Recently Google CEO Sundar Pichai was asked to testify to the Congress about possible biases in Google's algorithms. This isn't the first time Google has been accused of bias, and it probably won't be the last. Google claims there is no bias, but many conservatives argue that Google is biased against them.
With our expert knowledge of search engine optimization (SEO) And the huge amount of data from Ahrefs, we wanted to see if we could spot a bias from Google by examining data for popular conservative and liberal news sites. If you are new to Ahrefs we are one of the best SEO Seriously Big Data Tools on the Web.
Google makes hundreds of improvements to its ranking algorithms every year. Most of these go unnoticed because they are small. However, from time to time there is a major "core" update that affects a large percentage of search results. When Google gave us the dates of these updates, we figured we could look for bias by examining the organic traffic to popular liberal and conservative news sites before and after these updates.
For example, here is the estimated organic search traffic to Fox News since 2015. Each line represents a Google Core update:
Looking at this data for a website doesn't say much, however, so we've done the same and for the best conservative and liberal news sites. We pulled these from AllSides Media Bias Ratings (left and right bias ratings). Here is a list of these websites:
Conservative news sites:
- New York Post
- The last refuge
- Drudge report
- The federalist
- Orange County Register
- The epoch times
- Washington Times
- Christian broadcasting network
- National review
- City Hall
- The Mark Levin Show
- The Rush Limbaugh Show
- The national interest
- The gateway expert
- Red state
- PJ media
- Washington Examiner
- Fox News
- Christian today
- Zero hedge
- The daily caller
- The daily line
Liberal news sites:
- US news & World report
- The Washington Post
- Hustle and bustle
- NBC news
- Hollywood Reporter
- Los Angeles times
- Yahoo news
- Al Jazeera
- Rolling Stone
- The edge
- The New York Times
- ABC news
- CBS Local
- The guard
- CBS news
- The Atlantic
Before we get into the results, I should share a bit about Ahref's data. We have hundreds of millions of search terms and huge amounts of clickstream data. We use this data to estimate organic traffic by looking at the various search queries, the positions websites occupy in search results, and the click locations of the users. For the core updates, we decided to examine the traffic at the start of the Google core updates and the traffic 14 days later. This will give Google time to roll out the changes across the various data centers. It also gives us time for our data to reflect the changes.
Our data is normalized in the sense that the volume is averaged over 12 months. So this should mainly take into account seasonality, with elections being an exception as they do not take place every year. We won't see newer stories or search topics early on, either, but we should get any popular searches and associated clickstream data later.
From 2015 to the present, there was a decrease in average traffic for the top news sites in each category during the Google Core Update periods.
Conservative overall traffic decline: -2.65%
Liberal overall traffic decline: -1.78%
These numbers are actually very similar and not statistically significant when you consider that we consider traffic from 50 websites and look at a 6 year period. Before the last election in 2016, the impact on both categories was about the same. If you look at the results for the previous year or so before the 2020 elections, you will find that the impact for both categories was roughly the same. The latest update seems to be better for conservative websites.
If we look at each of the data points, both conservative and liberal news sites have seen positive and negative effects on each of these Google Core updates. Each box diagram below shows the top websites in each category. Let me reiterate that every single update had winners and losers for both categories. Whether a site wins or loses on a core update tends to depend more on its quality than anything else.
While we cannot conclude from this data that there is no bias in Google search results, we can say that no new bias has been introduced on Google Core updates in the last 6 years.
Is there any traffic distortion?
One of the things that we noticed is that liberal websites are definitely getting more traffic than conservative websites.
Now the question arises why this is so. Does this show a potential bias in front of our keyword record? Let's see if we can explain the traffic difference.
Amount of content
In terms of the number of pages indexed, liberal news sources have more than 8x more indexed pages than conservative news sources. In fact, the graph for traffic participation is almost identical to the one above. As a result, the best liberal news sites generally have more chances of ranking for other things than the best conservative news sites.
Brand or brand traffic
Brand traffic for CNN and Fox News is roughly the same, meaning a similar number of people look for them specifically in organic search. Brand traffic makes up a smaller percentage of, however CNNOverall traffic, probably because they just have more content. CNN has ~ 2.5 times as many indexed pages as Fox News, giving them more chances of ranking on different things.
Other explanations for traffic differences
With just Google search data, it's difficult to definitively determine why liberal websites get more traffic than conservative websites. Conservatives may be using Google less. It's also possible that Conservatives will get more news from TV, Apps or social media than google. All of this may be true, but without additional data these statements cannot be confirmed and are only guesswork.
More ways we could have looked for bias
If we looked at specific examples of queries, we would surely have found examples of bias. The problem with using individual examples is that there is an inherent bias in the person doing the analysis. The data is complex and you need a good way to determine if the websites and content are relevant to specific queries. Unfortunately, this would be difficult and extremely subjective, which is why we have decided not to pursue this path any further.
We could also have looked at link data or links from other websites to these news sites. Again, I have no doubt that this data is biased as the more popular and higher ranking websites with more pages would tend to naturally get more links. We have already proven this in our backlink growth study.
We could have looked at specific examples of search terms that Google suggested through their autocomplete system, but we already know that they remove a lot of negative terms from these results. Again, this is already biased, and it would be difficult to prove a malicious bias here.
Everything is inherently biased. While our data shows that Google Core Updates did not appear to be biased in one way or another over the past few years, we still cannot say for sure that there is no bias elsewhere in the Google system.
Do you have any questions about this data? Ping me on Twitter.