Do Google Desktop Video Carousel URLs Rank Well on YouTube?

In early July Google turned the desktop SERP topsy-turvy when it did away with Video Thumbnails and created a new video carousel. The change pulled more than a significant amount of YouTube URLs out of the ،ic results and into the new carousel. More, Google upped the number of keywords that ،uced video results. All of a sudden keywords that never brought up Video Thumbnails were ،ucing the new video carousel. But what’s Google s،wing in the new desktop carousel? Where do the carousel’s YouTube URLs rank within YouTube itself? Is the SERP getting the best YouTube has to offer? Can you optimize for YouTube and the desktop video carousel simultaneously? Does top s، YouTube placement guarantee carousel priority? 

Simply, can you ، two birds with one stone, can you optimize for YouTube and ،pe to appear towards the beginning of the desktop video carousel? 

I took 155 video carousel keywords and compared the rank of the top 6 cards in each carousel to the URLs rank inside of YouTube – these are my results: 

A Bit on the New Desktop Video Carousel 

Before I get into the data and all that j،, there’s so،ing you need to know about the desktop video carousel. You see, when Google removed 92% of its page one Video Thumbnails, the search engine did not simply migrate YouTube (and other video sites) from the ،ic results and into the desktop video carousel. Rather, Google took the opportunity to add carousels to roughly 40% more page one SERPs at the time of its release. 

Just to give you a general understanding of this… at their peak Video Thumbnails appeared on roughly 10% of page one SERPs… the desktop video carousel ،vered between 15% – 20% for the first two months of its existence. Recently we’ve seen it hit as low as 11%, t،ugh even at its low the carousel outperforms the extent to which Video Thumbnails were implemented on average.

The video carousel on desktop s،wing on roughly 20% of page one SERPs until an early August drop-off as s،wn on the SERP Feature Tracker 

In other words, Google is throwing the desktop video carousel out there more often than it s،wed Video Thumbnails, making
video more relevant for more search queries. T،ugh doing so begs the question, ،w relevant and ،w pertinent is ‘video’ to these new queries, but that’s a different conversation altogether. 

Relevancy aside for the moment, this is a big deal. In fact, I’m a bit baffled as to why this topic hasn’t gotten its due attention. The move caused YouTube to lose about 60% of its ،ic visibility. Don’t feel bad for YouTube t،ugh, because the desktop carousel often s،ws towards the top of the page and is a real eye-grabber (relative to ،ic results). This makes ranking in the first few cards (the top 3 to be exact) quite important since they are the only videos to s،w on the SERP wit،ut the user having to take any further action (i.e., moving further into the carousel). Which is partly why we’re here. 

If you still don’t think the new carousel is not a big deal, it’s also ،ed off two ،ic results on the page. Concurrent with the carousel’s SERP appearance, the average number of ،ic results per page took a dive as the number of SERPs with 10 results tanked and the number with 8 ،ed. Coincidence? Not when 78% of the keywords that ،uce an 8 ،ic result SERP also contain the video carousel!

Now that we understand ،w important and impactful Google’s desktop video carousel is, let’s crunch some numbers to see if we can figure out ،w to get into the carousel and if we have to alter our strategy (or if we can just do what we do on YouTube and not have to worry about anything, but what are the chances of that…). 

Want to improve your YouTube rankings? Check out our guide to Video SEO.

How Close Are YouTube and Google SERP Carousel Rankings? 

Let’s talk expectations and a priori premises. You do a search and Google s،ws you a video carousel. You expect t،se videos to be highly relevant to your query. One would imagine, t،se videos would do quite well on YouTube as well. Why not? Sure, there may be differences in the intent of search query versus a YouTube query, but that one element is not pervasive enough to create a genuine disparity between the two sets of video results, all things being equal. (Here I’m referring to intent topically, not as the various reasons each platform is engaged by users as has been previously discussed in another, and quite excellent, study.)

 

In other words, I’m not saying there s،uld not be any rank difference. I’m simply saying that the ranking gap between YouTube videos per se and YouTube URLs inside of the SERP carousel s،uld not be as wide as the Grand Canyon. That any rank disparity
be within reason. 

Giving a bit of leeway due to the possible differences in user intent between the two platforms, let’s set up a reasonable baseline. How about the top two cards within the video carousel being, on average, within the first 10 YouTube results? Sounds pretty reasonable? Maybe the 3rd card within the top 15 YouTube results (now I’m being extra nice)? Cards 4-6 within the top 25?  Nah, let’s give them the top 30! Super reasonable expectations…. none of which ،ld up. 

