The 2021 Paper That Still Explains Why Google Beats Bing
A peer-reviewed academic review of search engine optimisation puts hard numbers on something most of us feel but rarely quote: Google holds 73.02 percent of desktop search users while Bing sits at just 9.26 percent. The paper, "Search Engine Optimization: A Review" by Firas Almukhtar, Nawzad Mahmood and Shahab Kareem, was published in Applied Computer Science (vol. 17, no. 1) in March 2021. I think it is worth revisiting now, because the mechanics it describes still explain why one engine wins and the other does not, and those mechanics quietly underpin the AI search tools everyone is chasing.
What the research actually found
The core finding is a comparison of how Google and Bing rank pages, and where they diverge. Both engines use many of the same signals. What separates them, the authors argue, is how those signals are weighted and applied.
The paper credits Michael Basilyan, a senior program manager at Bing, with naming content quality as a key priority in Bing's ranking phase. Bing weighs three attributes: the topical importance of a page, its meaning, and the consistency of the content. In plain terms, does the page answer the query, does the context fit the user, and can the content be trusted.
Google's 200-plus factors versus Bing's tighter net
On the Google side, the researchers note you cannot reduce ranking to a handful of levers. They reference SEO consultants citing more than 200 ranking factors, including keyword use, site structure, site speed, time on site, number of inbound links and inbound link quality.
The most practical distinction in the whole paper is this: Google's algorithm excels at matching synonyms and related terms to a query, while Bing needs more precise keyword matching to return accurate results. That single line explains a decade of ranking behaviour, and it maps directly onto how large language models now interpret meaning rather than exact strings.
The timeline behind the two engines
The history matters because it shows how long these habits have been baked in. Larry Page and Sergey Brin launched Google in 1998. Bing arrived in 2009 as the successor to Microsoft's MSN Search, then Windows Live Search, then Live Search.
| Attribute | Bing | |
|---|---|---|
| Launch year | 1998 | 2009 |
| Desktop search share | 73.02% | 9.26% |
| Query matching | Strong on synonyms and related terms | Favours precise keyword matching |
| Stated ranking factors | 200-plus signals | Topical importance, meaning, content consistency |
| Recency bias | Balanced | Favours pages tied to recent events |
| Flash content | Tends to rank lower | More likely to rank higher |
| Domain age | Less weighting noted | Weighs domain age more heavily |
The paper's own example illustrates the split. Searching for India's official tourism site, Google returned the official site first, while Bing led with news articles before the official page. Bing's tilt toward recency and domain age is a real difference in judgement, not just presentation.
How it works, and why it still matters for growth teams
Here is the part I keep coming back to. The authors define SEO as the mechanism by which a page is improved to maximise the frequency and quantity of organic traffic. They stress that ranking for keywords is fine, but meeting customers at each point of the purchasing process matters more.
From my observation, that framing has aged better than most 2021 marketing advice. The engines that reward meaning over exact keywords are now the same engines feeding AI overviews and chat answers. If you have followed how search behaviour is shifting toward conversational, non-English majority audiences, the lesson is identical: write for intent and comprehension, not for string matching.
Three actions the research points to
- Build for meaning, not exact strings. Google rewards synonyms and related terms. Cover a topic completely so the page answers the real question, not just the phrase typed.
- Treat content quality as a ranking input, not a nicety. Bing's three questions, can we trust it, is it comprehensive, is it well presented, are a clean checklist for any page audit.
- Measure beyond position. The paper names page length, pages per visit, mobile traffic, bounce rate and returning visits as signals worth tracking in Google Analytics, alongside Search Console query data.
What about AI search?
This is where a 2021 paper becomes surprisingly current. The signals it describes, trust, comprehensiveness, topical relevance, are precisely what generative engines draw on when they decide which sources to cite. In my opinion, the teams treating AI search as a brand new discipline are missing that the underlying quality bar has not moved, it has just become more exposed. If you want to see how that plays out on the commercial side, the way Google rewards well-structured product data is the same principle applied to feeds.
The concrete action is unglamorous but reliable. Audit your top pages against Bing's three questions and Google's emphasis on meaning, fix the weak ones, and stop chasing exact-match keywords that no modern engine, search or AI, actually needs.
Tags