Last updated on
In its early days, Google depended primarily on plain text data and backlinks to determine rankings, using monthly updates known as the Google Dance.
Today, Google Search has evolved into a complex system with numerous algorithms designed to deliver content that aligns with user needs.
SEO has become largely a numbers game, focusing on:
We also consider third-party metrics, such as search visibility or attempts to replicate PageRank. However, the core of SEO typically revolves around these quantifiable metrics.
These metrics are crucial because they are often how we, as SEO professionals, are evaluated and compared against competitors using various tools.
Clients aim to improve their rankings and increase organic traffic, which, in turn, should boost leads and sales.
When selecting target keywords, there’s a temptation to prioritize those with the highest search volumes. However, the intent behind a keyword is far more important than its search volume.
Additionally, keywords with low or no search volume are often dismissed as having no “SEO value,” but this perspective can be misleading. The actual value of these terms depends on business-specific factors and requires thoughtful analysis.
This consideration is a crucial, yet often overlooked, aspect of content creation. While ranking for a specific term is important, the content must also be relevant and meet user intent.
In 2006, a study by the University of Hong Kong identified two primary search goals at the foundational level:
Further categorization reveals two main types of search intent:
Lagun and Agichtein (2014) investigated the complexity of users’ online tasks, using eye-tracking and cursor movement data to assess user satisfaction and engagement with search results. Their study highlighted how user attention patterns vary with task complexity (the mental effort required) and search domain (e.g., health and finance searches often receive more scrutiny compared to sneaker shopping).
Search engines are continuously improving their ability to understand both types of search intent. For instance, Google’s Hummingbird and Yandex’s Korolyov and Vega represent advancements in this area.
Numerous studies have explored the intent behind user queries, which is evident in the types of results Google provides.
In 2016, Google’s Paul Haahr delivered an insightful presentation on how search results are generated from the perspective of a ranking engineer.
The “highly meets” scale discussed in the presentation mirrors the criteria outlined in the Google Search Quality Rating Guidelines.
Haahr’s presentation delves into fundamental theories, such as how users searching for a specific store (like Walmart) are generally more interested in finding their nearest location rather than the company’s headquarters in Arkansas.
This concept is reinforced in the Search Quality Rating Guidelines, particularly in Section 3, which describes the “Needs Met Rating Guidelines” and their application to content evaluation.
The rating scale used ranges from Fully Meets (FullyM) to Fails to Meet (FailsM) and includes flags for content issues such as pornography, foreign language, loading problems, or content that is upsetting or offensive.
Raters evaluate not only the websites displayed in search results but also special content result blocks (SCRBs), also known as Rich Snippets, and other search features beyond the traditional “10 blue links.”
A particularly noteworthy section in the guidelines is 13.2.2, titled “Examples of Queries that Cannot Have Fully Meets Results.”
This section explains that queries which are ambiguous or lack a clear user intent or dominant interpretation cannot receive a Fully Meets rating.
For example, the query [ADA] could refer to the American Diabetes Association, the American Dental Association, or a programming language from 1980. Because there is no single dominant interpretation, a fully satisfactory answer cannot be provided.
Recently, Google has been giving more prominence to Reddit in its search results.
A 2011 study explored the potential of community-based question-answering (CQA) platforms to enhance user satisfaction with web search outcomes.
The research involved data from an unnamed search engine and an unspecified CQA site, employing machine learning models to predict user satisfaction. Key data points included:
The study concluded that factors like the clarity and completeness of answers were significant indicators of user satisfaction.
Despite these findings, there remains a perception that Reddit may not be a valuable addition to search results and should not be prioritized.
Due to the variability in language, many queries can have multiple meanings. For instance, [apple] might refer to either a technology company or a fruit.
Google addresses this challenge by categorizing queries based on their most likely interpretation. This interpretation helps define the user’s intent.
Query interpretations are divided into three main categories:
Dominant Interpretations
The dominant interpretation represents the most common meaning users intend when they perform a specific search.
Google’s search raters are instructed that the dominant interpretation should be evident, particularly after additional online research.
