Why “Buckets” are critical to SEO?

People are very often mystified and confused with the term SEO (Search Engine Optimization). However, SEO is a measurable and clearly quantifiable process where all you are doing is giving signals to Google notifying your value to them. This isn’t too different from the signals we wish to put forth in our resume and during an interview with a prospective employer. We want to stand out from the crowd and want to be hired for our unique traits and/or our superior qualifications. 

In a similar way, the job of any SEO process is to get you in the top 10 rankings of a Search Engine Result Page (SERP) for a particular keyword or search term. In our analogy, think of the SERP as the open job positions (with 10 open positions in this case). The Keywords can be thought of as the particular job you are trying to apply to. Therefore, your SEO process is not too different from trying to be the best fit for an open job position over others doing the same. You have to stand out and make yourself unique!

The main difference is that instead of scrutinizing yourself to a subjective evaluation, where you might win over the employer with your wit and charm; Google uses complex algorithms to determine your place in its rankings table. Hence, your job as an SEO is to be better than your competitors in meeting Google’s 200+ ranking factors. The better your overall score with the rankings, the higher your website will rank. There is no more myth to this. 

Now, these mathematical algorithms can be visualized as being a collection of empty “buckets”. No, they are not buckets filled with water or being used to build sandcastles! These buckets are representations of various factors used to rank your site. The bucket method was already patented in the earliest days of the Internet. 

The patent states that the bucket-oriented route planning method is:  “A method of planning optimum routes on the basis of successively selected subsets of the total topographical and traffic information, so-called buckets, which method anticipates which buckets will be of importance in the near future for the calculation of the navigation data, and navigation system comprising a route planner for carrying out such a method.”

Simply put, Google algorithms divide your website ranking factors into various “buckets”. It is needless to say that some buckets are more important than others. For example, one bucket could be for the quality of your site, one for trust, and one for authority. These are the most important ranking factors used by Google. Hence, if you want to rank better you ideally want as many buckets as possible to be “filled”, and especially the 3 mentioned above. 

Ranking better than your competition on Google can thus be summarized as a process of either: 

  1. Making your website quality better (writing unique, engaging content being the top priority).
  2. Improving your trust factor by linking to higher-authority websites.
  3. Making yourself the authority in your niche. This is done by ideally creating content (video, blog, etc.) that is share-worthy and effectively utilizing your fan-base to promote your products/services. 

When you have implemented these above strategies, Google will accordingly assign you a PageRank. Your goal obviously is to get into the top 10 ranks on the Search Engine Result Page (SERP).

There are various other factors or “buckets” as well, but they generally fall into these 3 categories stated above. 

How Search queries use buckets? 

Now that you know what to do to get your website ranking better, it is now time to know why this is the case. There are several patents that can explain to you how sites are ranked. The whole world of search engines was created mainly to address a user’s needs for informational search queries. After being founded in 1998, Google took this concept to a new level by being the first search engine to provide value to the user’s search queries. This was unlike its rivals (Yahoo for example) that capitalized only on ad revenue. One of the ingenious methods deployed by Google was its cluster architecture(1)  that utilized the theory of algorithmic “buckets” to divide each search query. Basically, every query (for example ‘shoes’) was divided up among different servers that dealt with a particular topic (in this case footwear or apparel). This enabled Google to come up with only relevant search results and in the fastest way possible. The early implementation of cluster architecture made Google the Search Engine leader by the mid-2000s. 

After crushing the competition, Google went on to define the SEO process as serious science. It was no longer just about answering search queries, but about understanding user intent. This has been a general trend in the advancement of search engines as can be seen by this patent. Google has been perfecting this with its improvements in NLP (Natural Language Processing) and Artificial Intelligence. They came up with many algorithm updates in the 2010s addressing the issue of user intent and getting closer to the ideal each time. The latest among these updates is referred to as the BERT update. BERT, or Bidirectional Encoder Representations from Transformers, to be more precise, is a further improvement on simple “buckets”. It is about understanding search terms in relation to each other and taken as an entire phrase, rather than just a bucket of words. Hence, the bucket of words has been broadened to include entire phrases (based on human search intent), thus reducing the time it takes to search particular terms. In simple words, by studying human typed search query patterns, Google has gotten one step closer to deciphering what is in our minds. This has great implications in improving search accuracy and giving faster and more reliable results for long-tailed keywords. 

The Evolution of Google’s Search buckets

2019 was a critical moment for Google updates and many fundamental changes were done to Google’s search algorithms. Google initially had an algorithm where it would classify a whole website within a “bucket”. Google viewed all the webpages and content of the website in terms of a “flower” structure (shown below). Google would club all 200 search ranking factors into 1 bucket (with each ranking factor seen as a petal of the flower). Each of these ranking factors (content quality, trust, authority, etc.) carried individual weight and their combined weight is what eventually factored into the rankings. Therefore, simply explained, the “heaviest bucket” would be ranked #1, the second heaviest as #2 and so on.

Before Flower After Flower

(sample client before)                                   (after Search3w applied algorithm changes)

Afterwards, with more improvements (latest being the BERT update), additional ranking factors such as “related terms” came into play. However, the biggest change is that instead of clubbing everything into one “bucket” Google algorithms would now categorize a website in separate “buckets”. So, a website would now be divided(2)  into separate “sub-buckets” like “citations”, “links”, “content” etc. The factor that counted now was not the “size of the flower” or the “heaviness of the overall bucket”; but its arrangement and structure. Hence, in order to rank for a particular keyword, or search term, your relevancy to that term, the relevancy of your content etc. would be more important than all your website factors taken overall. Google’s latest algorithms are keener on specificity rather than a one-size-fits-all competition, where only a “heavier” website or website with more content and the strongest link building wins out. Google algorithms can now easily track the most relevant content for your user search query based on ranking for the “appropriate bucket”.

(1) Do no confuse between server cluster architecture and ranking algorithmic clusters
(2) For example, when Google spiderbot finds the business address, it place it in the website bucket of NAP. When is find image it placed in the IMAGES bucket connected to the universal search branch

Some patents to further look at 

US Patent 5170353 US 5170353

US Patent 8433702 US 8433702

US Patent 20090021572A1 US 20090021572A