Video SEO in 2024: the SERP in the Netherlands

Our in-house video analysis

New opportunities for video SEO? Our data specialist Justen didn't need to be told 2 times. He developed a new script for our in-house tooling in no time, enabling us to collect and analyse all data relating to SERP videos. This allows us, for example, to gain insight at a glance into:

• Which search terms displayed videos
• How many videos were displayed per search term during the period
• How often a particular video was displayed
• How stable the video results were
• The titles under which the videos appeared in the search results
• Which channel the videos came from
• The URL of the videos displayed
• For YouTube videos: what the autocomplete suggestions were for each search term

We highlight the key insights from our Dutch case for you.

Case: fashion brand keyword tracking

Our case consisted of tracking of 733 fashion-brand-related search terms, from pure branded-terms (‘stone island’) through to combinations with products (‘stone island jacket’). Fashion, a visual segment, in which there might just be lots of videos could appear in the SERP, or so the reasoning went. That was a little disappointing. 

Frequency: how often was a video displayed?

The analysis examined the search results from the past 30 days. Of the 733 keywords, only 53 on which a video was displayed at some point during that period, 7% of the total.

Emphasis on ‘at some point’: for only 1 term (‘ugg platform chestnut’) was in 100% of the time, a video was visible, but the average was at 56%. In other words: of the 30 days, a video appeared on on average 17 days, a video result for a keyword; on the remaining days, there were no video snippets. 

Localisation: which language were the videos in?

In total, 53 keywords, then, for which we encountered video snippets. Although the tracking was predominantly Dutch-language, with one exception, the keywords that ultimately remained were predominantly ambiguous. Think of Dutchified but actually English combinations with ‘blazer’, ‘hoodie’, ‘cargo pants’ and pure brand names.

We saw the same ratio reflected in the video results themselves. 77% of the videos was in the English, the remainder 23% consisted of a mix of different languages, from (a little) Dutch to even Japanese. Perhaps this is related to the ambiguous nature of the 53 keywords, or perhaps video snippets in the Netherlands are (still) not so region-specific.

Platforms: video distribution by channel

Another thing that stood out was the distribution across the channels from which the videos originated. Our hypothesis was that, in particular, the short videos offering a ‘snapshot of the atmosphere’ would perform well for the fashion industry. Think Instagram Reels, TikToks or YouTube Shorts. In reality, the opposite was true. Around 88% of the videos came from YouTube, just under 10% came from TikTok and the rest came from Facebook and Instagram (the latter accounting for a mere 1 in total).

The videos shown were also 88% of the time, normal (horizontal) videos, usually between 5 and 10 minutes long in this case, rather than the more fleeting YouTube Shorts we had expected.

Relevance: did the videos actually match the search terms?

The 53 terms from the fashion brand tracking that had video snippets in the SERP together yielded a total of 210 video results in total. Only 44% of these were not noise.

In our keyword tracking, which contains many brand names, there were various terms with more than one meaning or search intent. For example, the fashion brand ‘Caruso’ produced many (Italian) music videos, while fashion brand ‘Begg & Co’ was interpreted by Google as a reference to a business podcast by a certain Mike Begg, not what we were looking for.

All in all, there were 93 video snippets left that were actually relevant to the fashion industry. These were mainly:

• Unboxing videos
• Reviews
• Try-ons
• Collection overviews (including fashion shows)

In addition to brand names, terms such as ‘review’, ‘unboxing’, ‘collection’, ‘new arrivals’, ‘try-on’, seasons (‘Fall/Winter’, ‘F/W’) and years frequently appeared in the titles, as did, unsurprisingly, ‘fashion’ and ‘outfit’.

Quality: how did the platforms distinguish themselves?

Finally, we looked at the possible signals Google uses to determine which of the thousands of available videos may or may not rank in the SERP for relevant keywords.

Are these, for example, clips from popular channels with a large number of subscribers? Are they videos with a large number of views and/or likes? Videos with a high level of interaction in the form of comments? When assessing regular organic search results, Google takes external validation into account, for example in the form of domain authority or reviews. We therefore thought it logical that this would also apply to video snippets.

The answer, however, was no, no and no. Interestingly, the videos we encountered were predominantly from (very) small and unknown channels without any notable subscribers, and rarely had many views or any likes or comments.

The videos were, however, 9 of the 10 times relatively recently posted Most were no older than 6months. In our case, this may be due to the fast-moving nature of the fashion industry: after all, there is already a new collection six months later.

The good news: you do not need to build an entire YouTube career before you can rank with your video content in the SERP. The bad news: you have limited opportunities to influence your chances of achieving a position in the SERP.