Common Interpretations
Queries can often have several common interpretations. For example, [mercury] could refer to either the planet or the element.
In such cases, Google cannot offer a result that “Fully Meets” the user’s intent. Instead, it provides a range of results covering both interpretations and varying intents.
Minor Interpretations
Many queries also have less common interpretations, which can be influenced by regional differences.
Occasionally, minor interpretations may shift to become dominant ones if real-world events generate significant public interest in the new interpretation.
The “Do, Know, Go” framework categorizes search queries into three types: Do, Know, and Go. These categories help determine the kind of results Google delivers to users.
Do (Transactional Queries)
“Do” queries are those where users want to complete a specific action, such as making a purchase or booking a service. This type of query is particularly significant for e-commerce websites, where users might be searching for a particular brand or product.
Device action queries, which involve tasks performed on smartphones and other technologies, also fall under “Do” queries. Their importance has grown with the increasing role of mobile devices in our lives.
The launch of the first iPhone by Apple in 2007 revolutionized our interaction with handheld devices. Smartphones became more than just phones; they provided internet access on our terms.
Although earlier technologies like 1G, 2G, and WAP existed, it was the advent of 3G around 2003, along with the introduction of widgets and apps, that transformed our behavior and significantly enhanced internet accessibility for a broader audience.
Device Action Queries & Mobile Search
By May 2015, mobile search had overtaken desktop search globally across most sectors. As of 2024, nearly 60% of web traffic now comes from mobile and tablet devices.
Google has adapted to this shift by emphasizing the need for mobile-optimized sites and transitioning to mobile-first indexing, reflecting the growing importance of mobile-friendly content.
The rise in internet accessibility has also led to more frequent searches driven by real-time events.
Consequently, Google estimates that 15% of the queries it processes each day are entirely new and have never been encountered before. This increase is partly due to the broader global reach of smartphones and the expanding internet penetration.
Mobile devices are increasingly shaping not only how we search but also how we engage with the online world. Currently, 95.6% of global internet users aged 16-64 access the web through mobile devices.
One important aspect of mobile search is that users may not always complete their queries on the same device. From my experience across various sectors, many mobile search queries are primarily used for research and informational purposes. Users often switch to a desktop or tablet later to finalize a purchase.
According to Google’s Search Quality Rating Guidelines:
“Due to the challenges of using mobile phones, Special Content Result Blocks (SCRBs) can significantly help users complete their tasks quickly, especially for certain informational, local, and transactional queries.”
Mobile search is a major focus in Google’s Search Quality Guidelines, with Section 2 entirely dedicated to it.
Know (Informational Queries)
A “know” query is one where users seek information about a specific topic. These queries are closely associated with micro-moments.
In September 2015, Google introduced the concept of micro-moments, which have emerged with the rise in smartphone usage and internet access. Micro-moments occur when users need immediate answers to specific queries, such as checking train schedules or stock prices.
With the ability to access the internet anytime and anywhere, users expect brands and real-time information to be equally accessible.
Micro-moments are evolving, and know queries can range from straightforward questions like [how old is Tom Cruise] to more complex inquiries that don’t have easy answers.
Know queries are typically informational and not commercial or transactional. Although they may involve product research, users are not yet ready to make a purchase. Examples of purely informational queries include [how long does it take to drive to London] or [Gabriel Macht IMDb].
While these queries might not hold the same commercial weight as direct transactional queries, they still offer value to users, which is what Google prioritizes. For instance, a user planning a vacation might start with a search for [winter sun holidays Europe] and then refine their search to specific destinations. If your website provides the information they need, they may eventually reach out for more details or inquiries.
Featured Snippets & Clickless Searches
Rich snippets and specialized content blocks, such as featured snippets, have long been crucial for SEO, and securing a spot in these areas can significantly boost traffic to your website.
However, appearing in position zero might sometimes result in users not clicking through to your site. This means you miss out on potential traffic, user engagement, and ad impressions.
Nevertheless, landing in these prominent positions is highly valuable for click-through rates and offers a fantastic opportunity to introduce new users to your brand or website.
