaluating writers, look at three things: fluency in English, quality and length of their existing samples, and whether they’ve demonstrated they can research and explain technically difficult topics. Someone who has written about cybersecurity can probably handle your niche. Someone whose entire portfolio consists of travel listicles and restaurant recommendations probably cannot, and the month or two it takes them to get up to speed is a month or two of content that doesn’t rank.

For the kind of volume production that actually moves the SEO needle, consider hiring writers full-time rather than relying entirely on freelancers. For your own blog, freelancers are usually fine. For a content marketing agency serving multiple clients at scale, full-time employees deliver more consistency, which turns out to matter quite a lot.

Your virtual assistant, incidentally, is a resource you’re probably underusing. There is an enormous amount of tedious SEO work that shouldn’t be consuming your team’s attention: gathering contacts for link-building outreach, uploading articles to WordPress, basic administrative tasks that expand to fill whatever time you give them. Delegate these. Your writers and editors should be writing and editing, not maintaining spreadsheets at ten o’clock at night.

 

Authority, Niches, and the Focused Approach

Google rewards topical authority. A website with a hundred articles on digital marketing is considered more authoritative on that topic than one with ten, and Google is measurably more likely to reward it accordingly. More critically: thirty articles on accounting will outrank ten articles each on accounting, sales, and business management. Pick one niche. Establish authority there first. Then expand. The alternative is spreading your content team thin across multiple topics and achieving mediocrity in all of them, which is a surprisingly popular approach.

While we’re talking about not wasting effort: technical SEO matters, but you shouldn’t get lost in its infinite detail. Google won’t penalise you because your website doesn’t load in three milliseconds or because a single page is missing a meta description. Get the fundamentals right: run fast, create good content, get backlinks. You’ll rank, even if your implementation isn’t surgically perfect. The fundamentals are load speed, content volume and quality, and backlinks. Everything else is tuning.

Google isn’t the only search engine that matters everywhere, either. If you’re targeting China, your priority is Baidu. Russia means Yandex. These are not edge cases in their respective markets; they are the dominant search engines, and treating them as footnotes to a Google-first strategy is a mistake with consequences.

Two persistent myths worth dismissing before we go further. Voice search is not relevant and, despite a decade of confident predictions to the contrary, shows no signs of becoming so. It is simply too impractical for most search queries. You can safely deprioritise your voice search optimisation strategy, to the extent that you have one. Similarly, SEO is not dead, regardless of the article published about this every January since approximately 2012. Ignore those pieces. They are filed under “content that exists because someone needed to publish something.”

 

Keyword Research: Do It Once, Do It Properly

Make your initial keyword research as comprehensive as possible. Aim for one hundred to three hundred keywords, depending on the niche. This gives you a realistic picture of the competitive landscape and allows you to prioritise content in a sensible way rather than making it up as you go along, which is the approach more commonly taken than anyone admits.

Start every keyword research session with competitive analysis. Extract the keywords your top three competitors are already ranking for and use those as your foundation. Build on them with tools like UberSuggest. Then gather backlink data on the top three ranking articles for each keyword and use this to prioritise: focus on keywords with lower competition, good traffic potential, and medium to high buyer intent.

On the subject of tools: you need either SEMrush or Ahrefs. The subscription cost, roughly ninety-nine dollars per month billed annually, is justified by the time they save on research and auditing. However, and this is an equally important point, don’t overdo it with SEO tools. There are hundreds available, and if you’re the kind of person who accumulates SaaS subscriptions the way other people accumulate books they’ll definitely read someday, this is a live risk for you.

For organic SEO, the core stack is: SEMrush or Ahrefs, Screaming Frog, RankMath or Yoast SEO, and whatever outreach tool you prefer (such as snov.io). Two optional additions worth trying: Surfer SEO for content briefs and on-page optimisation, and ClusterAI for simplifying keyword clustering. That’s the list. Resist the temptation to extend it indefinitely.

