Ever since the dawn of search engine optimization (SEO) as an industry, experts have stressed about the importance of having the right keywords embedded in your blog content in order to rank high on Google.
Many say that keywords are "do or die" when it comes to ranking. There are sophisticated checklists that tell you how to optimize your keywords and even dedicated tools that specify exactly how many times you need to spell out a particular keyword in order to rank.
Does that mean spending an extra hour to mention the keyword "plumbing service" 17 times on my article will get me #1 on Google?
We're here to tell you that this form of keyword optimization is a waste of time.
Keyword optimization is different from keyword/intent research, which is still an important part of SEO. We're saying that creating a list of keywords is fine, but optimizing their frequency on your content is not the way to go.
Before you object and recall how a particular keyword optimization you did two years ago got you first place on Google, hear us out!
In this article, we're going to demystify SEO keyword optimization and why you shouldn't focus on it. We'll first dive into how search engines like Google used to work, which lets us understand why SEO experts are obsessed with keywords. Then, we'll explain what's different today, and what you can do tomorrow to keep your SEO strategies future-proof.
Great content is about answering user questions. That's because a search engine's primary goal is to bring relevant and quality websites to answer your search intent. This is reflected in Google's mission to organize the world's information.
So why are SEO experts talking about keywords - a seemingly unrelated concept - for ranking success?
To understand the reasons, we need to take a brief historical look at how search engines came to be.
Computers were relatively slow and dumb back in the early 2000s. Imagine if you were Google. Your job is to sort through millions of websites and to give the user the most relevant ones.
Figuring out if a website or a piece of content is relevant to the user's query is a big challenge.
In order to figure out the relevance of a website, programmers had to resort to heuristics - a set of simple rules - to see if a piece of content's topic matched the user's query.
One simple heuristic is to check if your content's topic answered the user's query was by counting keyword frequencies.
For example, if someone types "beaches in CA" and your article contains that phrase, it is probably relevant to that query. If you were searching for the topic "flight discounts" and an article mentions "flight" 100 times, it must be about flights!
So we've figured it out, right? Not quite.
This "naive rule" has an obvious drawback. In most languages, including English, there are common words and phrases that are overused. According to Zipf's law the most frequently occurring word, "the", accounts for almost 7% of all words used in English. The second and third place for most frequently used words are "of" and "and".
These common words dilute an article's topic. It's obvious that "the" is not the primary topic of your content. So search engine programmers resorted to more sophisticated ways to re-weight these keywords.
The classic way to combat the "keyword dilution problem" is called term frequency-inverse document frequency (TF-IDF). Without getting into the math, this fixes the naive keyword counting strategy by re-weighting each keyword by how important it is relative to other websites.
So a keyword becomes more important for your content if it is mentioned a lot in your content, but not on other websites. This helps Google determine the unique topics you write about.
This means that terms like "the", which is used very often, but not unique between websites, should not be considered as "keywords".
Calculating the TF-IDF of your content is straightforward, and there are free tools out there to help you.
But how does this connect to SEO keyword optimization?
Now that we have an idea of how search engines back in the early 2000s work, we can understand why keywords were so important.
If we assume that Google uses an algorithm based on TF-IDF, then so long as we have the right keywords, we can be more relevant than other websites competing for a search query.
And being more relevant means higher ranking, higher traffic, and all that goodness. This is why SEO experts are so fond of keywords as they have, for more than a decade, been an important ranking factor.
So can I just put in "plumbing service" 1000 times and be done with SEO?
Like most things, if it sounds too good to be true, it probably is.
Even in the 2000s, Google implemented ways to circumvent these types of black hat SEO techniques that relied on keyword stuffing (repeating the same keyword 100 times). PageRank is a great example of an algorithm that mitigates some of these exploitative techniques by estimating the trustworthiness of your content based on backlinks. These additional ranking factors made it harder to game the keyword system.
In the 2010s, we had the Google Panda Update, which helped further deter low-quality sites from ranking high. Keyword stuffing and over-optimization can lead to your website being penalized.
From then, SEOs quickly evolved their strategy to avoid over-optimization, and instead focus on just-right optimization. And that's where certain SEO tools came in handy to recommend a specific number of keywords your content must use to avoid under or over-optimization - a form of smarter keyword stuffing.
So instead of using "plumbing service" 1000 times, you would use it exactly 17 times in your content.
But in the last few years, Google has evolved again, and the focus this time is away from keywords entirely.