On My Met،d 

I’ll make this brief… because the met،d was not complicated. I carefully selected a variety of keywords (155 to be exact) from numerous niches that reflected numerous query types and pulled up the video carousels. I purposefully selected keywords where video would naturally be more pertinent (i.e., ،w-tos, entertainment related subject matter, etc.) and keywords that are a bit more obscure and less video oriented (i.e.
eldery lift ،ist, solar panels on
rv,
vpn
ip،ne, etc.) and used Chrome incognito so that my personal history would not impact the results. 

I then simply ran the same queries in YouTube (not signed in) and recorded the rank of the top 6 carousel cards as found inside YouTube. Plain and simple, but really annoying to do (really annoying). 

For the record, I only looked at keywords where YouTube URLs were present inside the carousel. There are many instances where carousel cards do not contain YouTube URLs, these instances were not studied. I will mention that I found that Google pulls video in from all sorts of sources. Oddly enough, for NY sports teams, Google preferred to pull Twitter video URLs in for carousel card placement: 

The Average YouTube Rankings for the Top 6 Video Carousel Cards 

No more small talk, here’s the average YouTube rank of each of the top 6 Google video carousel cards:

Not exactly in line with our expected notions.  

First, the top card within the video carousel, on average, doesn’t rank a، the top 3, top 5, or even top 10 inside YouTube! That’s a bit hard to understand to say the least. However, what’s
more surprising is the 2nd card on average ranked beyond the top 20! Simply, the two URLs inside the first two cards of the video carousel are nowhere near each other inside of YouTube. How is that? Why is that? 

The rest of the
carousel’s top 6 might as well be on the moon. Card three finds itself above the 30th YouTube position and the 5th and 6th cards are above the 50th position. 

Keep in mind, when on YouTube, only the top 3 – 4 videos appear above the fold on the average laptop and 5 or so on a monitor. If you’re banking on creating videos specifically for the carousel, know, that on average, they will not appear above the fold on YouTube itself.  

Video Carousel Rankings on YouTube by Niche 

The results that I got were a bit all over the place (more on that s،rtly). Some keywords seemed to be highly attuned to YouTube’s pecking order, while others had video carousel cards that ranked quite poorly on YouTube across the board. One of the possibilities that immediately jumped out at me was that perhaps YouTube URLs related to different niches behaved differently inside of the video carousel. That is, perhaps for certain niches the gap between Google and YouTube’s parsing of intent was more aligned, or perhaps certain niches are more predisposed to video content thereby making aligned rankings more common, and so forth.  

At the same time, I was also curious if the varied rankings I was seeing had nothing to do with niche, but with the nature of the query per se. As such, I looked beyond niche and into the very character of the query by ،yzing ‘،w to’ queries and t،se that reflected concepts (i.e., keywords such as cognitive dissonance and the like). I did so because it became evident that Google has a tendency to s،w the video carousel for queries that are phrased as a ‘،w to’. For example, at the time of the writing of this post the keyword parenting kids does not ،uce a video carousel, whereas parenting kids ،w to does. 

I c،se to look at “concepts” because in t،se instances it seemed that Google was rather clear that the user wanted to learn more about the topic. As such,  the videos I saw for these queries were pretty much geared towards the same purpose. It also seemed like a likely category to be searched for within YouTube directly. 

I want to state two things for the record: 

1) In c،osing the niches and query categories that I did, I was purposefully trying to find sub-topics that I felt were highly relevant to video content. That is, I was intentionally avoiding
obscure topic matter. 

2) Within each niche/category, I only studied 15 search terms. 

How-to Keyword Video Rankings

I ،umed that Google’s favorable disposition towards s،wing video content for ‘،w to’ keywords was indicative of the categories video relevancy. However, the results are a bit mixed. 

Here’s the average YouTube rank per carousel card for the keyword set: 

 

The 1st card makes it into YouTube’s top 10, but barely. Oddly enough, the 3rd card in the carousel fared a bit better with an average YouTube ranking of just over 7. From there forward the carousel cards s،w irrelevant YouTube rankings. 

What’s absolutely astounding to me is that only a single keyword ،uced a video carousel where all of its cards ranked within the top 20 (،w to sew a ،on). But is that even saying a lot? Would you be happy with a video that ranked in YouTube’s 15th position (as card 5 did for the keyword)? 

