Michael Martinez over at SEO Theory just wrote another outstanding article on technical SEO, what is SEO, and what are ranking signals. Most of all, it’s about what are not technical SEO tasks or SEO tasks at all, and what are not ranking signals. It’s called, “Is There a Single Most Important Technical SEO Task?” There’s a lot of critique, humor, insight, and education to be had there. I started out writing a bullet-point summary of it and it turned into this thing – several lists followed by my own ranting thoughts on some of the ideas. I hope this advances the conversation somewhat.
On SEO Standards, Definition, & Metrics
- There are no SEO standards.
- There is no SEO manual.
- Technical SEO cannot be defined.
- SEO cannot officially be defined.
- The universal answer to any question in SEO is “it depends….”
- SEO hasn’t changed in 20 years: It’s still about links and content.
- How we define SEO has changed merely.
- Michael Martinez’ current definition of SEO:
“The practice of analyzing search engine protocols, actions, resources, and guidelines for the purpose of improving Website compliance and performance in search results.”
- Most “SEO activity” is inefficient – tasks that waste time and money.
- There are no reliable SEO metrics. (Google has denied using “domain authority.”)
SEO Does Entail:
- Publishing content
- Getting links for it
- Making sure that search engines find the content
- Making sure search engines see the links to the site
- One ranking signal has been publicly documented ever: PageRank (implied).
- How PageRank is calculated today we don’t know.
- Lists of ranking factors are snake oil – there is no real list of ranking factors.
- HTTPS is not a ranking factor if you don’t know how much it affects ranking.
- There is no optimal website design – a huge range of website designs rank well.
These are Not SEO Tasks
- Link research – Finding links to help with Google rankings isn’t SEO because you don’t know which links help rankings, or when they stop helping.
- Creating relationships with influencers – again, you don’t know which links improve search results.
- Defining site user personas.
- Website design – because there’s no optimal design, SEO can merely enhance website design.
- Improving page speed – only 1% of websites need to improve performance per Google, so it doesn’t optimize for search. (Sidenote: prefetches don’t improve page load speed; they add unnecessary processes to Web servers.)
- Following industry news.
- Managing bounce rates – Google and ex-Googlers say that click rates and bounce rates are not used to determine rankings.
- Adding an XML Sitemap to your robots.txt – It isn’t required for SEO although it’s a good move (you could also submit it to Google via Search Console, or even embed a link to a sitemap on a Web page).
- Writing ridiculously long blog posts.
My Thoughts on All This
I’m known for occasional rants, so I love a good rant and this is a great and thoughtful one. I think the article is brilliant and packed with insights. I agree with the standards, definition, and metrics. Although regarding the argument that SEO hasn’t changed in 20 years… I’d say: In broad terms, SEO has not changed, the definition changes depending on whom you talk to, as Michael Martinez explained. But, I’d say that some of the details and some of the methods for doing SEO have changed.
Yes, we still publish content to rank. But 8 years ago, shitty content would more routinely rank than it would today. Content standards have shifted and how SEO content is created has shifted. Content farms don’t rank as reliably or as long now. How the content is written (as part of SEO) has changed somewhat to be more natural. Adding video and structured data can anecdotally improve rankings now and I believe that was not the case 20 years ago. The content should not be full of duplicate crap today, and canonical tags can help with that today.
How people “get links” has made a shift as well. Michael points out that we don’t know what links improve rankings. However, a lot of research has been done to demonstrate what types of links can do harm… and that seems to have changed as well in relation to various Google penalties.
Ranking Signals Need Not be Proven to be Strongly Circumstantial Correlations
I love his points on ranking signals, and that there are no definitive ranking signals. The article repeats in many ways that we don’t know specifically what helps rankings, so anything that cannot be definitively measured should not be considered to help: “…you are just throwing stuff at the wall and hoping that something sticks.” Hah, well said, but of course nobody knows definitely what helps. That’s why there are a bazillion articles by folks in the industry trying to research and figure that out and prove hypotheses. Are any of them definitive proof? Perhaps not, but there’s plenty of strong hypothetical evidence regarding what may and may not be ranking factors. While many long articles are just self-serving fluff, many writers in our industry do put in diligent work to attempt to show cause and effect.
As for those items that he submits are not SEO tasks….
Link Research Can Help SEO
I think that link research does contribute towards SEO. Of course we don’t know which links help rankings. But we do know that dubious links (as defined by Google) will hurt Google rankings, and we have seen correlations between increasing quantity of links and ranking improvements. So attempting to research which links might be helping your competitors or such is likely to contribute positively towards improving rankings if done thoughtfully.
Personas May Help Keyword ID
I don’t advocate spending a lot of time on persona research, but it can be helpful to your SEO keyword research to know your audience, particularly for business-to-business SEO which needs to reach out to different roles within a company. I see no waste in an SEO spending limited time brainstorming which types of key terms may be searched for by different interested parties. In some scenarios, this will surely improve your keyword research.
Web Design Testing Can Help SEO and Conversion
Agreed, web design isn’t directly SEO. But as Michael notes, crawlability is part of SEO, and that is part of good web design. Also, we all know that a lot of websites just suck for users. Or suck on mobile. And if a site sucks for users, why would search engines want to rank them well? Google reps often emphasize the importance of having great content and building websites that users find easy to navigate.
Even if we don’t know specifically what Google is looking for in a functional website, we do know that Google cares about usability in some form or another. That’s where testing comes into play. If we advise a designer to make some changes to a page and we can measure strong rankings improvements related to that page shortly after that, then we can postulate with good odds about cause/effect performance impact. And if we’re apparently but not definitively improving rankings, I consider that SEO work.
Page Speed, Influencer Relations, Industry News
- On page speed, I appreciate the reminder that this is only important for about 1% of websites.
- About creating relationships with influencers, see my take about links above.
- Following industry news is not doing SEO, right, but it sure can help us do SEO… or it can kill a lot of time on fluff and filler info, depending on the articles.
Bounce Rates & Click Through Testing Can Help Conversion
About managing bounce rates and click-through rates, see above about web design and improving usability. The same sort of testing can apply to meta descriptions and page titles. And site usage metrics can sometimes help to inform better conversion rate optimization. CRO isn’t SEO, right, but as an SEO firm, our main goal really is to help the website’s bottom line – the revenue. So it makes sense to help them convert better too, in addition to driving more traffic. If successful, it will sure help the SEO firm be seen in a better light by its clients.
I hate ridiculously long blog posts myself. But shit, I just passed 1400 words now. That’s a wrap.
I consider these tasks useful to SEO (my additions are in orange):
- Thoughtful keyword research to identify relevant and valuable key terms
- Publishing content that helps and appeals to users while targeting the useful key terms
- Research to find which links may damage rankings and which ones might be helpful
- Try to acquire links for the content (target links that you think may be helpful)
- Review and test site usability and try to improve site usage metrics
- Make sure that search engines can crawl the site and find the content
- Make sure search engines see the links to the site
- Review site usage metrics and paths, then test to try to improve site conversion rate