Google started as a simple search engine in the late 1990s, allowing users to search the internet for information. However, over the past two decades, Google has become more than a search engine. Today, Google aims to directly answer users’ questions rather than simply providing links to web pages. This transition reflects Google’s ambitions to evolve from a search engine to an “answer engine” that can have natural conversations with users.
In this lengthy article, we will explore Google’s transformation in depth. We will cover:
- A brief history of Google and its early days as a search engine
- The launch of Google Assistant and other AI-driven products
- How Google is leveraging knowledge graphs and semantic search capabilities
- The introduction of featured snippets and direct answers
- How Google aims to understand intent and context behind queries
- The rise of conversational interfaces and voice search
- Challenges facing Google as it aims to become an answer engine
- What the future may hold for Google as it continues to evolve
By the end, you will understand how Google has transformed over the past 20+ years and its ambitions to become an AI-driven answer engine. This journey has profound implications for Google and the future of search and information retrieval.
A Brief History of Google’s Early Days as a Search Engine
Google was founded in 1998 by Larry Page and Sergey Brin, two PhD students at Stanford University [1]. The company started as a research project with a mission to organize the world’s information and make it universally accessible and useful [2].
At the time, existing search engines like AltaVista, Lycos, and Yahoo! relied on basic keyword matching, which often produced irrelevant results. Page and Brin aimed to fix this problem through their new search engine, Backrub. It analyzed relationships between websites through link ranking results based on the PageRank algorithm [3].
After renaming their engine to Google, the company launched google.com in 1998 from a friend’s garage in Menlo Park, California [1]. Google’s spartan homepage stood out from cluttered portals like Yahoo! with its focus on search [4]. Features like autocomplete, and relevant results helped Google deliver a fast, streamlined search experience.
Google’s PageRank algorithm proved superior at delivering relevant results compared to rival search engines. By 2000, Google was handling 18 million daily searches [5]. Its clean interface, speed, and accuracy made it a hit with early internet users.
Key milestones in Google’s early history as a search engine include:
- 1998: Google is founded by Larry Page and Sergey Brin and incorporated in California [1]
- 1998: The company launches the initial version of Google search at google.com [4]
- 1999: Google receives $25 million in funding from venture capitalists [1]
- 2000: Google launches AdWords, an online advertising platform [5]
- 2002: Google launches Google News, an automated news aggregation service [6]
- 2004: Google goes public, raising $1.67 billion in its IPO [1]
By the mid-2000s, Google was firmly established as the world’s leading search engine. But the company set its sights far beyond just organizing web pages. Next, we’ll explore Google’s transformation into an “answer engine” providing direct responses to queries.
The Launch of Google Assistant and other AI Products
In 2016, Google signaled its evolution from a traditional search engine into an “answer engine” with the launch of Google Assistant [7]. Google Assistant is an AI-powered virtual assistant that can engage in two-way conversations with users.
Unlike the Google search engine, which simply returns links to web pages, Google Assistant can understand natural language requests and provide direct answers by tapping into Google’s vast knowledge graph. For example, users can ask questions like “How old is the President?” and get a specific response without clicking results.
Google Assistant marked a major milestone in the company’s shift towards conversational interfaces. It built on Google’s earlier virtual assistant experiments like Google Voice Search and Google Now [8].
Other key launches on Google’s journey to becoming an answer engine include:
- 2011: Google launches Knowledge Graph, a knowledge base that enhances search results with semantic information [9].
- 2016: The Google Home smart speaker is released, allowing users to interact with Google Assistant using their voice [10].
- 2017: Google Lens image recognition tool launches, letting users search what they see through their smartphone camera [11].
- 2020: Google announces the Multitask Unified Model (MUM), its next-generation AI system aimed at complex, multifaceted queries [12].
- 2021: LaMDA, Google’s Language Model for Dialogue Applications, hints at the company’s progress in conversational AI [13].
Through these offerings, Google aims to understand the intent behind queries and provide increasingly comprehensive answers without users needing to click results or reformulate searches.
Leveraging Knowledge Graphs and Semantic Search Capabilities
Its evolving ability to parse meaning and context is central to Google’s shift from a search engine to an answer engine. Key to this is the Knowledge Graph and semantic search capabilities.
Launched in 2012, the Google Knowledge Graph is a knowledge base containing 500 million entities and billions of facts about people, places, and things [14]. It draws information from sources like Wikipedia, CIA World Factbook, Wikidata, etc.