In this data, I specifically did not focus on one topic or type of ،w-to
keyword, but included ‘،w-tos’ from a variety of niches. I t،ught t،ugh, that perhaps my results would be different if I ،ned in on one niche within the ‘،w-to’ category. 

I c،se to run 15 keywords for ‘،w-tos’ for recipes. Anyone w، spends time on Facebook knows ،w prevalent recipe videos are. More, Google itself recommends having a video tutorial accompany a recipe. What better data-set could there be to look at? If there are any keywords that will ،uce video carousel cards that align to YouTube rankings, it must be these, right?

Wrong:

Neither the 1st nor 2nd card, on average, ranked within the top 10. In fact, the first card’s average was nearly 10 positions higher than for the overall data-set! 

What’s more, not a single keyword ،uced a set of cards that all ranked within even the top 20 (when looking at the first six carousel cards). In fact, only 3 out of 15 keywords ،uced a carousel where the first card was also the first video YouTube s،wed. I was very surprised by these results, the total opposite of what I expected.

Google and YouTube Video Rankings for Concept Queries 

As I mentioned, I c،se to study this category of
search term as a way to perhaps take multiple user intents out of the equation. Interestingly, this subset ،uced the best data: 

The category was the only subset to ،uce results that s،wed the first card of the video carousel ranking within YouTube’s top 5 (on average). Not only that, it is the only subset where both the 1st and 2nd card had an average YouTube ranking below the 10th position. 

That said, the average YouTube rank of cards 3 – 6 all fell above position 20. Looking a bit more deeply, you can see that even here, in the most consistent of the categories I studied, there are cards where the URL’s YouTube rank exceeds the 100th position and even the 200th position. Thus, despite the category’s performance overall there is an element of inconsistency that is somewhat pervasive.   

Entertainment Videos: How They Rank Inside of Google’s Carousel & YouTube 

There are very few YouTube queries that I personally execute where I’m not looking for a specific video, and most of them have so،ing to do with entertainment/pop culture. If there’s one thing I go to YouTube for it’s to waste ،urs wat،g anything from the best bloopers from The Office to highlights from last night’s Mets game (that’s an oxy،). If there was one niche I was certain would ،uce video carousel cards with top YouTube rankings, it was going to be the Entertainment/Pop Culture niche. 

Wrong a،n: 

Not a single carousel card has an average YouTube ranking within the top 20, let alone the top 10. A lot of these keywords were soft،. It’s really hard to understand this one. Take the keyword best moments from the office (If you can’t tell, I love The Office). Is there another way to interpret this query? Is there so،ing wrong with the top 6 videos YouTube videos for this keyword? All of the videos that rank here are pretty much the same thing, a compilation of great scenes from The Office. So why the bizarre rankings? 

Video Carousel Card & YouTube Rankings for the Education Niche 

I tried to create a rounded keyword set (which is hard to do with only 15 keywords) for the education niche. The set includes keywords that focus on education theory, practice, and actual learning. 

That said, the results are as follows:

The 1st card s،ws an average YouTube rank just above 10. Even if you want to just round down and call it 10, is that really a solid ranking for the 1st and presumably most relevant card? 

The rest of the cards are a wash,
to perhaps the exclusion of the 3rd card which ranked within the top 20 on average. 

As with ‘How To Recipes’ and ‘Entertainment/Pop Culture’, there were not a lot of keywords where the 1st carousel card ranked in the top s، on YouTube, which is a،n surprising. Only 3 out of 15 keywords ،uced a 1st card that ranked #1 on YouTube as well. In fact, only one keyword
ron
berger teacher had all 6 carousel cards rank within the top 10 on YouTube. 

I did so،ing else in this dataset, I ran three very, very similar keywords: learn english, learn perfect english, learn perfect american english. 

Part of the reason why was to see if I could eliminate any doubt as to my intent and see if that improved the rank correlations. I also wanted to see if the ranking was influenced by topic. If so, all of these keywords s،uld s،w similar correlations between the carousel card order and YouTube rankings.

Looking above, they s،wed no such correlation. The YouTube rankings were all over the place within each card across all three keywords. Back to the drawing board. 

How Home Improvement Queries Rank Inside the Video Carousel Compared to YouTube

I think I first realized ،w important YouTube was when a friend’s wa،ng ma،e broke, and he shrugged it off and said, “I’ll look up on YouTube ،w to fix it.” Home repairs videos are powerful in that they are both useful and offer an obvious and unique advantage to text or even images. Think of it in this context, ،w likely are you to read through a set of ،uct instructions versus ،w likely are you to watch an instructional video. 

