Why Keywords Stuffing Doesn’t Work Anymore and 2 Lessons That I Have Learnt

Fat man stuffing food

When I first started out as a SEO practitioner, keyword stuffing was something I would actively work on to ensure my projects rank. For example, if I am targeting “SEO Singapore” as a keyword, I will ensure that the exact same phrase “SEO Singapore” will appear in the meta title, meta description, headings, content, images’ alt text of the page.

It used to work 2-3 years ago.

During that time, I could observe the significant increase in my SEO projects’ rankings. But the improvements did not last, and the rankings began to drop.

Then I start to realise and understand when I started to ask “why”…

First, we need to know about LSI keywords, which means Latent Schematic Indexing. LSI is a method that Google uses to identify the context, the topic about the content.

You can read more about LSI here at Backlinko’s blog.

But one of the key lessons that I’ve learnt is not from Backlinko’s blog but it is from a Google’s white paper published in 2016 titled “Improving topic clustering on search queries with word co-occurrence and bipartite graph co-clustering”.

You do not have to read the whole paper, just read the title – 2 keywords: Co-occurrence and Bipartite Graph Co-Clustering.

What is Co-occurrence?

Co-Occurrenceis the technique of clustering words to canonicalize variations, including misspelled and plural forms as well as to remove stop words, and generates topics by looking for words that co-occur frequently with topic anchors in the given set of queries.

What is Bipartite Graph?

A bipartite graph, also called a bigraph, is a set of graph vertices decomposed into two disjoint sets such that no two graph vertices within the same set are adjacent


Source from http://mathworld.wolfram.com/BipartiteGraph.html

To put it simply, it is 2 set of datas and how are they being put together to propose relationship. Note: it is not how they are related to one another.
In the example below, we can see the author had used queries and URLs as separate nodes. They then used impressions and clicks data to measure “edge weights” and induce query similarities.
Edge weight

With both techniques, we can infer that

    1. We do not have to be bothered about the exact phrasing of the keywords. We just need to ensure different part of the phrase are appearing near to one another.
    2. Impressions and clicks data are important in Google to determine the topic cluster of your content. The older the content is, the more likely it will be able to rank higher.
    3. Maybe we can make the ranking rises faster by leveraging on AdWords, so Google will be able to collect more impressions and clicks data.

Finally the second lesson that I learnt from this experience is be focused on delivering value, information that help the reader. Think deeper about what will readers like to learn more about the topic you are writing.

Don’t stuff keywords. Stuff value.

If you want to learn more about SEO, you can find out more about my upcoming SEO Course.
Find out more about my SEO services, please click here.

Ken Toh

Principal SEO Consultant at Thinking Notes
Ken graduated with a Master’s Degree in Information Systems from Nanyang Technological University, specializing in Human Computer Interaction and Usability Engineering. Beside WordPress development, Ken provides professional SEO consulting services to SMEs and has helped his clients improved their businesses significantly through top page ranking results.