Google Asking Searchers Which Result is Better

Google Asking Searchers Which Result is Better

It’s no secret that Google has been using human quality raters to fine tune its algorithms. It’s also true that Google search results pages occasionally asks the searcher what they think of the results page overall. But Google has recently taken this to the next level by asking users to specifically compare two search results and choose which one is best.


Now, for certain results pages, a form appears to the right of the search results that asks:

Which result do you prefer?

Visit both pages before choosing.

–                    [one result with a link to the page]

–                    [another result with a link to the page]

–                    Both results are equally good

–                    Neither result is good for my needs

The pages are chosen from the middle of the search results page; they are not the first and second results.

This survey is interesting because it is not only asking searchers what they think of the results page, but how two specific pages compare with each other. A “Learn more” link explains that the ratings won’t impact the search results directly, but will be used to test changes to the algorithm.

By introducing such side by side tests, it’s clear that Google is collecting more accurate statistical data, suggesting that more updates are on the way. How the algorithms will make use of this information isn’t entirely clear, but it strongly suggests that the search engine is looking for more obscure correlations and patterns that dig deeper than links and keywords.

If you thought Panda was Google’s one and only flirtation with machine learning, think again.

In reality, Google has a growing love affair with machine learning algorithms. For example, Google is currently selling a new prediction API to firms. It takes advantage of cloud-based machine learning tools which firms can use to develop their own spam detection, diagnostics, document and email classification, customer sentiment analysis, and more.

If Google is marketing machine learning algorithms, it’s hard to believe they aren’t planning to use them in future search upgrades. Research At Google has it’s own Artificial Intelligence and Machine Learning team, with 263 publications on the subject.

It’s interesting to speculate about what’s going to happen when Google combines the data collected from these user surveys with their machine learning capabilities. I suspect the impact on SEO won’t be small.

What are your thoughts?

Original Source:


Recent Posts

Start typing and press Enter to search