One other thing about metrics: don’t get sidetracked by vanity metrics. Traffic graphs that go up and to the right are visually gratifying, and there is nothing wrong with enjoying them. But they are meaningless if the traffic doesn’t convert to your actual product or service. What you care about is whether SEO is affecting your bottom line. Stay focused on that, and you’ll be ahead of a significant proportion of people doing SEO.

If you’re struggling to make SEO work and can’t identify why, hire an expert. An SEO consultant or agency has been there before, probably more than once, and may catch issues that aren’t visible to someone without that pattern-recognition. This is not a failure. It’s rational allocation of resources.

The community, incidentally, is worth engaging with. SEO professionals are generally helpful and accessible. Facebook groups like SEO Signals Lab, SEO & Content Marketing, and The Proper SEO Group will get you a dozen useful answers to a well-framed question. Google’s own Search Central blog is essential reading for staying current with algorithm changes, of which there are many, and some of which have the potential to significantly revise your understanding of what is working.

Finally, once you’ve earned rankings on your best keywords, consider running PPC ads against those terms as well. Users who see your ad first are measurably more likely to click your organic listing. It amplifies what you’ve already built, at a cost that tends to be substantially lower than buying the traffic outright.

 

Local SEO: Different Rules Apply

If you’re doing local SEO, your focus should be service-based landing pages rather than content. An accounting firm in London needs pages targeting /accounting-firm-london/ and /tax-accounting-london/ rather than a globally targeted article on financial accounting theory. Ranking nationally for “financial accounting” will not generate a single local client. It will generate traffic from students writing essays and curious professionals who will never become customers, which is an excellent way to get impressive numbers on a dashboard that mean absolutely nothing to your business.

Local SEO has its own rules, several of which contradict general SEO advice. It’s worth studying separately rather than assuming the general principles transfer cleanly.

 

Content and On-Page SEO: Writing Things That Actually Rank

Aim to write articles that are roughly fifty percent longer than the best-ranking article on a given keyword. This is a general rule rather than an absolute one. In highly competitive niches where top-ranking articles are already exhaustive, going longer for its own sake adds noise rather than value. There is a natural ceiling for certain topics, and going past it doesn’t help anyone.

The honest counterbalance to “longer is better” is that shorter is sometimes correct. Some queries want brevity. If someone searches for how to tie a tie, they want a short, clear guide. They do not want “The Ultimate Guide to Tying a Tie: 11 Best Tips and Tricks for 2022,” which exists, I suspect, somewhere on the internet right now, and which nobody has ever read to completion. Know what your reader actually needs and deliver that, not a word count.

Written content is not always the right answer. For procedural topics where the user needs to see something done, video will significantly outperform text. Someone learning how to deadlift correctly is almost certainly looking for a video. Text will rank, but it will struggle to serve the user as well as the medium they actually want.

For the mechanics: follow basic optimisation practices on every page. Include the keyword in the URL, use correct heading hierarchy, add meta descriptions, deploy internal links properly. RankMath or Yoast handles the checklist automatically. Create proper content outlines for your writers that specify the target keyword, related keywords, the article structure, and the title. Sending a writer a keyword and asking them to knock themselves out is a reliable way to get content that misses the point. Outlines are a small investment that dramatically improve the quality of what comes back.

When hiring writers, look for people who specialise in SEO content specifically, not just good writers generally. Excellent general writers frequently struggle with SEO content because the constraints are unfamiliar and the optimisation instincts take time to develop. Beyond that, look for writers who either have direct niche knowledge or, failing that, who have demonstrably explained technically complex topics in the past. A writer who has covered cybersecurity can probably tackle your niche. A writer whose entire portfolio is travel content probably cannot, or at least not quickly.

Write at the right level of assumed knowledge for your readers. A post on advanced tax strategy doesn’t need to define what an income statement is. A post on income statements needs to assume nothing. Get this calibration wrong in either direction and you’ll lose the reader you were trying to reach.

