The complexity of modern day SEO

Is this effective for SEO? It depends…

When talking to SEO’s you’ll find a diverse range of opinions amongst them, ask one a question and you’ll often find the words “it depends” in an answer. It’s a common joke amongst SEO’s but it has real substance in their works, because very often, recommendations will absolutely depend on so many factors, that will influence another

The community is full of people who are more than happy to tell you how they think it is.

Some will parrot what others say, some will do their own research, and some will just make it up as they go along, and chat fluent Turdish.

Excepting those who just make it up as they go along, there are good reasons for this.

For those who parrot what others say, it’s mainly related to how the people they echo, have built trusted reputations over time; people trust what they say, and don’t feel the need to re-invent any wheel so happily follow along. If someone has provided a good guide for how to do a thing, then others will often follow, especially if it’s supposedly tried and tested.

Now you might argue that there’s little wrong with this approach, especially if it works. However, you’d be missing a trick, as would those slavish adherents to their credo. By all means, take the best of what someone is saying, but don’t take it as rote and assume that there’s this one size fits all approach, because there isn’t and through testing and trying alternatives, you might find an edge (you might not either but hey, learning!).

The Diversity of Query Spaces, Niches and Search Engine Results

Niches and query spaces vary in all manner of ways. Query intent can be informational, commercial (transactional) , navigational (branded) and different niches have different types of content and or levels of competition. Not all SERP’s (search engine results pages) are equal.

Informational Queries

Informational queries are used to find information on a specific topic. These queries often include keywords such as “what,” “how,” or “why,” and are typically used when a user is looking for information on a specific subject.

Commercial Queries

Transactional queries are used to complete a specific action, such as making a purchase or booking a reservation. These queries often include keywords such as “buy,” “order,” or “reserve.”

Navigational Queries

Navigational queries are used to find a specific website or webpage. These queries often include the name of a website or brand, and are typically used when a user already knows the name of the website they are trying to find.

Aspects that influence user SERP’s and Queries

In addition, queries and the results returned, are often influenced by personalisation (device, user history, behaviour), geolocation (local, regional, international), competitiveness of the SERP, and the user journey.

Query Personalisation

Suggestions and SERP’s will vary based upon search history or known factors related to age and gender, location, device type (mobile vs desktop), what they clicked on previously and whether they are likely to want to see it again perhaps. Other factors like earnings, social status, disposable income etc may also be at play, but these are also likely to be anonymised, and due to their sensitive nature, downplayed or mentioned obscurely if at all in a public sense.

Geolocation

SERPs will be shaped by region and results compiled of results that are influenced by factors that the algorithm has determined are important. TLD’s, linking patterns, href lang

Competitiveness

How competitive the SERP is, who is there, why are they there, what do they bring to the party and why, who mentions them, what criteria do they meet all influence a competitors presence

User Research Phase

SERPs might be refined based around previous search history, AI’s like MUM and BERT will seek to refine results to give users what they expect and may well fold these factors in to other searches if similar patterns are noticed in sufficient numbers.

So to create, or write a simple piece of SEO content that says in order to succeed in X, just do Y will ultimately fall short, as you need to understand how your content is likely to be interpreted by all of its users in order to rank; users like search engines AND people.

Making Content Helpful

Many people in search will utter the phrase, make content for people and not for bots.

Indeed, the recent so called ‘Helpful content update‘ was focused around just this thing. Which on the face of it seems relatively simple and easy to understand right?

Create good content for the people who are going to use it.

Done, simple, what’s the problem?

But simple it isn’t.

Far from it.

There are many hurdles to jump, many little caveats and nuances to consider.

It isn’t enough to just write a good informative piece that fits the query because these days ( if you can be bothered to go beyond being a nitwit and creating more than some keyword keyword rank me quick please drivel) then, achieving that seemingly good piece of content in itself, is really very very easy, on most topics.

Enter the AI

Anyone with half a brain can ask a Generative Pretrained Transformer to write some copy around a topic, and the AI will produce content that reads reasonably well. A few edits here, an alteration there and you’ll soon have an authoritative sounding piece of content that fits a particular query.

Here’s an example for a query about SEO Services in Kent created in around 30 seconds.

Now, that piece is pretty generic and in a site lacking authority, wouldn’t rank straight out of the traps, as it would lack the other X number of signals that search engines use to elevate content, but it does include words and phrases that make sense to humans and so if returned for such a query, it would make all of the right kinds of noises that someone less informed might expect, and whilst it doesn’t fit any tightly defined model of content made for search engines alone, it could potentially trigger a dampening factor, where other aspects were absent.

YMYL Queries and Content

Now, the query referenced above is relatively harmless in that no one is going to die or harm themselves through reading such a page, but what if the content was around health or finance, so called YMYL (your money or your life) content, what then?

At the very least, you’d expect the page to be authored by an expert, in an authoritative, trustworthy and accurate way.

So, let’s look at a health query and suppose that we wanted to inform a reader about how much insulin we should give to a 76 year old man with diabetes, weighing 12 stone.

Let’s ask a GPT AI this very serious question and see what it comes up with.

Now, I don’t know if that content is accurate, I’m not an expert in diabetes.

I don’t know if it’s correct or misleading either.

It reads like it knows what it’s talking about and mentions a few if’s and but’s and words that infer assurance and on the face of things it could be argued that it too is relatively harmless in that it isn’t saying that this person SHOULD do this or MUST do that, as it is pretty generalised, but if we consider how a GPT model works, then we might think twice.

How does a GPT Model Work?

GPT models work by using trillions of data points around words and generated text based on probabilities and training data, and can produce coherent sounding strings in ways that sound confident and knowledgable, but it’s also incredibly limited too.