“Go” queries usually involve searches for specific brands or known entities, where users are looking to visit a particular website or location.
For example, if a user searches for Kroger, showing them results for Food Lion would not fulfill their intent as effectively.
Similarly, if your client aims to rank for a competitor’s brand term, you need to consider why Google would direct users to their site when they are clearly searching for the competitor.
This is also an important factor to consider during rebrand migrations, including the implications and intent behind the new brand name.
Understanding the customer journey has long been essential for crafting marketing campaigns and designing websites.
While developing personas and planning user navigation are crucial, it’s equally important to grasp how users search and where they are in their journey.
The term “journey” often implies a linear path, with typical routes like landing page > form or homepage > product page > form. This linear perspective often guides how we design website architecture.
However, mobile and voice searches have introduced new complexities, significantly altering user behavior and decision-making almost overnight. For instance, Google acknowledged this shift in 2015 by expanding mobile-friendliness as a ranking factor, even before the well-known Mobilegeddon update.
These “micro-moments” challenge our traditional view of the user journey. Users now search in diverse ways, and Google’s evolving search results mean there is no single standard page of results.
To understand the stage of a user’s journey, we can analyze the search results provided by Google and review data from tools like Google Search Console, Bing Webmaster Tools, and Yandex Metrica.
It’s crucial to recognize that search intent and the search results displayed by Google can shift rapidly.
A notable example of this is the Dyn DDoS attack in October 2016. Unlike previous DDoS attacks, this one received widespread media attention, with even the White House issuing a statement.
Prior to the attack, searches for terms like [ddos] or [dns] would yield technical results from companies such as Incapsula, Sucuri, and Cloudflare, which were suited to a technical audience.
However, as the attack unfolded, the nature of these queries shifted from commercial or transactional to informational. Within 12 hours, search results transformed to feature news articles and blog posts explaining what a DDoS attack is.
This incident underscores the importance of not just optimizing for keywords that drive conversions but also for those that offer user value and relevance to current topics.
The Dyn attack demonstrates that while search intent can evolve quickly during major news events, such shifts can sometimes become permanent if they capture significant user interest and engagement.
After analyzing client Search Console profiles and examining keyword trends, we’ve identified a notable pattern over the past year.
For our home electronics client, we’ve observed a rise in queries beginning with “how,” “what,” or “does,” which have expanded on previous query types.
For instance, instead of the traditional query format like [manufacturer model feature], there is now a growing trend toward queries such as [does manufacturer model have feature].
Historically, search queries followed a fairly consistent pattern, allowing Google to develop effective intent classifiers. This insight aligns with Google’s patent for automatic query pattern generation.
To accomplish this, Google needs to annotate the query using various elements such as a language identifier, stop-word remover, confidence values, and entity identifier.
For example, as shown in the image above, the query [proxy scraping services] encompasses several related queries and variations. Although these variations haven’t been explicitly searched for, results for queries like [proxy services], [scraping services], and [proxy scraping services] may overlap significantly. Handling these overlaps could strain resources if the results for each query are returned separately.
This issue is particularly relevant as AI and evolving technologies may influence how users conduct searches. Providing additional context to large language models (LLMs) is crucial to meet our needs effectively.
Our language tends to become more conversational and detailed as we aim to be explicit about our goals, a point highlighted in Vincent Terrasi’s ChatGPT prompt guide.
If this trend becomes widespread, it could lead to changes in how Google and other search engines process query types, potentially altering the current search engine results page (SERP) structures.
On the flip side, the increase in diverse and abundant content on websites can significantly impact and shift search behavior.
Additionally, the way platforms promote new tools and features plays a crucial role in these changes. For instance, Google’s high-profile, celebrity-endorsed campaigns for Circle to Search illustrate how marketing can drive shifts in search dynamics.
As machine learning continues to advance, coupled with Google’s evolving algorithms, we can expect changes in search results pages.
This evolution might prompt Google to experiment with new search result page formats (SCRBs) and other features across various sectors, such as financial product comparisons, real estate, or automotive.
Original news from SearchEngineJournal
