Digging through the archive I found a post about optimizing your website for search engines. No byline I could trace, no clear date, just sitting there quietly giving advice to whoever wandered in from Google.
It is, without exaggeration, one of the most confidently wrong documents I’ve ever read. Not because whoever wrote it was careless. It was solid advice when they wrote it. That’s the interesting part. Every line of it made sense once, and now most of it would actively hurt you.
So rather than delete it, here it is with commentary. The original text is quoted. My notes follow each bit.
Search engine optimization is crucial to your traffic flow. Having your website listed on the first page of a search engines results will give you maximum exposure. And, the closer you are to the top of the page the greater the number of visitors you will receive.
Still broadly true, but the ceiling has moved. Ranking first used to mean you got the click. Now an AI Overview often answers the question above you and the user never scrolls. You can hold position one and watch your traffic fall anyway. “First page” was a finish line back then. Now it’s a qualifying round.
Not all search engines operate exactly the same. But, if you follow the simple outline below you will do just fine. This will give you an overview of what most search engines are like and what they are looking for.
This sentence dates the post harder than anything else in it.
There was a genuine plural field once. AltaVista, Lycos, Excite, Ask Jeeves, Inktomi, and they really did behave differently enough that you optimized for each one. Writing “most search engines” in 2026 means Google, and then a rounding error.
Although here’s the twist. The field went plural again, just not the way anyone expected. ChatGPT, Perplexity and Gemini all pull from the web and they do not agree with each other about who to cite. So this line came back around to being true, for entirely different reasons.
- Are your keywords in your title tags? (This includes the filename and page title)
Fine. Still do this. Just write the title for a human first and stop cramming.
- Are your keywords in your content? (Especially the first and last paragraph of each page)
Half right, wrong reason. Front loading your main point matters enormously now, but not because a robot counts words at the top. It’s because AI systems lift most of their citations from the opening chunk of a page. Bury your answer under 400 words of preamble and you don’t get quoted.
The last paragraph thing is dead. Nobody weights your closing paragraph.
- Do your keywords accurately describe the theme or subject of your page?
This one aged beautifully. It’s basically asking “is your page actually about what you say it’s about,” which is the entire foundation of how search works now. Google stopped matching strings and started understanding topics. Whoever wrote this was closer to the future than they knew.
- What words will users type into search engines to find your website?
Still the right question. The answers just got longer. People type sentences now, and increasingly they don’t type at all, they ask. Your headings should look like questions, because that’s what people are actually asking.
- Does the page contain JavaScript? If so, how much of it precedes content? (CSS and JavaScript is nice, but engine robots don’t have a clue as to what to do with it and often give up if there’s no content nearby.)
Completely obsolete, and I love it.
There was a real era where a crawler hit a wall of inline JavaScript and just gave up. Googlebot renders JavaScript now. Entire frameworks exist that ship almost nothing but JS and rank fine.
The underlying worry did survive though. It just moved. Nobody cares where your script tags sit, but they very much care how long your page takes to become usable, which is what Core Web Vitals measures. Same anxiety, new metric.
- If your website is a business, have you made sure your business address and other information is easily available on each page? (Directories look for this when reviewing business sites.)
The advice is right, the reason is extinct.
“Directories look for this when reviewing business sites” is a sentence from the era of human editors at Yahoo and DMOZ deciding whether you deserved a listing. Actual people, reading your site, deciding your fate.
Consistent name, address and phone across the web still matters a lot for local search. But no human is reviewing you. Google is matching entities across sources and quietly deciding whether you’re the same business it saw somewhere else.
- Did you optimize several pages of your site for certain targeted keywords? Be sure that your page titles reflect your keywords and be sure to repeat them in the body text of the page.
“Be sure to repeat them in the body text” is where this goes wrong. That instruction, followed enthusiastically, produced two decades of unreadable content. It’s the direct ancestor of every article that says “best running shoes for flat feet” nine times in six paragraphs.
Worse, targeting several pages at the same keyword just makes them fight each other. Merge them into one good page.
- Are there plenty of text links on your pages for robots to follow? (They don’t follow image map links or options in drop down menus.)
