Within Search Ranking

Why the top search result feels right

Higher search results often attract clicks because they are easier to see, not only because they are better answers.

On this page

  • What position bias means in search
  • How ranking changes clicks and trust
  • How to check beyond the first result
Preview for Why the top search result feels right

Introduction

When people use a search engine, the first result often feels like the best answer. Sometimes it is. But research shows that top-ranked results attract attention and clicks partly because they are placed at the top, not only because they are more relevant. This tendency is known as position bias. It is one of the most important ways search ranking influences everyday interaction with AI-driven systems.

Position bias illustration 1 Position bias matters because search engines do not merely display information; they arrange it. Once results are ordered, many users give extra attention and trust to the highest positions. As a result, ranking can shape which sources are read, which explanations are considered credible, and which information remains unseen. [ACM Digital Library]dl.acm.orgACM Digital LibraryClick-Conversion Multi-Task Model with Position Bias…Position bias, the phenomenon whereby users tend to focus on h…

Position bias is the tendency for people to click items that appear higher in a ranked list simply because they are easier to notice. Researchers studying search behaviour define it as a situation where the probability of a click is influenced by position independently of the result’s actual relevance. [ResearchGate]researchgate.netAn experimental comparison of click position-bias modelsApr 25, 2026 — This paper focuses on explaining that bias, modelling…

This does not mean top results are poor. Search engines invest heavily in ranking useful pages, and higher results are often genuinely relevant. The key point is that clicks reflect two forces at once:

  • The quality or relevance of the result.
  • The visibility created by its position.

Because those influences are mixed together, a click is not a perfect signal that a result was objectively the best choice. [Google Research]research.google.comGoogle ResearchPosition Bias Estimation for Unbiased Learning to Rank in…by X Wang · 2018 · Cited by 363 — We propose a regression- ba…

A simple thought experiment illustrates the issue. If two equally useful pages swap positions on a results page, the page moved upward will usually receive more clicks. The change in behaviour comes from placement, not from any improvement in content.

How ranking changes clicks and trust

Position bias works because people have limited attention. Search results are scanned quickly, and users often stop looking once they find something that appears satisfactory.

Eye-tracking studies have repeatedly found that users concentrate most of their visual attention near the top of search results pages. Higher results are examined earlier and more frequently than lower ones. [MDPI+2ResearchGate]mdpi.comA framework is created to examine the effectiveness of eye-…Read more…

Large-scale behavioural data show the same pattern. In one study of real-world search activity, more than 97% of Google clicks and more than 99% of Bing clicks occurred on the first page of results. Users overwhelmingly interacted with what appeared first rather than exploring deep into the rankings. [PMC]pmc.ncbi.nlm.nih.govPMCYou are how (and where) you search?Comparative analysis…by A Urman · 2023 · Cited by 50 — We found that on both engines participants clicked disproportionately more on t…

Trust plays an important role as well. Research examining how people evaluate search results found that users often place confidence in the search engine’s ordering itself. A well-known eye-tracking study reported that participants tended to trust Google’s ranking more than their own independent assessment of alternatives. [Information Research]informationr.netInformation ResearchWe still trust in Google, but less than 10 years agoby S Schultheiss · 2018 · Cited by 85 — Their eye-tracking experi…

This creates a subtle feedback effect:

  1. The search engine ranks a result highly.
  2. Users notice it first.
  3. Users infer that it is probably important or trustworthy.
  4. The result receives more clicks because it was highly ranked.

The ranking becomes a cue that influences judgement before the page has even been read.

Position bias illustration 2

Visibility is not the only reason

Researchers also note that presentation matters. Titles, snippets, formatting and other visual features can make some results appear more attractive than others. Google researchers have described this as a form of presentation bias that operates alongside position bias. A result may gain clicks because it is higher on the page, because its summary looks appealing, or both. [Google Research]research.google.comGoogle ResearchBeyond Position Bias: Examining Result Attractiveness as…by Y Yue · Cited by 265 — Our ex- periments conducted on the G…

This is why click counts alone can be misleading. A heavily clicked result may be genuinely useful, unusually eye-catching, highly ranked, or some combination of all three.

Why position bias matters for understanding AI

Search ranking is often one of the most familiar AI outputs people encounter. Position bias reveals that the influence of AI is not limited to generating answers. It can also arise through ordering and presentation.

Researchers have shown that ranking choices can affect users’ knowledge, beliefs and decisions. When people rely on search results to learn about a topic, the information placed near the top gains an advantage in visibility and influence. [PMC]pmc.ncbi.nlm.nih.govThe Impact of Search Engine Selection and Sorting Criteria on…by A Allam · 2014 · Cited by 187 — This study is concerned with demon…

Importantly, this influence does not require deception or manipulation. Even a ranking system designed in good faith can steer attention simply because humans must decide where to focus first. Position bias emerges from the interaction between the ranking algorithm and normal human behaviour.

