Ask a room full of SEOs whether something is a best practice, and there’s a good chance everyone will agree before anyone asks the obvious question: Why?
We’ve accumulated plenty of these recommendations over the years: Use one H1. Keep title tags under a certain character count. Put your target keyword in the title, H1, and first paragraph. Improve your Core Web Vitals until the report turns green. Use schema. Create fresh content. None of these recommendations came out of thin air.
Many were based on reasonable observations, testing, or guidance that was relevant at the time. Some still are.
The problem comes when an idea gets repeated long enough. The reasoning disappears, the recommendation remains, and “this has historically been useful to consider” becomes “this is what you have to do.”
That’s how SEO best practices become dogma.
How ‘best practices’ are born (and why they stick)
Most SEO best practices start with a reasonable observation. Someone notices a pattern, runs a test, shares a successful tactic, or follows search engine guidance.
The idea gets repeated in articles, conferences, tools, and audit templates. Over time, “this worked under these conditions” becomes “this is how SEO works.”
That’s how a recommendation becomes a checklist item.
Once it does, questioning it can feel irresponsible. A crawler flags 2,000 pages with multiple H1s, so fixing them feels productive. The issue is visible, the recommendation is familiar, and there’s even a number attached to it.
But what problem are you actually solving?
Track, grow, and measure your visibility across Google, AI search, social, local, and every channel that influences buying decisions.
The difference between ‘still true’ and ‘still useful’
Someone will question a long-standing recommendation, and someone else will immediately respond: “Well, that’s still a best practice.”
Maybe. But something can still be technically true without being useful enough to prioritize.
Take Core Web Vitals, for example. We all know a slow page can create a poor user experience. Nobody’s arguing otherwise. But that doesn’t mean every URL sitting outside a recommended threshold deserves an engineering project.
If a high-traffic product or conversion page is slow enough to affect user behavior, that’s worth investigating. If a low-traffic resource page misses a benchmark by a fraction of a second, the business case may be much harder to make.
Yes, clean HTML is generally preferable to messy HTML. Descriptive headings can help communicate page structure. Image alt text is important. Page speed matters. The question isn’t whether these things are good. The question is whether fixing a particular issue is the best use of your time right now.
That’s a much harder question because it requires context.
A publisher dealing with delayed JavaScript rendering may have a serious discoverability problem. An ecommerce site with thousands of duplicate URLs created by faceted navigation may be wasting significant crawl resources. And a service business with three H1 elements on a handful of pages styled differently by its CMS may have absolutely nothing to worry about.
The recommendation is the same. The impact isn’t.
And that’s why “best practice” has become an increasingly inadequate way to prioritize SEO work. We can’t optimize for compliance over outcomes by checking boxes instead of asking harder questions about impact.
Dig deeper: Stop saying ‘best practice’ and start bringing proof
How to decide which SEO work matters
A better way to escape SEO dogma is to change the questions we ask.
1. What evidence suggests we should change what we’re doing?
This should be the starting point for more SEO work.
Google changes constantly. Search changes constantly. Someone will always have a new theory about what matters. That doesn’t mean every announcement requires a response.
Before changing a strategy, rebuilding a page template, or assigning development resources, look for evidence that the issue is affecting visibility, traffic, discoverability, or business goals.
Sometimes the right response to a change is action. Sometimes it’s observation. Occasionally, the right response is to leave things alone.
That may be the least marketable SEO advice in history. It’s still advice more teams need to hear.
2. What problem does this solve?
“Best practice” is often used as a substitute for explaining why something should be done.
My favorite example is the familiar “multiple H1 warning.” A tool flags multiple H1s and labels them as a Warning of Medium-high priority. The recommendation says to fix them, so they get added to a backlog of site update requests.
OK. But what problem are we solving?
Is the page structure genuinely confusing? Is important content being communicated poorly? Is the HTML preventing search engines from understanding the page?
Or does the page simply have multiple elements that happen to use an H1 because of how the CMS and design system were built?
Those situations shouldn’t automatically receive the same response.
Before fixing something, define the outcome you expect. If you can’t explain the potential outcome beyond “the tool says this is an error,” the issue may deserve another look.
3. How much of the site do those issues actually affect?
SEO audits are very good at making problems look large. A crawler can tell you that a site has 10,000 issues, and that can certainly sound alarming. But issue counts rarely tell you anything about impact.
Ten thousand missing alt attributes on legacy images with little search value may have limited SEO consequences. One broken rendering implementation on a high-value page template can affect thousands of URLs. One canonicalization problem can create more damage than an audit full of warnings.