The Knowledge Graph powers Google’s semantic search features. When users search for entities like “Tom Cruise,” Google can now provide a sidebar knowledge panel with key facts rather than just search results [14]. This helps Google understand the contextual meaning behind queries.
Importantly, the Knowledge Graph also enables Google to make connections between related entities [15]. This means users can explore broader concepts through search, not just look for specific web pages.
Google also leverages semantic search capabilities like RankBrain, which interprets word meanings and contexts to better match search intents [16]. Launched in 2015, RankBrain helps Google handle never-before-seen queries by understanding meaning, not just keywords [16].
These semantic capabilities aim to move beyond keyword matching and result rankings. They help Google interpret billions of searches in context every day, bridging the gap between search queries and results [17]. This brings Google closer to understanding true user intent.
The Introduction of Featured Snippets and Direct Answers
In its mission to become an answer engine, Google has also introduced featured snippets – summarizing key information in response to searches without requiring a clickthrough.
Featured snippets first appeared above Google search results in 2014 [18]. These summarizing excerpts aim to provide immediate answers so users don’t need to visit web pages. Snippets draw information from third-party sites and condense it into an easy-to-consume format.
Google also directly answers some queries with knowledge panels, calculator conversions, dictionary definitions, and more [19]. The system can simply provide a voice response to Google Assistant’s voice queries.
Featured snippets and direct answers demonstrate how Google is evolving beyond search to fulfill users’ informational needs without clicks or page visits. Studies show that nearly half of Google searches end without users clicking results [20].
Key examples of Google’s direct answers include:
- Calculator: Converts queries like “35 divided by 5” into an answer.
- Dictionary: Defines words and terms like “ephemeral.”
- Knowledge panel: Provides direct answers to entity queries like “population of London.”
- Weather: Responds with a forecast when users search for weather.
- Sports scores: Surfaces live scores for sports teams and ongoing games.
- Unit conversions: Automatically converts between units, like “50 kg to pounds.”
These efforts save users time and aim to make Google feel more conversational by providing information directly.
Understanding Intent and Context Behind Queries
Traditional search engines relied solely on matching keywords to produce relevant results. Google aims to further interpret the intent and contextual meaning behind queries.
This is enabled by artificial intelligence techniques like natural language processing (NLP), neural networks, and machine learning. Tools like Bidirectional Encoder Representations from Transformers (BERT) help Google better understand language [21].
Specifically, Google is trying to identify whether queries have:
- Informational intent, seeking knowledge or facts
- Navigational intent: wanting to visit a certain website or page
- Transactional intent, aiming to perform an action like purchasing a product [22]
This helps Google determine whether to provide direct answers, web results, or take action.
Google can also parse context and meaning. For example, the query “mercury” could refer to the planet, element, automobile brand, music artist, or more. Google aims to understand which one is based on contextual signals [15].
Identifying intent and context helps Google move from simplistic keyword matching to more human-like comprehension. The next search phase aims to match keywords and deeply understand them.
The Rise of Conversational Interfaces and Voice Search
In another shift away from the traditional search box, Google is adapting its engine for conversational interfaces and voice search.
Google maintains over 70% market share in voice search, fueled by the rise of smart speakers like Google Home [23]. It is now training its algorithms on these spoken queries, which are often more natural and conversational [24].
Google Duplex, the eerily human-like AI system for booking appointments over the phone, further demonstrates Google’s advancements in conversational AI [25]. Features like Continued Conversation keep micro-conversations going without [26].
The rise of voice search encourages people to talk to Google as they would another person. Processing these natural language queries requires Google to advance its semantic capabilities and contextual understanding.
Voice queries also often begin with pronouns like “What is…” or “Tell me…” demonstrating how users increasingly view Google as an informative agent rather than just a search box [27]. Google aims to provide spoken answers, not just links, in response.
Google is also exploring search improvements targeted specifically for phones, such as Lens search and multimodal experiences combining text, voice, and touch [28]. Conversational interfaces are key to the future of search.
Challenges Facing Google in Becoming an Answer Engine
Although Google has made significant progress in evolving into an answer engine, major challenges remain:
- Answering open-ended queries. Google excels at factual questions but struggles with more complex, open-ended queries [29]. Teaching AI to reason and think critically is an immense challenge.
- Scaling knowledge. Google’s knowledge graph contains trillions of facts but pales compared to the scope of human knowledge and experience [30]. Teaching AI common sense and general world knowledge poses difficulties.