For AI systems, this creates a challenge. Engineers often use clicks as feedback when improving rankings. Yet clicks themselves are affected by position bias. A result may receive many clicks because it was shown first, making it difficult to separate genuine relevance from the advantage of visibility. This is why search researchers devote significant effort to measuring and correcting for position bias in ranking systems. [Google Research+2ResearchGate]research.google.comGoogle ResearchPosition Bias Estimation for Unbiased Learning to Rank in…by X Wang · 2018 · Cited by 363 — We propose a regression- ba…

How to check beyond the first result

Position bias does not mean top results should be ignored. Many top-ranked results are excellent. The risk is assuming that rank alone proves quality.

A few habits can reduce the influence of position bias:

  • Open more than one result before deciding.
  • Compare sources that appear at different ranks.
  • Pay attention to the content of a page rather than its position.
  • Notice when snippets, headlines or formatting are shaping your first impression.
  • For important topics, especially health, finance or public affairs, check whether multiple independent sources agree.

These steps do not eliminate bias, but they help separate the quality of an answer from the convenience of its placement.

Position bias illustration 3

The key lesson

Position bias is a reminder that visibility and quality are not the same thing. Higher-ranked search results receive more attention partly because people naturally focus on what appears first. Eye-tracking studies, click data and behavioural research all point to the same pattern: ranking affects what users see, what they trust and what they choose to read. [MDPI+2PMC]mdpi.comA framework is created to examine the effectiveness of eye-…Read more…

Understanding this mechanism makes search ranking easier to evaluate critically. The top result may indeed be the best answer—but its position is one reason it feels that way. [Information Research]informationr.netInformation ResearchWe still trust in Google, but less than 10 years agoby S Schultheiss · 2018 · Cited by 85 — Their eye-tracking experi…

Amazon book picks

Further Reading

Books and field guides related to Why the top search result feels right. Use these as the next step if you want deeper reading beyond the article.

eBay marketplace picks

Marketplace Samples

Example marketplace items related to this page. Use the search link to explore similar finds on eBay.

Using USA

Endnotes

  1. Source: dl.acm.org
    Link: https://dl.acm.org/doi/10.1145/3539618.3591963
    Source snippet

    ACM Digital LibraryClick-Conversion Multi-Task Model with Position Bias...Position bias, the phenomenon whereby users tend to focus on h...

  2. Source: pmc.ncbi.nlm.nih.gov
    Link: https://pmc.ncbi.nlm.nih.gov/articles/PMC4004139/
    Source snippet

    The Impact of Search Engine Selection and Sorting Criteria on...by A Allam · 2014 · Cited by 187 — This study is concerned with demon...

  3. Source: researchgate.net
    Link: https://www.researchgate.net/publication/200110550_An_experimental_comparison_of_click_position-bias_models
    Source snippet

    An experimental comparison of click position-bias modelsApr 25, 2026 — This paper focuses on explaining that bias, modelling...

  4. Source: research.google.com
    Link: https://research.google.com/pubs/archive/3bace79f9bcead0b20dec31e2a0878346ad2fb0d.pdf
    Source snippet

    Google ResearchPosition Bias Estimation for Unbiased Learning to Rank in...by X Wang · 2018 · Cited by 363 — We propose a regression- ba...

  5. Source: mdpi.com
    Link: https://www.mdpi.com/2078-2489/11/6/300
    Source snippet

    A framework is created to examine the effectiveness of eye-...Read more...

  6. Source: researchgate.net
    Link: https://www.researchgate.net/publication/2917730_Eye-Tracking_Analysis_of_User_Behavior_in_WWW-Search
    Source snippet

    The goal is to gain insight into how users browse the...Read more...

  7. Source: researchgate.net
    Link: https://www.researchgate.net/publication/221517357_What_are_you_looking_for_An_eye-tracking_study_of_information_usage_in_Web_search
    Source snippet

    An eye-tracking study of information usage in Web searchIn order to better understand how and why users were clicking on search results...

  8. Source: pmc.ncbi.nlm.nih.gov
    Title: PMCYou are how (and where) you search?
    Link: https://pmc.ncbi.nlm.nih.gov/articles/PMC10155157/
    Source snippet

    Comparative analysis...by A Urman · 2023 · Cited by 50 — We found that on both engines participants clicked disproportionately more on t...

  9. Source: research.google.com
    Link: https://research.google.com/pubs/archive/36363.pdf
    Source snippet

    Google ResearchBeyond Position Bias: Examining Result Attractiveness as...by Y Yue · Cited by 265 — Our ex- periments conducted on the G...

  10. Source: researchgate.net
    Link: https://www.researchgate.net/publication/221022091_Beyond_position_bias_examining_result_attractiveness_as_a_source_of_presentation_bias_in_clickthrough_data
    Source snippet

    Beyond position bias: examining result attractiveness as a...24 Sept 2018 — In this paper, we examine result summary attract...

  11. Source: researchgate.net
    Link: https://www.researchgate.net/publication/322969246_Position_Bias_Estimation_for_Unbiased_Learning_to_Rank_in_Personal_Search
    Source snippet

    Traditional click models aim to extract the ‹query...Read more...

