If you’re not a digital marketing expert, you might think that TF*IDF sounds complicated. Despite its mathematical foundation, anyone can use TF*IDF to improve written content on their websites - beginners and experts alike.
TF*IDF helps you to create relevant website texts for your users, and at the same time shows Google that your webpage is relevant enough to be ranking high up in the search results.
In practice, TF*IDF is useful when creating and improving website texts for two main reasons:
- It shows you the most relevant keywords for your topic, so you know which terms to add or remove from your content to improve your rankings.
- It gives you keyword, structure, and topic inspiration.
You can use TF*IDF for all of your website texts – from blog articles, to product pages, career pages, company information, and of course product descriptions. In this article, we’ll focus on how you can use this formula to help improve or create the perfect product description.
What is TF*IDF?
TF*IDF stands for term frequency, inverse document frequency. Term frequency is the number of times a term appears in a document. This is divided by the total number of terms in the document. The formula looks like this:
Luckily you don’t need to understand exactly how the formula works. The logarithms ensure that common words such as “and” or “the,” that would in theory end up with a high term frequency (TF), are not given a high score.
Inverse document frequency is the number of documents containing the term, divided by the total number of documents (corpus). This is the formula:
Again, the logarithm ensures that only relevant words will end up with a high TF*IDF value.
How is TF*IDF relevant for rankings?
So how will this complicated formula help you improve your website texts? When you publish texts on your website, you should aim to rank for a certain keyword if you want users to be able to find your page via organic search, rather than by paid advertising. Your website texts should therefore be optimized for this keyword. This keyword is often referred to as a “focus keyword” or a “target keyword”.
A TF*IDF analysis shows you which terms have the highest TF*IDF value when searching for the focus keyword, showing you what the most relevant terms for this keyword are. These are the terms you should be including in your content if you want to rank well. If you use Ryte’s Content Success for your TF*IDF analysis, this data comes from the first page of the Google search results.
For Google, relevancy of texts is even more important than ever, as Google’s aim is to provide the best possible content to their users. In 2011, they introduced the Panda update to weed out weak content. Just two years later in 2013, the Hummingbird update was introduced. This placed even more value on the semantics of texts so that search results would be able to answer users’ questions more accurately. Relevant content is therefore more important now than it was before, and TF*IDF helps you find out what are considered the most relevant terms by Google.
A TF*IDF analysis is also a great way to get keyword inspiration and can help you structure an article or product pages. For example, you can use terms with a high TF*IDF score as headings for a webpage, and relevant text under the headings will come naturally.
TF*IDF is a great way to get inspiration for working out what to include in your texts so that Google considers you relevant, increasing your chances of ranking highly.
How to make the perfect product description
When you write product descriptions, the user has priority - you shouldn’t just be thinking about the ranking potential of your pages. Writing great product descriptions is important because if they’re persuasive and interesting, your customers are more likely to convert. You’ve probably heard from many digital marketers that there’s no such thing as optimizing for search engines; you only need to optimize for the user. If your website is easy to use with texts that are enjoyable to read, KPIs such as time on page and bounce rate (i.e. the classic SEO metrics that are usually cited as ranking factors) will improve. This will send Google signals that your website is great, leading to better rankings.
You can find some useful tips regarding what to consider when writing product descriptions in this article from Kissmetrics. These are some of the things you need to think about:
- Make sure you know who your target audience is and create descriptions that are relevant to them.
- Make sure you describe the features and benefits of the product.
- The tone of the description should reflect your brand personality.
- Users should easily be able to scan over it.
- Very important: your product description should be persuasive!
How to use TF*IDF analysis for your product descriptions
Now that we’ve established how TF*IDF works, its benefits, and the importance of writing good product descriptions, it’s time to put the two together.
There are many tools you can use to carry out a TF*IDF analysis. Ryte Content Success is one of them. With this tool, you can carry out a TF*IDF analysis based on a specific language and region, you can directly compare your content with your competitors’, and you get easy writing assistance with the Content Optimizer. Let’s look at an example to see how you can implement TF*IDF practically in your product descriptions.
Example of using TD*IDF analysis
You have a shop selling furniture, and you want your couch to rank in the top spot in Google or at least pretty high up. Obviously, there are some fierce competitors that will be hard to beat (I can think of a Swedish one for example). You also have to bear in mind that search results are often full of adverts, meaning there might not be more than 5 organic search results on the first page.
Run a TF*IDF analysis for “couch”. With Ryte, you can set the target language and region of your TF*IDF analysis, particularly useful if you want to appeal to a local target group. These are some of the top terms I found when carrying out a TF*IDF analysis for the focus keyword “couch”, based on the search results for the United States.
Figure 1: TF*IDF analysis for “couch” from Ryte
Of course, you can see on the results that important terms for this keyword included “sofa,” “sofas” and “couch”, but you don’t need to be told to include too many of these fairly basic keywords. Instead, look at the unique or unusual keywords. For example, this product description from Ashley furniture contains “leather,” “fabric,” and “cushions”, and “upholstery”.
Figure 2: Product description from Ashley (source)
Be careful not to use a certain term too much - this might be seen by Google as spam. If you’re concerned about using a certain term too often, use the Content Optimizer from Ryte to find out. This will provide you with suggestions regarding which terms you should add, and which ones you should remove so that your text can be as optimized as possible for your target keyword.
Figure 3: Content Optimizer from Ryte’s Content Success
TF*IDF has the power to help you create and improve content in many ways. Giving you keyword inspiration for your product descriptions is just the beginning: TF*IDF can help form the basis of your entire content strategy. If you have a blog, the unique terms suggested to you by a TF*IDF analysis can provide inspiration for article ideas, for example how-to guides, or useful information on related topics.
Conclusion: TF*IDF analysis for online shops
Using TF*IDF is a fantastic but simple way to help improve your website texts. It helps you cater toward both search engines and the user. Firstly, you provide your user with relevant and engaging texts, meeting their expectations and fulfilling their needs, which in turn provides positive signals to search engines. Secondly, you show Google’s algorithm that your page is relevant for the focus keyword, increasing your chances of ranking highly. If you don’t use TF*IDF yet, now is the time to start!