- Understanding context and intent. Despite advances in natural language processing, Google still struggles to fully understand the context and intended meaning behind ambiguous or uncommon queries [31].
- Weeding out misinformation. As Google provides direct answers, it risks disseminating incorrect or biased information. More rigorous vetting processes are needed [32].
- Preserving privacy. Google’s vast troves of user data raise growing privacy concerns, especially as its products become more human-like in their understanding [33]. Developing ‘privacy-preserving’ AI is critical.
- Maintaining search engine heritage. As Google evolves, it must balance its search engine foundations, like fast page loads and website crawling, with its expanding role as an answer engine [34]. Search underpins Google’s knowledge.
Despite progress, Google has yet to achieve human-level language mastery. Perfect comprehension of user intent remains elusive. However, incremental improvements in semantic capabilities are helping bridge this gap.
The Future: What’s Next for Google’s Evolution
What does the future look like as Google continues evolving from a search engine to an answer engine? Here are some possibilities:
- More conversational interfaces. Voice assistants like Google Assistant will keep improving through speech recognition and synthesis advances. Expect multi-turn dialogue for more human-like interactions [35].
- Deeper personalization. By integrating personal data, Google may eventually move beyond generalized information to provide bespoke answers tailored specifically to individual users [36].
- Predictive search. Google may begin anticipating users’ needs, providing information, or taking action without a query. This could include reminders, notifications, and personalized recommendations [37].
- Intelligent agents. Future iterations of the Google Assistant could act as personalized agents that know users intimately and help with tasks like planning travel, shopping, and more [38].
- Integration with ambient computing devices. As home devices with integrated voice assistants proliferate, Google will increasingly deliver search results conversationally while on the go, not just via screens [39].
- Focus on multimedia. Future searches may rely more heavily on images, audio, and video than text. Google is already enhancing its algorithms to process visual information [40].
Some experts predict search itself may evolve into a predictive, omnipresent function that feels more like having a superintelligent assistant than querying a search box. But challenges remain around bias, transparency, privacy, and security [41].
While the future is uncertain, Google’s vision is clear: evolving search from retrieving relevant web pages into a conversational experience that intuitively understands and answers user needs. The coming years promise to redefine how we discover, interact with, and consume information. Google aims to lead this shift.
Analysis and conclusion
Google’s objective is to deliver search results without requiring clicks. They scrape content, anchoring you to their domain with ‘people also ask,’ thereby stripping away even more content. It has transitioned from a search engine for discovering websites to an answer engine. Yes, we can secure a top 3 position in SERPs and still receive zero clicks. The situation is only set to worsen.
The new strategy involves creating content on other platforms, not your domain. This mirrors what Google does, akin to social media platforms (Twitter, Facebook, etc.): they avoid showcasing posts with links in feeds, and with ‘x/twitter,’ they even obscure URLs to appear as an image, not a hyperlink.
Similar to these platforms, Google relocates all our content to its domain. That’s why I’ve discontinued SEO efforts. It’s a futile mission. I engaged in it for a decade. The landscape has evolved, and those with the audience are not motivated to direct them towards you.
If you activate AI labs, you’ll notice images sourced from other websites at the top. This necessitates numerous clicks to access the website. An AI result generates content from various sources, then prompts ‘see ‘x’ for more details,’ requiring two clicks to reach the website (often not the 1st SERP), followed by ads, and then ‘people also ask,’ which comprises more appropriated content. Additionally, on mobile, the second results page is replaced by Google Discover.
It’s simply bewildering. I can’t fathom that Google still portrays itself as a search engine. Furthermore, the notion that ‘Google’s useful content update hasn’t discovered hidden treasures yet’ is downright laughable. Really, Google? You’re suggesting that the ‘hidden gems,’ which you initially introduced, will resurface at some point? Remarkable…
Google has primarily directed its efforts towards profits by generating margins from tools it formerly provided for free. Contemplate the ‘free forever’ email we enjoyed for years, now accessible only for a fee (recently increased). Universal was supplanted by GA4, necessitating the legitimate use of BigQuery for identical functionalities. Consider also the alterations to AdWords (now Google Ads), where long-tail queries that once cost a few cents now command a dollar, influencing search result quality based on profit, etc? I couldn’t comprehend why Google would jeopardize its reputation as synonymous with the Internet for a slightly higher profit. Yes, we’re discussing billions and billions in profits… but wouldn’t preserving the brand as the epitome of the Internet is paramount?