The number of issues found isn’t a KPI. Neither is the number of issues closed. Scale matters, but only when the affected pages, templates, or user journeys matter.
4. What happens if we don’t?
What’s the actual risk of leaving the issue alone?
- Will important pages fail to get indexed?
- Will search engines struggle to access meaningful content?
- Will users have a worse experience?
- Will organic visibility likely decline?
Or will the website continue to perform exactly as it does today?
Imagine two recommendations. The first is adding meta descriptions across a ton of legacy blog pages that receive almost no search traffic and aren’t central to the site’s content strategy. The second is misconfigured canonical tags causing search engines to ignore the canonical directives entirely, leading to cannibalization.
Both might show up in an audit. Both could technically be called SEO issues. Only one has an obvious connection to discoverability and a tangible consequence to inaction.
Not every imperfect implementation represents an opportunity cost. Sometimes a fix is simply a fix looking for a problem.
Dig deeper: 10 technical SEO audit mistakes that lead to bad recommendations
5. Does the reason it worked apply here?
SEO advice is often presented as universal because it’s easier to package. But websites aren’t interchangeable.
A publisher and an ecommerce site can have the exact same technical SEO warnings and completely different priorities. For a publisher, crawl efficiency and content freshness might matter enormously. For ecommerce, duplicate URLs, faceted navigation, or product availability states can create problems across thousands of URLs. A local service business may have far fewer pages and far more to gain from fixing its information architecture or conversion paths.
Someone else’s solution doesn’t automatically belong on your website.
6. Are we solving the problem that brought someone here?
For years, SEO strategy has often started and ended with the SERP. What keyword do we want? Who ranks? What do they have that we don’t?
That analysis still has value. The problem is what happens when competitive analysis becomes competitive cloning. If everyone looks at the page ranking above them, adds what they’re missing, and calls it optimization, every result eventually starts looking suspiciously similar.
The better question is what the searcher is actually trying to accomplish. Sometimes:
- That means answering the obvious question.
- It means giving them the comparison they haven’t found elsewhere.
- The opportunity is a better tool, clearer explanation, or resource that helps them take the next step.
The goal shouldn’t be to reverse-engineer every feature of the page currently ranking above you.
Google’s results can tell you what already exists. They don’t always tell you what users still need.
7. Where is our audience searching when they decide who to trust?
For a long time, the SEO conversation has centered on convincing Google that a website deserves visibility.
People now encounter brands through more sources than ever. They read recommendations from publications. They ask questions in communities. They watch videos. They use social platforms as search engines. They receive answers synthesized by AI systems that may draw on sources far beyond a company’s own website.
A brand’s website remains important, but visibility increasingly depends on the broader footprint surrounding it. That should change how we think about what deserves investment.
An expert contribution to an industry publication can put the brand in front of people who may have never found its website through Google. Original research can earn citations and create authority beyond the website where it was first published. Even social channels can provide something SEO reports often can’t: unfiltered language from people describing their problems in their own words.
These signals are reminders that traditional search optimization was never supposed to become a substitute for understanding how people discover information.
8. What are we already doing that isn’t performing?
The SEO industry has a talent for accumulating tactics. A new recommendation appears, so it gets added to the process. Then another, and another.
Eventually, teams are maintaining an enormous collection of “best practices” because removing something feels riskier than keeping it.
But every task has a cost. Every low-impact optimization consumes time that could have gone somewhere else. Every unnecessary development ticket competes with work that may actually affect the business. Every piece of content created because a keyword exists is content that someone has to maintain.
Sometimes the smartest SEO decision isn’t adding a new tactic. It’s retiring an old one.
Dig deeper: Technical SEO testing: How to build a stronger experiment
See where your brand appears, where it doesn’t, and exactly how to win more visibility across search, AI, local, social, and every channel that matters.
Maybe the only real best practice is ‘think for yourself’
None of this means best practices are useless. They got us here, and they still have a place. But we have to ask if they still deserve a place on the list.
Our job is to figure out which changes actually matter. That means being willing to question the things we’ve always done. Sometimes, it means admitting that the thing we’ve always done worked before, but doesn’t anymore.
And it means building things that have value beyond a single search engine: useful content, recognizable brands, strong products, good customer experiences, credible expertise, and relationships beyond a single platform.
That’s not particularly revolutionary advice. It is, however, much harder to package into a neat checklist. Maybe that’s exactly why we need more of it.
https://searchengineland.com/when-seo-best-practices-become-seo-dogma-487718