Impact analysis of Google's 2020 December Core Update

Google's December 2020 Core Algorithm Update was a big one! At first, I didn't want to dig. But then, curiosity pulled me in. It wasn't disappointed. If I had to make a definite statement, I'd say that the algorithm looked predominantly at the desktop User Experience. But, as always, there is much more going on.

I'll start with summarizing my observations and finish with data of the 123 most exciting domains (some of the world's largest sites). All data is from SEMrush and covers US mobile and desktop. I manually checked performance in other countries here and there as well but not systematically for all domains.

Let's dive in.

The roll-out

On December 3rd, Google officially confirmed the release of a "Core Algorithm Update". You know how it goes.

https://twitter.com/searchliaison/status/1339279864742830081

As I always say, it's worth waiting until the update fully rolled out instead of firing early shots. This one took two weeks!

I found that the update rolled out globally at the same time. However, the impact varied across countries and devices. A detail that's often forgotten.

For example, Linkedin equally gained in the US, France, and many other countries (see below).

Impact

In some cases, the December core update was a reversal of the May update, in other cases, a continuation. A lot of domains I cover in this analysis already came up in my May Core Algorithm Update analysis. Most reversals consisted of a domain losing traffic in the May Core Update and recovering in December, even though I've also noticed some domains that recovered much earlier.

To give you a taste of domains that lost traffic in May and recovered in December:

• nfl.com
• creditkarma.com
• nerdwallet.com
• pexels.com
• verywellhealth.com

Some sites that already lost traffic in May lost even more in December. For example:

• drugs.com
• fandango.com
• 9gag.com
• howstuffworks.com

The update was a heavy one. On desktop, 33 out of 123 sites I took a closer look at had a +10% increase, 11 had a +20% increase, and 29 sites dropped beyond -10%.

The biggest relative drop happened on asos.com on desktop: -39.63%. On mobile, it hit Fandango the hardest: -19.29%.

The biggest relative gain came to gossipcop.com on mobile (154%) and desktop (+226%). I'll explain why in a minute. Verywellhealth.com came in second place with +130% on desktop and +57% on mobile.

In absolute numbers, Wikipedia lost the most traffic: almost 100M on desktop, which makes up a loss of "only" -2.79%. It's not the only dictionary site that lost a lot of desktop traffic: dictionary.com and wiktionary.com lost over 30M and 19M visits, respectively. So, I guess you could say that dictionary sites are having a hard time.

IMDB.com gained the most traffic with +84M on desktop and +32M on mobile. Instagram came in second for absolute desktop gains with +51M and lowes.com on mobile with +17M.

Mobile vs. Desktop impact

As I already alluded to, the impact was different on mobile and desktop. Most sites were impacted on desktop, but there are exceptions like cars.com, which saw bigger gains on mobile (+18.75%) than desktop (+10.55%). Some sites lost on mobile and won on desktop like Doordash or the other way round like Apple.

When we map the relative impact on mobile and desktop traffic on a scatter chart, you can see that the majority of domains saw a similar impact on mobile and desktop. Just a few venture out.

A mobile subdomain bug - or a hint?

In my research, I noticed a weird bug with mobile subdomains (example: m.domain.com) showing up in the desktop index. Facebook's mobile subdomain, m.facebook.com, gained a lot of traffic on desktop and saw a huge spike on December 3rd in many international markets. Remember, this is the mobile subdomain in the desktop index.

The weird thing is that this doesn't happen in every market. I noticed it in BR, FR, ES, IT but not DE or IN.

The same "bug" occurred for

• m.kohls.com in the US and CA
• m.aliexpress.com in the US
• m.fourseasons.com US, which tanked in May
• m.alibaba.comm.barnesandnoble.com

Look at the Google Search Console screenshot below from a big domain (trust me, it's huge). It's not just 3rd party rank trackers that caught this.

What does it mean? It's probably a bug that results from a major recrawl of the web. Google recrawls large parts of the web to reassess relevance and authority for core algorithm updates (see for yourself in the crawl stats report in GSC: crawl rate spiked in early December and most recrawls had the purpose of refreshing Google's index instead of discovering new URLs). It seems that some mobile subdomains accidentally made it into Google's desktop index. Could this be a hint that subdomains, in general, were reassessed? I think it's likely.

One theory I have is that Google reassessed how it differentiates between the relevance of subdomains and root domains. How that surfaces in the search results, is yet to be seen.

User Experience becomes more important

One commonality between winners and losers I observed is user experience. What I mean with UX is not Core Web Vitals, even though that's how Google quantifies user experience with an impact on rank coming next year.

Instead, I look at user experience in this context as design, layout, ad load, and how easy it is for users to get their needs met. I don't know exactly how Google measures this, but there seems to one or several factors that make an impact.

Let's look at some examples!

Answers.com is one of the winners of the updates while ehow.com lost. Ehow's UX is subpar: the ad load is obnoxious and the answer you're looking for is really hard to find (see below).

Answers.com, on the other hand, has a very clean design with minimal ad load. Notice how the design looks like a mobile site with a very narrow content area.

When we look at a third player in that vertical, howstuffworks.com, we can see another example of an ad load that's detrimental to the user experience.

One example: About 50% of the content on https://recipes.howstuffworks.com/wagyu.htm is a Taboola Feed (ads). Taboola distorts the mobile experience as the ads are too wide for the screen (see below).

Another example of a harmful user experience is wish.com. The site shows a popup to anyone visiting that you can't get around. In plain terms: you can't even get to the content. As a result, wish.com tanked -24.63% on desktop and -10.33% on mobile traffic.

A different case is Gossipcop, one of the big winners of the list of domains I analyzed. The domain went through a complete redesign since it got hammered by the December 2019 Core update.

Notice how the new design is much cleaner and has a clearer separation of ads from the main content.

Speaking of outdated designs, another big loser of the update is fandango.com, which also has an ancient design.

But user experience isn't just how a site looks. It's also how easy it is for users to get their needs met (lingo from the Google quality rater guidelines), and that's where UX and content overlap.

The differences are often very subtle, which speaks for Google's ability to detect them. An example is Wiktionary (lost) vs. Merriam Webster (gained).

Wiktionary dropped in ranking for the keyword "defence" from #1 to #4. Merriam Webster jumped up to #2. The reason? Wiktionary has no definition of "defence". Merriam Webster does.

Penalties overlap with the algorithm update

The timing of manual or algorithmic penalties often coincides with bigger core algorithm updates.

I'm not sure if Asos was penalized, but they had issues with doorway pages and PBNs in the past. It seems that sites like petlifetoday.com or www.thearchitectsguide.com have been penalized in December while the update rolled out. It's hard to tell from the outside, but the traffic trajectory hints at it.

Conclusion

We're not living in times of linear correlations anymore. There is no single factor that determines whether a site wins or loses in a core update.

In the December core algo update, I observed that authority and UX seem to be a common theme between winners and losers. My story is that Google shows more authoritative brands at the top of the search results and tries to push domains with poor UX further down.

Lastly, here's the data for you to explore:

Desktop winners and losers

Mobile winners and losers