  12. Source: cacm.acm.org
    Link: https://cacm.acm.org/research/are-people-biased-in-their-use-of-search-engines/
    Source snippet

    voring items at the top of results lists, though they also seek out high-relevance items...

  13. Source: researchgate.net
    Link: https://www.researchgate.net/publication/261844807_Incorporating_vertical_results_into_search_click_models
    Source snippet

    (PDF) Incorporating vertical results into search click modelsNov 18, 2014 — We also performed eye-tracking analysis to study user's real...

  14. Source: researchgate.net
    Link: https://www.researchgate.net/publication/361415642_Search_engine_effects_on_news_consumption_ranking_and_representativeness_outweigh_familiarity_on_news_selection
    Source snippet

    (PDF) Search engine effects on news consumptionThe ranking effect is so widely accepted that heavy weights are used in studies that inves...

  15. Source: researchgate.net
    Link: https://www.researchgate.net/publication/341908333_Eye-Tracking_Studies_of_Web_Search_Engines_A_Systematic_Literature_Review
    Source snippet

    A framework is created to examine the...Read more...

  16. Source: researchgate.net
    Link: https://www.researchgate.net/publication/323753668_Position_Bias_in_Recommender_Systems_for_Digital_Libraries
    Source snippet

    Position Bias in Recommender Systems for Digital LibrariesFor example, position bias in search rankings strongly influences how many clic...

  17. Source: google.com
    Link: https://www.google.com/intl/en_us/search/howsearchworks/how-search-works/ranking-results
    Source snippet

    How Does Google Determine Ranking ResultsDiscover how key factors such as meaning, relevance, and quality are used to generate how websit...

  18. Source: research.google
    Link: https://research.google/pubs/position-bias-estimation-for-unbiased-learning-to-rank-in-personal-search/
    Source snippet

    The extracted bias can improve the learning-to-...Read more...

  19. Source: informationr.net
    Link: https://informationr.net/ir/23-3/paper799.html
    Source snippet

    Information ResearchWe still trust in Google, but less than 10 years agoby S Schultheiss · 2018 · Cited by 85 — Their eye-tracking experi...

Additional References

  1. Source: anneschuth.nl
    Link: https://anneschuth.nl/assets/hofmann-effects-2014.pdf
    Source snippet

    up complexity to cover assumptions that may affect evaluation outcomes based on the resulting clicks on...Read more...

  2. Source: github.com
    Link: https://github.com/NovaSearch-Team/position-bias-in-IR
    Source snippet

    NovaSearch-Team/position-bias-in-IR: An Empirical Study...This repository accompanies our paper, which investigates position bias—a pers...

  3. Source: amazon.science
    Title: modeling position bias ranking for streaming media services
    Link: https://www.amazon.science/publications/modeling-position-bias-ranking-for-streaming-media-services
    Source snippet

    by M Ruffini · 2022 · Cited by 9 — In this paper, we present a novel position bias estimator that overcomes these limitations: it can be...

  4. Source: youtube.com
    Link: https://www.youtube.com/watch?v=qlpOeJMaEp8
    Source snippet

    The Psychology of Search: Leveraging Cognitive Bias in SEOGarrett Sussman, Director of Marketing at iPullRank, explores how cognitive bia...

  5. Source: merriam-webster.com
    Title: POSITIO N Definition & Meaning4 days ago — 1
    Link: https://www.merriam-webster.com/dictionary/position
    Source snippet

    an act of placing or arranging: such as a: the laying down of a proposition or thesis b: an arranging in order 2. a point of view adopt...

  6. Source: shadecoder.com
    Title: position bias a comprehensive guide for 2025
    Link: https://www.shadecoder.com/ja/topics/position-bias-a-comprehensive-guide-for-2025
    Source snippet

    Position Bias: A Comprehensive Guide for 2025Jan 2, 2026 — Position bias affects how users perceive and interact with lists, search resul...

  7. Source: research.tudelft.nl
    Title: This phenomenon – known as position bias –
    Link: https://research.tudelft.nl/files/96007152/3404835.3462851.pdf
    Source snippet

    TU Delft Research PortalExploring Why Biased Search Result Rankings Affect User...by T Draws · 2021 · Cited by 87 — when search result r...

  8. Source: youtube.com
    Title: Unbiased Learning to Rank: Counterfactual and Online Approaches
    Link: https://www.youtube.com/watch?v=BEEfMrn9T9c
    Source snippet

    Doubly-Robust Estimation for Correcting Position-Bias in Clicks for Unbiased Learning to Rank...

  9. Source: oreilly.com
    Title: Eye Tracking: How Users Scan Results Pages
    Link: https://www.oreilly.com/library/view/the-art-of/9780596809133/ch01s07.html
    Source snippet

    The Art of...Users' eyes are drawn to bold keywords, titles, and descriptions in the natural (“organic”) results versus the paid search...

  10. Source: youtube.com
    Link: https://www.youtube.com/watch?v=bmhClxLzzxI
    Source snippet

    Recent Advances in Unbiased Learning to Rank from Position-Biased Click Feedback...

Topic Tree

Follow this branch

Parent topic

Search Ranking Why search results are an AI output

Related pages 2