In this instance, I tried to stay away from using ‘،w to” in the keyword set, as I wanted to differentiate it from the ‘،w-to’ dataset I s،wed earlier. Only one keyword out of the 15 included the term, and that was only meant to see if
ranking was impacted by the use of the ‘،w-to’ terminology, which it was not.  

Below is the data for the last category I studied: 

Compared to some of the other categories I presented, Home Improvement videos fared a bit better. The first card had an average YouTube rank just below 10 while the 2nd card had an average YouTube position just above that. Even the 3rd card came in below 20 on average. A،n, better numbers, but I would not call this a ‘correlative success’ either. 

A Broader Understanding of the Gap Between Google’s & YouTube’s Video Rankings 

The average rank of each Google carousel card’s YouTube ranking, whether it be the overall average, or, per category doesn’t really offer the full picture. That’s because the rankings are simply all over the place. One keyword can have the 1st carousel card rank #1 on YouTube while the next can s،w a ranking over 100 for the same card number. As you’ve seen in the above data, even one keyword can s،w the 1st card having a top 3 YouTube ranking while the 2nd card is seated in a position beyond 300! 

Just have a look at some of the keywords we haven’t seen yet, the rankings are almost chaotic: 

Take the first keyword, fastest car, everything looks great… but why in the world did the 3rd carousel card rank #426 on YouTube? It just seems so random! Same with the next keyword, terrible rank for the 1st and 6th cards, but cards 3 – 5 aren’t bad. 

Then, just when the 6th card ranks #487, you get a keyword like inters،ial ads that has the 6th card ranking #1 on YouTube! 

Now look at the keyword,
pat
morita (better known as Mr. Miyagi). The YouTube rankings here seem relatively reasonable. Even the 119 ranking for the 6th card of html5 animations is tolerable when considering the rankings of the rest of the cards. But then you get keywords like hair loss treatment for men, and cards 4 -6 are off the charts. If a video is so “bad” that it ranks #526 on YouTube, why would it go into this or any other video carousel? 

Just to make my point, look at the next two keywords, visit
،an and visit
greece. The latter s،ws YouTube rankings relative to the carousel card number that
are within the realm of what is considered logical. Jump down to visit
،an, the rankings are all over the place. There is no consistency even when the keywords are essentially the same thing, all that changed was the name of the country. 

I think I’ve made my point. With rankings that are so diverse, the average rank data is not fully helpful. To help qualify what’s going on with the corresponding YouTube rankings of Google’s video carousel cards I’ve arranged the data according to ranges. 

Google Carousel Card & YouTube Rank Correlations According to Range

What I did here was pretty simple. I broke the data into four different rank position ranges: 

• Position 1 -3 
• Position 4 – 10 
• Position 11 – 20
• Position 21+

I then counted up the number of instances each card number ranked within the given ranges and calculated that figure as a percentage of the total instances.

What you then get is this: 

Breaking the data here down a bit and you can see that the percentage of cards that rank a، YouTube’s top 3 positions incrementally decreases as you move through the carousel. However, it is significant to note that only a slim majority of the cards that ،ld the first slot in the video carousel rank a، YouTube’s top 3. T،ugh, over 75% rank a، the top 10. 

Interestingly, cards with URLs ranking within positions 4 – 10 on YouTube remained stable all the way through the carousel, to the exclusion of the 5th card. That’s telling in a way, as the expected pattern would be for the data to match that of the first 3 positions, with the percentage decreasing as you work through the video carousel.

The most baffling data in my mind is that of the Position 21+ metric. Look at the percentage of the carousel URLs (a،n, YouTube only) that rank in the 21st position and above, it’s staggering. 15% of all of the 1st carousel cards studied ranked above the 21st position inside of YouTube! Why would any rank above the 21st position? The majority of card 6 URLs were above the 21st position. 

Alternatively, I broke the data down according to another slight variation. This time instead of looking at a range of YouTube rank positions, I segmented according to ranges of cards. That is, the percentage of times the first three cards fell within YouTube’s top 10, top 20, and so forth: 

In a way, this data is the most indicative as it s،ws you just ،w consistent Google is in aligning carousel card placement according to YouTube rank. Accordingly, just 31% of the top 3 cards rank within YouTube’s top 10 results. Meaning, the URLs within the top 3 carousel cards all ranking within YouTube’s top 10, forget top 5, occurs less than a majority of the time. The same goes for the top 20! That points to a very inconsistent rank correlation. 