Two tools worth keeping in your writing workflow: ProWritingAid catches structural errors and is free at the basic tier, and the Hemingway App pushes you toward readability. Neither replaces a human editor, but both catch things that human editors miss when they’re on their fourth piece of the day.

Write compelling headlines. A headline with a number, a clear outcome, and a year where relevant will consistently outperform a generic descriptive title. Compare “101 Productivity Tips [To Get Things Done in 2022]” with “Productivity Tips Guide” and ask yourself which one you’d click. The answer should inform every title you write from here on.

Format your blog posts to be scannable. Avoid walls of text. Use the structural and visual elements the format supports. And on the subject of images: use relevant images. Relevant is doing a lot of work in that sentence. Not stock photos of office workers smiling at laptops in environments that no actual office has ever resembled, but graphs, charts, screenshots, and diagrams that support the content you’ve written.

The Skyscraper Technique is worth understanding and applying, but only to content where it actually makes sense. Pick topics that are genuinely important, where you’re creating something meaningfully better than what exists. Applying the Skyscraper to mediocre content topics is a waste of everyone’s time, including the time of the internet, which is otherwise quite busy.

For seasonal content, keep the year out of the URL slug and update the headline annually. Keep it /saas-trends/ rather than /saas-trends-2022/, change the headline each year, and preserve the accumulated authority of the single URL rather than starting from zero with each annual edition. If you want a clean record for each year’s report, then yes, include the year in the URL. Pick one strategy and stick to it.

Don’t write two articles on the same topic. Content cannibalism confuses Google about which page to rank and generally results in neither ranking particularly well. Pick one, make it excellent, and update it rather than creating a companion piece that competes with it.

Don’t overload articles with outbound links. Including sources is good. Including twenty outbound links in a thousand-word article begins to look like spam, even when the links are all legitimate. There is a point at which helpfulness tips over into the appearance of something else entirely.

Pay attention to the “People Also Ask” section in search results and optimise for the Google Featured Snippet. Both represent additional SERP real estate you can claim. Make a list of PAA questions related to your topic and ensure your article addresses them, either naturally within the content or in a dedicated FAQ section at the end.

Reverse-engineer viral content in your niche. Look at what has performed well on Reddit, Hacker News, Facebook groups, and Buzzsumo. Create something in that territory that is significantly better. This is not plagiarism; it is understanding what your audience responds to and doing it well.

Review and update your top-performing articles annually. Rankings are not permanent. The article that earned the top position last year will eventually be displaced by someone who one-upped it, and that process accelerates if you never revisit the content. Most companies, having finally reached position one for a keyword, leave the article entirely alone. Then they get outranked and are surprised. Check your top performers once a year and improve them wherever possible.

Finally, experiment with your click-through rate by testing different headlines for articles with low CTR. What constitutes a good CTR varies significantly by keyword and position. Sometimes position one drives fifty percent of the traffic. Sometimes it’s less than fifteen percent. Test, measure, and adjust.

 

Link Building: The Part Everyone Gets Wrong In Different Ways

Do you need backlinks to rank? The answer, like the answer to most SEO questions, is that it depends on your niche. In low-competition niches, superior content can rank without backlinks. In high-competition niches such as VPN, insurance, or financial services, content quality is the table stakes and links are what separate the first page from everything below it. Your competitors all have good content. What they don’t all have in equal measure is backlinks, and that’s where the ranking difference lives.

In some niches, paying for links is unavoidable. Gambling, CBD, and similar categories operate this way. In those cases, you need either a significant link-building budget or a very creative approach: viral infographics, newsworthy research, data-backed content that earns coverage organically.

My favourite link-building principle, and the one most consistently ignored, is to build relationships rather than links. The difference between spamming a hundred sites with “hey, can you link to my article?” and developing genuine relationships with a dozen sites in your niche is the difference between a two percent reply rate and a sustainable link-building operation that compounds over time. Spam damages your reputation, which makes all subsequent outreach harder. It’s a poor investment even when it technically works.

