Google’s algorithm has drastically changed over the years. Or should we say Google has started to think more like a human now? Google’s latest core update in June saw a significant drop in traffic overnight, notably soft news and chatters were hit by the medic update.
So what hasn’t changed?
TF-IDF doesn’t sound cool enough to be included in your SEO tactics for 2019. But the result it provides is too unreal not to share. Even after every algorithm update, this one tactic seems to be less affected when it comes to identifying quality content.
Understanding how Google identifies valuable content is vital to stay ahead amongst the competition as Google’s SERPs continue to evolve.
This is why I’m about to take you through the most under-appreciated tactic for optimizing your content for SEO that will help to bring in conversions.
TF-IDF is Google’s way of identifying quality content based on the established expectation of in-depth pieces of content.
Sounds complicated? We’ll break it down:
What is TF?
TF stands for Term Frequency. Meaning, how often the term, or word, occurs in the content. The more often the term appears, the higher the weight.
Term frequency is calculated as: tf(t in d) = ?frequency
Term frequency is further divided, taking into consideration the length of the document.
What is IDF?
IDF stands for Inverse Document Frequency. Ok, this one sounds scary! But it’s not hard to understand. IDF shows you how important the term is.
IDF has two things to look out for:
- Scale down the frequently appearing words like ‘the’ or ‘is’ as they appear in most documents.
- Scale-up more unique terms like ‘future’ or ‘SEO’ to zoom in on more relevant documents.
Inverse document frequency is calculated as: idf(t) = 1 + log ( numDocs / (docFreq + 1))
TF*IDF is more like a keyword inspiration tool. It’s not a tactic driven by writing an article and multiplying three or five times a term to find the relevant keyword. TF-IDF is used to get keyword inspiration from terms that could be used for page enrichment.
For instance, If you do a TD*IDF analysis for smartphones, you could get terms like iPhone or Android. Interestingly, when you analyze the Asian market, you can see a high correlation for the word ‘waterproof ‘ since many phones sold in Asia are waterproof.
This is a prime example since a lot of people search for the term ‘waterproof’ along with ‘smartphones.’ So it’s essential to add this term to the page you want to optimize.
Why is TF-IDF a vital SEO tactic?
Google is getting smarter in refining it’s metrics to think more like a human. From songs you can’t get out of your head to the shopping list you want to spend your money on, Google pretty much knows everything.
Maybe this mathematical equation sounds like it’s from the ’80s, but when it comes to uncovering the semantic relevance of keywords, TF-IDF does a fantastic job in revealing the type of content Google values the most.
While standard keyword research shows what people are looking for, It won’t reveal the related keyword your competitors are using. This means even if your content is well articulated, the chances of going down the line against the top SERP results are high.
The cool thing about TF-IDF is that you can quickly rank your keywords to the top, and better keyword means more traffic and conversions. With the chance of a better ROI, adding this minor tweak to your strategy could add more dollars in revenue.
TF-IDF analysis has seen a drastic increase in snippet featuring. This is because we already know what words and phrases Google finds important, and allows our content to be a qualifier.
When to use TF-IDF Analysis
TF-IDF is used by content creators to identify their content gaps by comparing it with the top 10 search results. The analysis can also be used when developing new content so that it could potentially rank higher without any link building strategies. However, conducting TF-IDF analysis for every content sounds impossible. So which do you focus on?
- Content struggling to break the first page
Audit your blog content and discover which posts are stuck on page 2 or 3 on Google. If the content is already optimized for technical SEO and has good authority, your content will likely benefit from TF-IDF Analysis.
- Content dropping traffic or ranking
If there’s a drop in traffic or ranking for a well-performing blog, then chances are there’s high competition for the topic you are ranking for. Or of course, the ever-changing Google algorithms have had an impact. A quick way to find this is to compare your current SERP results with last years results, using a tool like Spyfu. This will give you an idea about new competition and compare the relevance of the content.
- Service pages struggling to rank
If your service or product page is not ranking well for the money terms, chances are, critical terms are missing in the page. TF-IDF can give a brief analysis of the terms you should be focusing on.
How to Conduct a TF-IDF analysis
Google understands the metrics the drive more engagement or how well the users are pleased with the content. Good news? TF-IDF can give you a decent insight into Google’s metrics. You can see what your competitors are up to, along with the quality and relevancy of the content they are providing to their audience.
How about an example?
Let’s consider a document containing 2000 words, where the word ‘SEO’ appears 40 times.
The term frequency of the word ‘SEO’ will be ( 40/2000) = 0.02.
Now consider you have 10000 documents and the word ‘SEO’ appears in 10 of these. Then the Inverse document frequency will be calculated as log(10000/10) = 3
Thus, the weight of TF-IDF will be the product of these quantities will be 0.02*3 = 0.06
This weighting score tells us how relevant our keywords are, which is neat and handy when you apply it to SEO.
Yes! It’s crazy trying to get the maths done to find TF-IDF for your pages. Fortunately, there are few free tools available to simplify the process, and Seobility’s TF-IDF tool is so far, my favourite tool. There is a text editing tool available where the content can be changed according to the tool’s suggestion.
With most of the TF-IDF tools, you can find top ranking search results for your keyword and see which phrases in the term is used often for the page to perform well.
Let’s take our previous example of how TF-IDF analysis found the hot topics ‘Smart-phone and water-proof’. Once you find the magic keyword, compare it with the top URLs you are competing against.
Using tools like Ryte, we can get a detailed breakdown of any high ranking keywords which competitors are frequently using in their content.
For instance, If I want to claim a top spot and get my smartphone business going in Asia, I’ll have to think about incorporating the term ‘waterproof’ and ‘smartphone,’ as the high search volume indicates people are after waterproof smartphones.
I can also rank for low hanging keywords, which are having high search volume since those keywords won’t be battling out with the competitors. Either way, this tactic will help improve the topic relevance and search results.
TF-IDF is more than just a curious acronym. It is an essential part of developing a great content strategy. It’s a way to learn how machines value your content and reverse-engineer the results to improve your current strategy.
Do you have any neat TF-IDF strategies? Share them in the comments below!