Recently, our team have been undertaking some research into image search and image SERP trends, when we found an extremely interesting development in the SERPs. Highly likely an algorithmic change.
When we looked at the SERPs for hoodies, we saw that there are some websites that shouldn’t be there. When we say shouldn’t be there, we mean based on everything we know about the algorithm in the fashion industry and e-commerce for the past 12-15 years.
Here, we have found a number of brands who have low visibility absolutely dominate well-established brands who have invested millions of pounds into their SEO.
These sites invariably are built on Shopify, have little to no written text, are JavaScript prominent and usually have poor keyword optimisation such as no title tags, links are few and far between in most cases. Everything we have found suggests that these sites should not rank.
Not only that, but the only thing we can see that explains this uplift and prominence in the SERPs is social media.
Sisters and Seekers (S&S) are the main outlier here, and one we can use to prove the point. At the time of writing this, S&S has no backlinks to their site worth mentioning and very little in terms of SEO optimisation, yet are ranking on page one for ‘hoodies’.
Delving into S&S…
SERPs:
This is the SERP from SEMRush for ‘hoodies’
This is their listing in the SERP
Features-wise, the listing has one keyword in the title, which hasn’t been rewritten by Google; this is the complete title.
They have some decent Rich Snippets but nothing out of the ordinary. Shopify is good at providing structured data out of the box.
Landing Page:
This is their landing page
- Looking at the above the fold, we have little text.
- The description is low on text
- The H1 is simply ‘Hoodies’, and this matches the title
- When zooming out, the images get bigger and the font gets smaller.
- There are no indications of expertise in the industry, such as award-winning products
- There is no mention of how long they have existed
- There is little information about delivery
- No on-page information about being seen in ‘X’ publication
This is what we call a social first website. It has been built to sell without any consideration for optimising for traffic.
Intent: These are women’s hoodies, yet they are ranking well for a term which is non-gender specific.
Links:
So, it must be backlinks, right? This is SEO. We know Google can understand the technical elements of websites much better now. It can infer from context, and this is Shopify after all, so the structure of websites and the underlying code is well-known and understood by search engines.
Looking at the landing page itself, the SERP overview tells a story.
It’s such a new page in SEMRush index, it has no page authority score, no referring domains and obviously no backlinks. Yet it gets an estimated 20K traffic across 2.3k keywords. How?
Google ranks pages not domains (or so it says) but according to a Google leak it doesn’t. According to reports by SearchLogistics & Digitaloft, Google has a domain-wide quality score, referred to as “siteAuthority” in the leaked documents, which seems akin to a domain authority metric, something Google had previously downplayed in favour of page-specific PageRank. This ‘siteAuthority’ is thought to potentially apply sitewide when determining rankings. So we ignore Googles public statements and follow the logic which is to look at the domain as a whole.
Their link profile can be classed as being good but lacking maturity. What this means is that they have some good links from reputable domains i.e. The Sun Ireland, The Sun UK, OK, Glamour Magazine. Yet they don’t have many of them, they are all relatively new in the past year or so. Nothing like the link profiles of the traditional websites in the ‘hoodies’ SERPs.
S&S have a trend going down the lost backlink route as opposed to growth…
There is nothing suggesting that this websites link profile is strong enough to override the poor content and relevancy signals of the landing page.
Ultimately S&S are a great case study as there are very few variables that can explain how they achieved these SERP rankings.
Ranking Fluctuations née Google Dance?
It is wise at this stage to explore whether S&S have been the beneficiaries of the Google Dance. This may be an old term but it’s the best way to describe the random fluctuations we see in the SERPs.
Once again think about the data, we have a high-ranking page, with no positive SEO signals, no backlink history. Maybe the URL is new? If this is the case the new URL may have an older, more authoritative URL pointing to it.
Looking at SEMRush page data and archive.org we have concluded that the hoodies page is not new.
Additionally, S&S haven’t just arrived on the scene, we can trace their ranking on page 1 for ‘hoodies’ going back to October 2024. They are pretty well established, organically speaking. There is no dance going on.
How are they achieving these results?
In my experience, Sisters and Seekers are what I would class as a new brand. I had not heard of them before seeing them in the SERPs. I’ve worked with many of the UK’s fashion retailers and have a detailed understanding of the UK fashion market from an SEO perspective. There are established retailers that have dominated the SERPs for years, ASOS, Boohoo, River Island, Pretty Little Thing, et al.