The most effective link-building tactics are the most straightforward: direct outreach, broken link-building, guest posting, the Skyscraper Technique, creating viral content, and guest posting with infographics. Creativity matters in the execution, not in inventing entirely new approaches. The best link-builders, incidentally, never write publicly about what they’re doing. If they did, half the internet would copy it within a week, the tactic would become cliché, and the advantage would evaporate. Be creative in your own execution and experiment until you find what works.

When doing outreach, give something in return. Free access to your tool, a reciprocal link from the guest post, something of value to the person you’re asking. This substantially improves reply rates because it turns a request into an exchange, which is how human interaction tends to work in most contexts.

Two categories to avoid entirely: link resellers and Fiverr. The person messaging you on LinkedIn with a Google Sheet full of links to purchase is offering you private blog network links and a pathway to a Google penalty. The links sold on Fiverr are, without meaningful exception, useless. Touching nothing from either source is the correct approach.

Not all links are equal. Quality links come from pages with genuine backlinks of their own, low outbound link counts, high domain authority, and topical relevance to your content. A link from an accounting blog to your article about pet care will strike Google as suspicious, because it should be suspicious. It’s the kind of thing that doesn’t happen organically, and Google has become quite good at noticing.

Data-backed content generates links organically and at scale. OKCupid used to publish research based on how people interacted with their platform, and it never failed to go viral. Each report was covered by dozens of media outlets, generating links they didn’t have to ask for. Original data and genuine research are among the most powerful link-acquisition tools available, and among the most underused.

HARO (Help a Reporter Out) is a platform that matches journalists with expert sources. You receive a daily email with journalists actively looking for contributors in specific niches. Pitch them correctly and they may feature you or link to your website. It is legitimate, underused, and works.

No-follow links are not useless. Google treats the no-follow attribute as a suggestion rather than an instruction, and documented cases exist of no-follow links contributing to improved rankings. Similarly, reciprocal links are not the black-hat horror they’re sometimes described as. One or two link exchanges with relevant sites in your niche is entirely harmless in the vast majority of cases.

Buying an expired domain with an existing backlink profile in your niche can give a new website a meaningful head start. It is worth investigating if you’re launching something new and want to avoid starting from zero.

Sponsored links marked with rel=sponsored don’t pass PageRank. Don’t pay for links from Forbes or Entrepreneur expecting a rankings boost. They won’t deliver one. The only person who benefits from that transaction is the one selling the link.

Promote your content organically. Posting your best pieces in relevant Facebook groups, on LinkedIn, and on Reddit generates traffic and, as a pleasant side effect, organic links from people who find the content useful enough to reference elsewhere. I’ve seen a single piece of content promotion generate close to a hundred backlinks this way, purely because enough people encountered the article and linked to it from their own sites.

Expert roundup articles deserve a mention here. If you want to build relationships with influencers in your niche, write an article featuring them and tell them about it. People are remarkably enthusiastic about being recognised, and this kind of outreach converts at a substantially higher rate than cold link requests.

.edu links are over-hyped. According to Google’s own John Mueller, .edu domains carry large numbers of outbound links, and Google ignores a significant proportion of them as a result. Pursue them if they arise naturally; don’t treat them as a priority.

If you’re a SaaS company, your customers are a link-building resource you’re probably not using. Build relationships with customers who operate in the same topical niche and help each other build links. It’s a legitimately unusual approach and legitimately effective.

Finally: don’t spam outreach for every article you publish. Reserve it for your most substantive content. Nobody is going to respond with enthusiasm to a request for a link to your “Top 5 Benefits of Adopting a Puppy.” The more indiscriminately you send outreach, the less effective each individual email becomes. Save the asks for when they’re worth making.

 

Technical SEO: Getting the Infrastructure Right

This section is long because there is a lot of ground to cover, but most of it is common sense made explicit. I will move at pace.

Run PageSpeed Insights on your website. While load speed isn’t technically a direct ranking factor, a page that takes five seconds to load when your competition loads instantly will experience higher abandonment rates, which affects you indirectly through user behaviour signals. Your crawl depth should be below four, meaning any page on your site should be reachable within three clicks from the homepage.