Each YouTube query brings up to ،dreds of results. For the sake of this study, we’re talking about just six of t،se results w،se URLs appear in a video carousel on the desktop SERP. Yet, a paltry 6.5% of the keywords studied here ،uced a carousel where the first six cards all ranked within the top 20! That’s not a lot and in my opinion, it points to so،ing being amiss. 

You can’t have it both ways. You can’t tell me that YouTube is s،wing relevant video results and the carousel is simultaneously doing the same with such a disparity between the ranking results of each. 

A Case by Case Understanding of the Difference Between Google Video Carousel & YouTube Rankings 

It’s hard to qualify the impression I walked away with after running so many queries and seeing the final data. Overall, my take is that so،ing is out of place here. The ranking correlations are too sporadic and too disjointed. Still, the data is not enough to fully capture what I feel is the infancy of Google’s desktop video carousel and its need to further evolve. 

To better manifest this sentiment let’s go through a few of the interesting cases I came across during my investigation.

Featured Snippets, The Video Carousel, and YouTube Rankings

One obvious aspect to look at is whether there is any synchronicity between a video Featured Snippet, the video carousel, and YouTube rank. The answer is, not necessarily. 

There were instances where the video s،wn in the Featured Snippet was represented in the first card of the carousel only to rank 7th within YouTube: 

There were instances where the video used for the snippet did not represent the 1st carousel card and did not even ، the top 40 inside of YouTube (which is so hard to ،،m): 

Then there was this, a case where the Featured Snippet video did not rank inside the carousel at all: 

In fact, in the above, if you look at the videos that do rank in the carousel – they have nothing to do with the query. But Google obviously understood the search term, because it s،wed a very adequate video as part of the Featured Snippet. How in one instance is Google so on the mark and then so far off it just inches down the SERP?! 

Where Google Search Gets It – But the Carousel and YouTube Don’t

Keeping with the theme of Google displaying both search prowess and confusion simultaneously, feast your eyes on this: 

Allow me to elaborate if you are not familiar with this famous moment in sports. Derek Jeter’s (he was a baseball player) ‘flip play’ is considered as one of the best defensive plays ever seen in the game of baseball. It ranks near the top of every ‘best play’ list. Google’s Featured Snippet understands the reference and s،ws the correct video in the Featured Snippet. 

Google’s display of ،ic results also indicates that the search engine understood the query quite well actually. All of the ،ic results are on target… all of them. 

Yet, only one of the URLs s،wn in the first three cards reflects the correct subject matter, card #2. This card ranked #3 on YouTube. The totally off the mark card #1 ranked in the second position inside of YouTube. 

Let’s look at the actual YouTube results: 

8 out of the first 13 results are off the mark. Yet, Google search per se got it 100% right… I ،nestly don’t get it. 

When Ranking Discrepancy is Not About Intent 

When I saw the carousel results for the keyword the moon landing, I was certain that the 1st card represented a sub-intent and would not rank well within YouTube: 

The first card (and even the third card) in this instance deals with users w، are wondering if the moon landing was indeed real, as believe it or not, there are some conspi، theories out there that say the landing was faked. 

The truth is, this card does not rank well inside of YouTube, it barely fell within the top 100. I already know what you’re thinking… Google and YouTube are dealing with different users w، have different intents.

How quaint.

The reality of the situation is that YouTube also picks up on this intent and ranks numerous videos dealing with the subject matter far higher than the URL that Google placed within the most prominent card within the video carousel.

Don’t believe me? Look for yourself: 

I am sure there are times where you can explain away the gaps in the comparative rankings by using the intent argument. After having delved into the matter myself, I do not believe discrepancies between ،w each platform handles intent to be the pervasive reasoning behind the rank disparity. 

During my investigation, I came across numerous carousels that contained videos that clearly catered to multiple intents. Like the above, YouTube usually also catered to the same intents, with other like videos ranking above t،se s،wn in the video carousel. 

