It's no secret that Google tweaks its algorithms on a constant basis, as well as introducing new algorithms, some of which have played huge havoc to the search results. Google's latest algorithm is called the Paraphrasing Algorithm. The type of technology behind it has already been seen in what are known as either featured snippets or the Google Answer Box.
Here's an example below.
Google's new algorithm can actually create "coherent" articles by combining your content with your competitor's content. Imagine the content that you invest money to create can now be mingled with your competitor's content.
So how does the paraphrasing algorithm work?
Content is summarised by the new algorithm, which then extracts out parts of the content that the algorithm considers irrelevant. It aims to answer questions asked by users, much in the same way as in the Google Answer Box. You type in a question and Google shows an answer, that may seem satisfactory and stop you from reading a webpage that contains a more complete explanation.
The paraphrasing algorithm then uses an extractive summary algorithm and an abstractive summary algorithm, which aim to reduce the original content to what the algorithm considers to be the most important parts (extractive), and then uses a type of paraphrasing (abstractive). However the end result is not always perfect. In fact, according to an article in Search Engine Journal, this new algorithm produces results that 33% of the extracted and paraphrased content contains fake facts.
The example below is from a recently developed client's website. The client is Professor Shane Higgins, who is a top Irish consultant obstetrician gynaecologist, who has been elected to be the next Master of Ireland's National Maternity Hospital. When creating the content for his website, our team decided to write and create the pages in anticipation of patients' questions. So in this instance, if a user searches for "What are birth defects?" they will see the results below.
However Google's algorithm, in this instance, has not delivered the most accurate answer. Instead of explaining what are birth defects, the algorithm has extracted other text from the same page. The extracted text tells users about some of the possible causes of birth defects.
If a user decides to visit the website, the user will see that the actual question: "What are birth defects?" is on the page, and not only this but it has an internal page link directly to the answer. So essentially the work was done both for Google and the user, however the algorithm didn't recognise it correctly.
What Google says about the algorithm?
In their research paper, Generating Wikipedia by Summarising Long Sequences, Google stated the following:
“We show that generating English Wikipedia articles can be approached as a multidocument summarization of source documents.”
The essence of the theory is interesting on one hand, but as an author, I also question it as a form of plagiarism. Would I want my research and style of writing mixed up with an author that I haven't collaborated with on purpose?
What does this mean for your company's SEO results
Certainly it emphasises that Google is going to "reward" authority content on websites and blogs, as it has for many years now and probably even more than ever before. Going back to our recent website for Professor Shane Higgins, for such a new website (two months live at the time of writing this post) Google trusted it as an authority very quickly indeed.
The majority of the pages within the website are long form content, of between 2000 to 3000 words. They are designed to answer questions easily. Of course each page contains images labelled with alt tags and titles. We didn't include as many images as we would normally do as the subject matter is very delicate, and therefore we intended to keep the website low key visually. The patients who attend Prof. Higgins have pregnancy complications.
Key takeaway: Original, highly useful long form content ranks even better when it answers questions.
Thin content, meaning content light in substance and genuinely useful information, is like to get pushed further down the rankings. The further Google focuses on this new algorithm and the whole concept of providing answers to its users, the more likely Google is to disregard as unimportant content that is thin. Often this type of content has been developed by companies who have focused too much on selling their services or products, and far too little on providing rich, informative content that is customer-centric.
Key takeaway: Focusing on customers' needs and predicting their questions is not only good for your business, it is more likely to be recognised by Google.
Being a bit nerdy in terms of research is the way to go. This comes naturally to me and lots of other writers that I know that truly love to research, analyse and then write their own original content that has something unique to say about an already treated subject. Going further with your research and finding something fresh to say on old topics is something that Google has given credibility to for many years. With Google wanting to present answers to questions in this way, this is set to increase even further.
Key takeaway: Do intensive research, analyse it and say something fresh, while always bearing in mind the questions that relate to whatever topic your article discusses.
Which companies will be positively affected by Google's new algorithm?
If your company's website and blog have recognised the need to focus on customers and their potential questions and needs, then you should be positively affected by the new algorithm.
If not, it is time to re-develop existing content and create a strong, intelligent content strategy that you need to start implementing sooner rather than later.