Use next-generation image formats wherever possible: WebP, JPEG 2000, JPEG XR. They compress significantly better than PNG or JPG. Lazy-load images on content-heavy pages so that images below the visible screen area aren’t loaded until the user scrolls down to them. Enable Gzip compression for your HTML, CSS, and JavaScript files. If your site is loading slowly because you’ve accumulated a hundred-plus external JavaScript files and stylesheets, minify, aggregate, and inline where possible. These are foundational performance improvements and there is no good reason to defer them.

De-index pages you don’t want Google crawling, such as internal app pages or knowledge base content, via your robots.txt file. Your site must be mobile-friendly; Google uses mobile-first indexing and there is no scenario in which this is optional. Install an SSL certificate if you haven’t already. It’s a direct ranking factor, it keeps your site secure, and an unencrypted HTTP website has no business existing in the current decade.

Use rel-canonical for duplicate content to indicate which version is authoritative. When permanently migrating pages, use 301 redirects to pass link equity from the old URL to the new one. Temporary redirects get a 302. The distinction matters for whether accumulated link value transfers, and getting it wrong can cause ranking drops that take months to recover from.

When A/B testing two page variants, use rel-canonical on both to indicate which is the original. Do not use AMP unless you’re a media company. For most websites, it creates more problems than it solves.

When linking to internal pages, use the target keyword as anchor text. This signals to Google what the linked page is actually about. Use Google Search Console to verify all your blog posts are properly interlinked, particularly new ones. Internal linking has a genuine impact on rankings and is consistently underinvested.

Bounce rate is not a Google ranking factor. A high bounce rate on some pages is entirely natural: recipe pages, simple how-to guides, pages that answer a specific question and are then closed because the user has what they came for. Don’t optimise for bounce rate as a proxy metric when it’s not measuring what you think it’s measuring.

Google will frequently ignore your custom meta description and substitute its own. Write one anyway. On the pages where Google does use it, it matters for click-through rate. The overhead of writing them is low; the downside of omitting them occasionally is real.

Monitor your backlink profile and disavow obviously spammy or PBN links. Google usually ignores them, but a competitor deliberately building toxic links pointing at your site can create problems, and disavowal is the mechanism Google provides for addressing this.

Keep URL slugs short and descriptive. Do not include dates in your URLs. An outdated date in the URL reduces click-through rates because readers prefer content that appears to be current, even when the underlying quality hasn’t changed. The URL is part of the first impression; make it a good one.

Social signals affect your rankings, but not in the way most people think. Shares and likes don’t directly move the needle. What does: your article goes viral, people search for it on Google, click through, and read it. That behaviour pattern signals relevance to the algorithm. The mechanism is indirect but real.

Crawl your site with Screaming Frog every couple of months to identify broken links, 404 errors, and other technical problems before they accumulate into something larger. Use WordPress as your CMS unless you have a compelling reason not to. For ninety-nine percent of websites, it remains the right choice. Its plugin ecosystem is unmatched, and the alternatives are almost universally worse for SEO in practice.

When checking article rankings in Google Search Console, check both Page and Query. Checking Page alone gives you the average ranking across all queries that page appears for, which is almost always misleading. You need the Page and Query combination to get the accurate picture for a specific keyword.

 

In Conclusion

I said at the start that a large proportion of SEO advice you encounter on the internet is based on personal experience, experiments, and data of questionable provenance. Everything I’ve written here falls into the “personal experience” category as well. It’s what has worked for me, drawn from years of doing this across different niches and circumstances.

If something in this post contradicts what you’ve found to work for you, that’s fine. SEO has never had a single universally correct answer. The algorithm is a moving target, the competitive landscape varies enormously, and what works in one niche may actively harm you in another. The only honest position is to keep testing, keep measuring, and keep updating your model.

If you’ve found a better way, I’d genuinely like to hear about it. You can reach me at [email protected].

All Posts