To the best of my understanding, Alphabet acknowledges that they will inevitably, and possibly sooner than later, be compelled to separate due to prevailing American political perspectives on monopolies. If you foresee the end approaching, it’s the opportune moment to amass as much profit as possible. Just a notion… but perhaps this also elucidates the flurry of rapid version changes over the past 30 days.
On another note, I initially focused on how Google heightened its profit margin through escalating costs. I hadn’t contemplated the alternative of augmenting margins by curbing costs.
Indexing the Internet incurs expenses based on the number of pages. Sorting query results also entails costs. Perhaps this is why some of us grapple with indexing our content. If the search volume is meager, Google simply doesn’t bother addressing those searches. There’s substantiated evidence for this line of thinking… the person who introduced ‘suggested searches’ or autocomplete suggestions when typing a query is brilliant for various reasons, including directing people towards filtered cached results. Reduced costs.
Regardless, if it holds that the company is compelled to separate, it could bring considerable relief to webmasters.
References:
[1] From the Garage to the Googleplex (accessed on 10/17/22)
https://about.google/our-story/
[2] Our Mission (accessed on 10/17/22)
https://about.google/mission/
[3] The Story of the Google Search Engine (accessed on 10/17/22)
https://www.thoughtco.com/the-story-of-the-google-search-engine-1991640
[4] Google’s Simple Homepage Evolution (accessed on 10/17/22)
https://gizmodo.com/googles-simple-homepage-evolution-1999-to-now-1657948028
[5] How Search Works (accessed on 10/17/22)
https://www.google.com/search/howsearchworks/crawling-indexing/
[6] Google Milestones (accessed on 10/17/22)
https://www.google.com/about/our-story/
[7] Meet Your Google Assistant (accessed on 10/17/22)
https://blog.google/products/assistant/meet-your-google-assistant/
[8] The Evolution of Google Assistant (accessed on 10/17/22)
https://www.creativestrategies.com/evolution-of-google-assistant/
[9] Introducing the Knowledge Graph (accessed on 10/17/22)
https://blog.google/products/search/introducing-knowledge-graph-things-not/
[10] Google Home: An Introduction to Google’s Smart Speaker (accessed on 10/17/22)
https://www.lifewire.com/what-is-google-home-4684985
[11] Google Lens: What Is It and How Do You Use It? (accessed on 10/17/22)
https://www.techlicious.com/guide/what-is-google-lens/
[12] A More Helpful Google, for You (accessed on 10/17/22)
https://blog.google/products/search/helpful-google/
[13] Google’s LAMDA – Towards Safe, Grounded, and High-Quality Dialog Models (accessed on 10/17/22)
https://ai.googleblog.com/2022/02/lamda-towards-safe-grounded-and-high.html
[14] Knowledge Graph: Now Powering Search On Over 1 Billion Queries A Day (accessed on 10/17/22)
https://www.blog.google/products/search/knowledge-graph-now-powering-search/
[15] Search Quality Evaluator Guidelines (accessed on 10/17/22)
https://static.googleusercontent.com/media/guidelines.raterhub.com/en//searchqualityevaluatorguidelines.pdf
[16] Google’s RankBrain Algorithm (accessed on 10/17/22)
https://www.blog.google/products/search/search-quality-highlights-35-changes/
[17] Natural Language Processing in Search (accessed on 10/17/22)
https://www.google.com/search/howsearchworks/crawling-indexing/nlp.html
[18] The Definitive Guide to Featured Snippets in Google Search (accessed on 10/17/22)
https://www.searchenginejournal.com/featured-snippets-guide-google-search/426490/#close
[19] Straight Answers Information in Search (accessed on 10/17/22)
https://www.google.com/search/howsearchworks/responses/
[20] Nearly Half of Google Searches Now End Without a Click to Other Content, Study Finds (accessed on 10/17/22)
https://sparktoro.com/blog/less-than-half-of-google-searches-now-result-in-a-click/
[21] BERT Explained: A State-of-the-Art Language Model for NLP (accessed on 10/17/22)
https://developers.googleblog.com/2018/11/bert-state-of-art-natural-language.html
[22] Identifying User Intent (accessed on 10/17/22)
https://developers.google.com/search/docs/guides/identifying-user-intent
[23] comScore Releases April 2021 U.S