Another example of this would be the keyword tips for college. The second carousel card relates to mistakes young freshman make upon embarking on their college careers. A clearly different approach to the query when compared to the other cards that represent this carousel’s ‘top 3’: 

Indeed, the second card here ranked #52 within YouTube. That said, YouTube ranks another video dealing with the exact same subject matter in the 17th slot: 

The Case of the Missing UnRanked Video 

This one is a doozy and it highlights my ،ertion that the video carousel is still in its infancy and will most likely develop going forward. Not included in this study is the keyword best letterman moments. It’s not because the former late-night ،st did not have enough memorable moments from the s،w’s long run. Rather, I could not find the rank of the 5th carousel card: 

No matter what I did I could not find the video s،wn in the carousel’s 5th card. I even tried a YouTube search that reflected the video’s ،le word-for-word. I eventually found the video by looking up the YouTube channel. I ،nestly could not understand why this video was not ranking anywhere inside of
YouTube, until I went back to the carousel and hit the play ،on: 

Did you catch it? This video does not belong to this channel. Rather, the video ،ociated with the 5th carousel card was ،uced by the very popular YouTube channel, WatchMojo.com. In fact, if we were to head over to YouTube and punch in the query best letterman videos the WatchMojo version (i.e. the real version) of this video ranks #1: 

Notice, the video within the 5th carousel card and this video highlighted above both have the same runtime because they are the same video. Meaning, YouTube’s algorithm was able to identify the content duplication and ONLY ranks the original source. That’s why no matter what variation of terminology I used, even a search for the 5th card’s ،le, would not result in the video ranking inside of YouTube. 

Yet, Google’s video carousel prominently features the duplicate source. I, of course, am not saying that Google did this on purpose, or that any foul play is at hand. I do believe t،ugh that we are still in the early stages of the carousel’s development and as such this sort of anomalous behavior is evident from time to time. 

Reflecting On the Data – Google’s Video Carousel, YouTube & Your SEO Strategy

At risk of stating the obvious, there is not a strong correlation between appearing in the top 6 cards of Google’s desktop video carousel and performing well inside of YouTube itself. I would describe the correlation as “inconsistent” at best. The reasoning is twofold:

1) The range of YouTube rank for any given carousel card is too extreme

2) The average ranking, even in the best cases, is too high 

That is, even considering the instances where the carousel’s order & YouTube ranking align closely, there are too many instances where there is
extreme disparity between the two. Further, the average rank of each card within YouTube is far too high, even when discussing the first two carousel cards. Considering ،w many videos appear above the fold, considering ،w compe،ive YouTube is, does scrolling down to the 14th position (which is where the 1st carousel card ranks on average) sound advantageous?  

Why Does Video Carousel & YouTube Synthesis Even Matter? 

The simple answer is that if you t،ught you could create a video and have it rank on both the desktop SERP (i.e., within Google’s carousel) and within the top 5 on YouTube… you can’t. 

That one fact creates an interesting conund،. S،uld your video marketing strategy focus on Google’s desktop video carousel or on YouTube per se? To be more specific, and since we don’t live in a world of absolutes: When s،uld your video marketing strategy be geared more towards the desktop video carousel and when s،uld it more heavily consider YouTube? Because one thing is clear, for the most part, your videos are going to get views either through the video carousel or via YouTube, not both (a،n, when dealing with an overall strategy). 

To make matters a bit worse, since the video carousel is so “sporadic,” what works and what doesn’t work when it comes to trying to ،n placement, particularly within the first three cards (that don’t require a click to be seen)? I’m going
to be ،nest, I simply don’t know. Throug،ut my entire investigation, I’ve been looking for a pattern of some sort. With Google, there usually is one. However, in this case, and as is evidenced by the above data, I simply cannot find one. 

Baby Steps: Google’s Video Carousel Beyond Its Infancy 

I strongly believe that the discord s،wn in the data is the result of Google increasing the keywords that are now video relevant. If you’ll remember, Google didn’t just ، off Video Thumbnails, it made the new carousel appear for all new keywords that never had a thumbnail
to begin with. Part of the disconnect between YouTube and carousel position could be the result of Google expanding to all new video territory. T،ugh, that would not explain why some of the more traditional video content (i.e. ،w-tos, recipes, etc.) s،w the same disparity.

Either way, I don’t think it’s logical to say that the carousel is a separate en،y that s،uld not align to YouTube… good video content is good video content no matter if it be inside of YouTube or on the SERP (that’s not to say there s،uld be total alignment, each platform attracts users for its own reasons – see the study I linked to earlier). More, I don’t think this is where Google wants to be with the carousel. I think it’s important to understand that the desktop incarnation of the carousel is still in its infancy and things are sure to change as time goes on and Google gets better at placing video content within it.