When I took this to some more qualified members of our social team and asked about S&S they said that this brand has blown up on social this year, particularly TikTok. That’s right you heard it SOCIAL! Shivers went down my spine. I’ve been banging on for long enough now that links are King. But alas…
Social Uplift = Brand Uplift = Non-brand Rankings
In algorithmic terms…
Social Uplift = brand searches = increased site quality score = non-brand authority
In a nutshell, it’s becoming clearer every data set I go into that social media success leads to an increase in people searching on Google for those brands, why this is the case, I do not know. TikTok has a shop, and on most other platforms, the links to the site are in bios. Nevertheless, more people are using Google to search for S&S at the same time as more people like and follow S&S on social media. There is a direct correlation.
Let’s hear what our Director of Social & influencers, Gina Deen, has to say about this…
“From a social perspective, this reinforces something I’ve long suspected: strong brand equity built on social platforms can do more than just drive engagement, it can directly influence how Google perceives and ranks a site. What’s particularly interesting to me is that social-first brands like Sisters & Seekers are organically cultivating brand affinity and advocacy at scale, which appears to translate into increased brand searches and ultimately, non-brand keyword visibility.
“The SEO world has always known that brand searches are a signal, but this feels like a shift in how those signals are being earned. Not through traditional PR or backlinks, but through community-driven, content-rich social growth. TikTok especially, is proving to be a powerful element of discovery and loyalty—people see the brand on social, search for it on Google, and convert on-site.
“For me, this calls for a tighter integration between social and SEO strategies. It’s no longer about purely link-building activity or technical audits, it’s about brand storytelling and how your audience interacts with you across the full funnel.”
Looking at the data
The story tells itself when we look at the data.
Here we see the trend line for S&S followers over time…
Here we see the search volume for the brand search ‘sisters and seekers’ over time…
Here we see the increase in visibility in Google for non-brand terms over time…
Finally notice the visibility of the hoody landing page over the same period of time…
What we see is fairly unequivocal, the trend lines follow the same patterns. There seems to be a direct correlation between social media success and SEO success.
It is not as if S&S sit in isolation either, social media success is having impact across other brands.
Adanola
Adanola are another new brand to me that our social team are fully versed in. They rank in SERPs which ordinarily they shouldn’t, just like S&S. According to Glimpse & AddSauce, Adanola has seen significant search interest growth and is mostly discussed on TikTok, aligning with its appeal to younger, fashion-forward audiences.
We can see their growth in followers ramp up in March 2024 by approximately 45%
The trend line for brand searches in Google Ads then follows the same pattern. There is a spike around Christmas, which isn’t matched by followers, but these are different data sets. Seeing an influx in search volume at Christmas is not unusual.
Brand
Ranking keywords in SEMRush that include ‘Adanola’ have increased along the same lines as followers and brand search volume.
Non-brand
All non-brand queries follow the trend lines showing that the influx of fans and interest improves visibility for commercial queries not relating to brand.
Hoodies
Much like S&S, Adanola also have seen success for hoody terms. Following the same trend lines. Proving that S&S do not sit in isolation.
SEO Prowess
There does seem to be a bit more in terms of optimisation of the Adanola site. As we see they are boosted by Shopify as the others are.
Their landing pages are optimised but there’s less on the page, like many of these brands they present a more boutique feel than the very content heavy fast fashion brands we are used to see dominate the SERPs.
Looking at this more scientifically, there is a correlation between social media success and Google’s idea of a brand. We all know Google loves brands, its algorithms often prioritise well-established brands due to factors like authority, trustworthiness, and user engagement signals. Users are more likely to interact with known brands, reinforcing their ranking, according to a study by iMark Infotech.
In addition. the recent Google leak (December 2024) uncovered by Candour suggests that Google applies a site quality score to a domain much like the PageRank/siteAuthority score. Yet it’s not based on links and looks at something else entirely, so how does it establish a site quality/authority score and what is it in the first place?
Traditionally, the way you go about quantifying what a brand is, using an algorithm was to count links and brand mentions. The more sites talking about a brand, the more that brand is.