Image maps. Genuinely delightful.
Crawlers handle normal navigation fine now. But the principle survives in a form that matters more than it used to. Your internal linking is how you tell search engines and AI systems which of your pages are important and how they relate. Most sites are still terrible at it. Old posts sit orphaned with nothing pointing at them, slowly rotting.
- Did you read each engine FAQ to make sure you avoid SPAMMING their database before you submitted your URL? (FAQ’s change often, so you need to keep checking back.)
This is my favorite line in the whole document.
You used to submit your URL. Manually. To each engine. There was a form. You filled it in and waited, sometimes weeks, to get indexed. And you read their FAQ first so you didn’t get banned for doing it wrong.
All of that is gone. Google finds you through links and sitemaps. There’s no form, no queue, no FAQ to check.
- Do you have any web sites linking to your website? Links from other web sites similar to your own will increase the relevancy of your website in search engine results.
Correct then, correct now. Links still matter. Relevance still matters more than volume. This one needs no editing at all, which after nine other entries is quite something.
Focus on your keywords and keyword density but don’t sacrifice your message. Utilize the meta tags keywords, description and title. Use your keywords when naming your pages.
Two dead things in one short paragraph.
Keyword density was never real in the way people believed. There was no magic percentage. But the myth was so widespread that tools got built to calculate it and writers genuinely tuned their prose to hit a number.
The meta keywords tag is properly, verifiably dead. Google said publicly in 2009 that it doesn’t use it for ranking. Bing has indicated it can be a spam signal. If your CMS still has a field for it, leave it empty.
Meta description and title tag are both alive and worth your time. Just note that Google rewrites your meta description most of the time anyway, so treat it as a suggestion.
One of the biggest things that search engines look for, after content and theme, are quality links to your website from reputable internet addresses. Just any links won’t do. They must be roughly the same subject matter and rank well on the search engines. This can be one of the most difficult parts of the optimization process but if your site is interesting, full of rich content and you keep growing your website they will come.
This paragraph could run today with barely a comma changed.
“Just any links won’t do” predates the entire link spam industry that Google spent fifteen years dismantling. Relevance over volume was right the whole time. The last sentence is a bit optimistic, links don’t just come, but the sentiment holds up.
Remember that while search engines are a primary way people look for web sites, they are not the only way. People also find sites through word-of-mouth, traditional advertising, the traditional media, newsgroup postings, web directories and links from other sites. Many times, these alternative forms are far more effective draws than are search engines.
And then it ends on the single most relevant paragraph in the document.
An entire industry proceeded to ignore this for twenty years and build businesses that depended completely on one referral source. Everyone who did that has spent the last two years discovering what happens when that source changes its mind.
Swap “newsgroup postings” for whatever you like and this is current advice. Better than current, honestly. Most SEO writing today still won’t tell you plainly that search shouldn’t be your only channel.
What Actually Survived
Ten tips, one long post, and here’s what’s left standing:
Know what your page is about. Know what people would type to find it. Get relevant sites to link to you. Don’t depend entirely on search.
That’s four ideas out of maybe fifteen. Roughly a 25% survival rate over twenty years.
The stuff that died shares a pattern. It was all mechanical. Density percentages, tag stuffing, submission forms, script placement. Instructions you could follow without understanding anything, which is exactly why they were popular and exactly why they broke.
The stuff that survived is all judgment. Is this page genuinely about what it claims? Would a real person want to link to it? Those questions can’t be automated, so they can’t be gamed into uselessness, so they’re still here.
Which is worth sitting with, because the current wave of advice has the same split running through it. Half of what people are shouting about right now is mechanical: schema markup patterns, prompt formats, the exact word count for a summary box. That half will read like image maps in ten years. The rest, being genuinely useful and being specific enough to quote, will still be here.
Worth remembering when you read anything about optimizing for AI, including things I’ve written. It’s also why so much AI generated content sounds identical. It’s optimizing for the mechanical half, and the mechanical half is always the part that expires.
Anyway. Twenty years on, the last paragraph of a forgotten post in an archive turned out to be the most useful thing in it. Don’t put everything into one channel. Somebody knew that before Google finished winning.