Zero-click search, that is why Google doesn't want to click on the found results?

Zero-click search, that is why Google doesn't want to click on the found results?

More and more searches are not driving traffic to the pages found, in other words, fewer and fewer people are clicking on the results. Why is this happening? Is Google showing only the ads themselves, so that none of the information is interesting enough for the user? Nothing could be more wrong.

Google and other search engines want to find information as quickly as possible. At the same time, it is in their interest to keep users within the domain and related products, where they can collect further data, display advertisements, or offer products.

For this purpose, search engines have long gone beyond the formula of a simple list of links to other websites. They show more and more information within their interface: from presenting profiles of famous personalities, through showing the weather, to answering more complicated questions.

bing_search_visualisation
google_search_weather
duckduckgo_search_wiki

Zero-click search is all searches that do not require clicks on the found items. The user finds the information he is looking for already on the results page.

Although the name says "zero", it is customary not to take into account clicks on animations or interactive elements of the search engine interface, as they do not lead the user beyond the domain / page with results.

As reports show, already in 2019 50% of all searches were of the "zero-click" type [1], while counting the data from the entire year 2020, it was already about ⅔. [2]

Although the data comes from various sources and cannot be directly compared without taking into account the methodology with which they were collected, the trend itself is noticeable.

As more and more searches are made by voice or from mobile devices, users are less and less likely to jump to specific results and expect an immediate response. Having a more accurate weather forecast to choose from with a map of moving low and high, and the more accessible one, shown in an accessible way, consistent with other search results, we will choose the last one.

Fragments with the answer

An increasingly common view in search engines is the fragment with the answer (called "featured snippets") appearing at the top of the list with the results.

google_search_featured_snippets

The algorithm finds answers in the content of the searched pages and displays them in the first place of the results, thanks to which the user immediately receives the answer.

This functionality statistically reduces the sum of all clicks on the linked pages [4] - the user does not need to take any further actions, as he has found the information he is looking for.

Critics note that search engines use third-party content to build their products on it. Without visiting the website with the publication, the creators are not able to monetize the traffic.

On the other hand, the possibility of appearing above all results is seen by SEO experts as an opportunity. The additional element affects the clickability of the rest of the results. Typically 26% of all clicks go to the top of the list. In the case of showing the fragment with the answer, it collects 8.6% of all clicks, while the first element under it gets only 19.6%. [4]

Although some of the techniques used do not require complex algorithms (e.g. displaying the weather from an external provider), the development of machine learning (ML), in particular in the area of ​​natural language processing (NLP), significantly contributed to the improvement of search results.

The use of BERT transformer architecture allowed Google to improve results in 1 out of 10 cases. [6] The algorithm enables a better understanding of the context of the analyzed content and user queries.

The same technology is behind huge language models, such as the famous GPT-3 or Switch Transformer, which can cost millions of dollars to train (the cost of GPT-3 training is about $ 4.6 million [7]). Models are characterized by extraordinary skills, such as writing entire articles or code that is deceptively similar to those produced by a human being.

Fight against monopoly

It is estimated that 94% of all searches are done through Google.

Google, as part of SEO, has been promoting pages that properly structure their data for a long time. Adding appropriate tags by programmers to show Google robots which element of the page is what allows them to get a higher position in the search results. Not without significance is the fact that now this information structured by content owners is used by Google to improve its product.

It is impossible not to mention the struggles of media giants with technological corporations. The conflict between Google and Facebook and the Australian government recently demanded compensation for local media for their advertising revenue losses.

Publishers have a real dilemma, because search engines and social networks use their content in their products, on the other hand, it is tech-giants that are the main source of traffic for their articles.

Google is already working on the idea of ​​a new, better search engine that will be able to generate expert responses on its own instead of displaying search results. [5] The mechanism is to be based on models similar to the GPT-3, learned on all websites available from the Internet, and also constantly updated on new articles. The algorithm will provide information in the form of statements while providing quoted sources.

Company knowledge search engine

The example of Google illustrates situations where the searched resources are generally available, belong to a huge number of independent entities (individuals, publishers, companies), and there are so many queries that they can be analyzed, thanks to which they are in some way predictable, repeatable, and even in most relatively uncomplicated.

By reversing the situation and placing the search process in the context of organizations / companies, we are dealing with a completely different problem. In this case, a zero-click search can significantly cut costs - reducing the time spent by employees on manual searching of various sources and clicking through subsequent documents.

One of the solutions dedicated to the problem of finding information within the organization is the BOTWISE we create. We encourage you to follow us in the media and subscribe to the newsletter. If you see a potential application for an intelligent search engine in your company, please contact us :)

Sources:
[1] https://sparktoro.com/blog/less-than-half-of-google-searches-now-result-in-a-click/
[2] https://sparktoro.com/blog/in-2020-two-thirds-of-google-searches-ended-without-a-click/
[3] https://blog.google/products/search/google-search-sends-more-traffic-open-web-every-year/
[4] https://ahrefs.com/blog/featured-snippets-study/
[5] https://arxiv.org/pdf/2105.02274.pdf
[6] https://blog.google/products/search/search-language-understanding-bert/
[7] https://lambdalabs.com/blog/demystifying-gpt-3/