If we look at the update trends at Google, semantics (the meaning of your content) is becoming increasingly important as a ranking factor, while keywords are becoming less relevant. This is due to the rise of AI and deep learning technology fueled by large amounts of computing power.
The game-changing technology here is that programmers no longer need to rely on simple heuristics like TF-IDF anymore. The problem with keyword-based approaches for search is that they don't take account of context.
For example, if I use the word "rankings" 10 times in my article, does it mean my article is talking about "search rankings", or is it talking about "basketball player rankings"? Words alone cannot provide context, and thus cannot derive meaning.
And so Google, with its large computing power, developed BERT, an AI system that learns meaning from text without needing programmers to define heuristics. BERT works by predicting text from the web and is capable of extracting the meaning of sentences from your content.
BERT has been integrated into Google search since late 2019, and now almost all English search queries use BERT instead of relying on keywords-based approaches. We closely followed Google's research and found that they published more improvements on BERT for search in 2020, indicating that in the near future semantics will be increasingly important.
BERT marks a paradigm shift for how search is done, and also how as SEOs we should optimize for search engines in the future.
But wait! Some SEOs would argue that their keyword-based strategies still worked last year. Why is this the case?
It's very possible that writing your content with optimized keywords act as a guide in aligning your content with the right topic. So it's not necessarily the precise keyword counts that led to good rankings, but rather, keywords implicitly force you to cover your topic in-depth, which correlates with higher rankings.
Some SEOs focus on more sophisticated keyword techniques such as latent semantic index (LSI). LSI may be one step up above keyword optimization, but it is still a primitive approach that doesn't really help ranking.
The truth is that no one besides Google really knows how the search algorithm works. So it's possible that certain queries may still rely on keywords.
What we can say is that from our experience delivering hundreds of articles to our clients, the articles that optimize for keywords are not significantly different from those that don't in terms of rankings. Other studies have also shown that there is no correlation between keywords in title tags and higher rankings.
It's important to watch out for red herrings, and A/B test your SEO strategies to make sure correlation is not mistaken for causation.
SEO in the decade of the 2020s will be fundamentally different from the last two because of the rise in AI technology.
AI has enabled Google to adjust its search algorithms to become more aligned with its goal - to deliver great content that answers the user's query. One can say that algorithms are becoming more human.
With all that's said, it begs the question - what should we optimize for today to rank high on Google?
We believe the best strategy is to stop playing the cat and mouse game against Google, but rather to co-operate with Google as a partner. So long as your goal is to provide answers to the user's search queries, future updates can only benefit your SEO.
Here are three simple steps to get you started:
It is essential to understand what your customers want and need in order to develop a winning strategy for your business. Your website will be more successful if you know exactly what customers are looking for and what queries they might type into Google.
There's no one-size-fits-all method to understanding the customer, but a good technique is to talk to them or interview them. Go on to online forums, Facebook groups, Reddit and see what your target demographic is talking about. These reveal a lot of insight.
From there, ask yourself these questions:
These are the starting points for your content ideas. Let questions drive your content.
Now that you have a list of questions, it's time to find the answers.
Most SEOs target information queries. These are searches that are going to be looking for general information such as "how to" or "walkthroughs."
In order to write great content for information queries, you have to be well versed in your topic and synthesize interesting ideas from various sources. Start by doing research and gain expertise in your domain. Google wants content that follows E-A-T principles (which stands for Expertise, Authority, and Trust).
To build expertise and trust in your content you need to view sources from different perspectives and synthesize that information. To add icing to your content, sprinkle in your unique perspective. One quick way to get a glimpse of similar sources is to use Jenni's topic ideation tool.
This groups all your sources by their topic information and gives you a birds-eye view of what competitor sources are discussing. With the right ideas, you can create a content brief swiftly.
Finally, get your keyboard and start writing! When writing, always keep your target audience in mind and write with a purpose. Here are some brief tips:
With that, you should have what it takes to get started on creating quality content - without needing precise optimization of keywords.
Before we conclude, we want to clarify that we're not saying you shouldn't do any of these keyword optimizations, but instead, you should spend your precious marketing time and budget efficiently.
Practice the Pareto Principle - spend a few minutes ensuring your content is on topic and that's it. Studies show that Google considers over 200 factors for ranking, so you definitely should not put all your eggs in one basket.
Focus the majority of your time talking to your customers, understand their questions, and answer them in your content.
And finally, focus on delivering value to the reader above all.