There is no fact checking, there are no provided references, there is no person with the knowledge and expertise, it’s just a well designed word spouter that makes things sound plausible.

Just imagine if someone created a whole domain full of diabetes content; a great URL, lots of GPT generated articles that discussed the topic, food and diet, blood sugar, testing kits, exercise advice, drug advice, etc and then imagine that this content was viewed as authoritative, lots of links perhaps, social media accounts, facebook shares, advertising initiatives, mentions in mainstream press and then suppose, that there was one, just one errant piece that gave advice on how to deal with hypoglycaemia, and then suppose that a person died as a result of trusting that work, simply because a search engine decided that the corpus of content was unique and informative and trustworthy.

Scary right?

So, the search engines decided a while back that it was incumbent on them to ensure that for some query spaces it was particularly important that the content they returned to their users was trusted and accurate, curated by experts and not likely to harm or kill their users, or make them incredibly poor.

Quality Rating and E-E-A-T

Years ago in around 2005, a Dutch journalist at the time Henk Van Ess , discovered a secret document detailing a set of rater guidelines used by Google to rate documents around certain queries.

People were employed by Google via a 3rd party and presented with a query, a few results, and were asked to score pages around certain criteria and asked to mark them as offensive if they believed them to be spam.

Here’s an excerpt that gives a flavour of what they were saying.

These systems remain in place today and are used by engines to refine and train algorithms designed to surface what they consider to be the best content for their users.

Of course, a lot as happened since 2005 and the documents and guidelines have been refined over the years to change with the times.

A recent iteration late 2022 saw E-A-T get an upgrade to E-E-A-T the extra E being experience, and whilst Google say that the actions of raters don’t directly influence rankings, the reality is that they do en masse, especially if a site happens to be devoid of core expected aspects.

Of course, it’s super difficult to algorithmically determine if a piece of content has been authored by an expert, but not impossible either. One can but surmise around how the finer details of this works in practice ( Marie goes into some detail here for those wishing to explore it further) but you can be sure that if your content piece becomes rank-worthy and begins to climb the rankings, then an algorithm or a person will sooner or later determine whether your site is worthy of that position, from the perspective of having sufficient or insufficient E-E-A-T.

Not all query spaces require good E-E-A-T but it doesn’t hurt to think of it more generally, whatever you do.

Let’s say that you are a Dog Trainer and you want to rank for Dog Training queries, then it just makes sense that your content lends a little more credibility to your capabilities. It might mean that you become accredited or affiliated to a particular organisation and expressly say so on your site, it might mean that you list a bunch of testimonials from people or customers you’ve worked with, anything that adds a little extra authority to who you are and what it is you do.

SEO Is A Complex Thing

For years I’ve said that SEO is really simple once you get the fundamentals in your head. I’m no Einstein, so if I can do it, so can most.
In no particular order of importance the more simple aspects are:

• research keywords important to your niche using tools that reflect them
• grab a good brand-able domain name
• get rock solid hosting
• research your competition
• interrogate the SERPs for opportunities
• create relevant informative content
• maximise sharing opportunities
• maximise engagement via good UI’s
• use solid internal linking structures that help your content to rank, consider a silo’ed volume approach
• use templates that make sense and enhance relevancy
• share content in spaces where interested parties might congregate online
• have social accounts to share your content on
• advertise your domain and content when and where you can via paid means
• use expert link builders to help expand your web footprint
  

However, as simple as it is, it is always changing and within that simple list there are lots of other nuances that you need to understand and keep up with, IF you expect to rank.

Google and Bing and most other search engines aren’t really about making SEO easy is my well trodden mantra, as ultimately they are paid advertising platforms, that exist to return a profit to shareholders which they do via their paid search advertising products.

If SEO was this instant easy win, then more businesses would invest more of their marketing budgets to SEO than they would paid search, so for search engines like Google, it’s a constant balancing act of generating profits (and more ad spend) versus delivering non paid results that users both trust, and want to click on.

A list of the more complex nuances of organic search performance include:

• The use of Schema markup and its ability to influence clicks from the SERP’s
• The use of images to generate clicks via in SERP positioning and carousel
• Defensive SEO – monitoring your brand or niche
• Negative SEO – what it is, why it’s important to know
• Server knowledge – perfecting set up
• Programming knowledge – Python, PHP, Javascript, HTML
• UX and UI creating content people like and want to share and use again
• Href Lang and internationalisation
• Canonicalisation
• Robots.txt
• Understanding GSC and how it can benefit your domain
• The benefits of analytics packages and how they can inform change
• The myriad of tools that exist, SEMRush, Ahrefs, Screaming frog
• Crawl budget – what it is, what it is not
• CWV core web vitals, why they matter
• Algo updates – when, where, why, how etc
• History – the birth and evolution of search, the forebears, who they were and why they died
• Link acquisition – methods, pro’s, con’s and pitfalls
• How html elements might be used to generate clicks
• Video content and SERP carousels
• Youtube SEO
• Elevating brand through Off Platform Optimisation (OPO) reddit, Quora etc
• Formatting content to be returned in news results
• Appearing in Local search results
• Appearing in Google maps
• Good structure versus bad
• Knowledge graph how to get in it
• Featured snippets how to get them and why they’re important
• E-E-A-T knowing what it is and why it’s important
• Paid search and how it can benefit organic
• Understanding search features and realising that many are a test bed for eventual use in paid, often to the detriment of organic

It’s not an exhaustive list and I’m sure I’ve missed a bunch, because the space is SO broad, but the list of complexities is only likely to grow, especially with the inevitable growth and refinement of AI tools and the ever shrinking organic real estate in SERP’s.

Finally…

If you got to here, then well done. Give me a mention on your socials. Call me names or be nice, I prefer the latter.