Then there is brand search, the more a brand is searched for the more popular that brand is. Again, traditionally this would have been inflated by Paid Ads which would generate internet searches, especially on out-of-home advertising for example. This has always been easy to prove but what wasn’t easy to prove was if that activity had any bearing on organic performance, especially for non-brand keywords. There has always been too many variables at play to be able to evaluate if one metric has a part to play.
Yet, there’s a difference between organic social brand searches and artificial, paid-for brand searches, right? Or is there?
Paid Media’s Impact on SEO
Anecdotally, paid ads are artificially creating interest, the activity has no longevity and could end when the ads stop running. There is little perceived value. Yet those ads still need to create engagement and that engagement proves value.
Whereas social media success is generally created organically, audiences are made up of advocates. The brand searches made as a result are more meaningful.
It makes sense to place more emphasis on brand searches made by advocates than those made by clever one-off advertising. This isn’t to suggest paid media is valueless, more that it has less impact on organic performance than social media in this setting.
This is where site quality score comes in, Google uses the method to figure out how good a website is by looking at how people search for it. This aligns with information gleaned from SearchLogistics. The leaked Google documents mention a system called NavBoost, which appears to re-rank results based on user click logs, categorising clicks by quality (e.g., “Good Clicks,” “Squashed Clicks,” “Unicorn Clicks”). This suggests that user interaction with search results is a key metric for evaluating a page’s relevance and quality.
Site Quality Score (SQS) Explained
SQS is a Google registered patent which can be found here. It is not necessarily a ranking factor or used in the algorithm in the way it is described. We can make assumptions based on advanced reasoning and match those with what we see in data from Google leaks.
Here’s the simple version:
1. Count Direct Searches
They count how many times people search for the website directly, using its name.
2. Count Searches Leading to Clicks
They count how many times people search for something related to the website, see it in the search results, and then click on it.
3. Calculate a Score
Using these two numbers, they calculate a quality score for the website.
Basically, if a lot of people search for the website directly or find it through related searches and click on it, that means it’s probably a good website. This method uses people’s search behaviour to judge the quality of a website. The concept described here aligns with how click data and user engagement are increasingly understood to influence perceived site quality, as suggested by the leaked information about NavBoost. (SearchLogistics). This goes a long way to explain why these brands are doing well with poor SEO metrics.
We can further see the importance Google places on this type of behaviour when we address other aspects of the algorithm. For example, the predicted site quality score.
Predicted Site Quality Score Explained
Predicted site quality (another Google patent) occurs when websites are new and have no data to go against. In this method, they predict a score based on what they know about similar websites.
It works like this:
1. Look at Existing Websites
They analyse websites that already have quality scores.
2. Find Word Patterns
They figure out how often certain words and phrases appear on good and bad websites.
3. New Website, Same Words
When they see a new website, they check how often those same words and phrases are used.
4. Calculate a Score
Based on the word patterns, they calculate a score for the new website, predicting how good it is.
Simply put, it’s like a chef predicting the quality of a new dish based on the listed ingredients and their prominence in the recipe. This helps Google quickly figure out if a website is potentially high-quality or not. In terms of what we know from the other leaks, while the documents don’t explicitly detail “predicted site quality score” in this exact way, they do indicate that for new pages, the PageRank score of a website’s homepage might be used until those new pages accrue their own PageRank (SearchLogistics). This implies a foundational score based on existing domain attributes. If this is done for link-level metrics, then it stands to reason it is also done for navigational-level metrics such as brand searches.
If so much effort is placed on predicting site quality and calculating site quality score, then it shows how links are no longer the be all and end all when it comes to site authority as a whole.
It further explains how brands like S&S are doing so well despite the significant lack of SEO prowess.
Going back to the topic of paid media, though we do know that SEO and PPC in combo have a positive effect on a business’ bottom line. Alongside other fantastic benefits of a full funnel solution and dedicated digital strategies. What I haven’t seen before is a website perform well organically with less than stellar SEO signals just because they spent a lot on advertising and generated a lot of brand search. Moreover, most websites that can spend tens of thousands per month on paid ads tend to have an SEO team working on organic traffic at a fraction of the cost. This muddies the waters and makes it hard to draw conclusions.
This then makes sense when you think about what Google is trying to achieve with site quality scores. Both channels (paid & social) increase brand search, so we know it’s not an increase in brand search that causes non-brand rankings to increase. Instead, we could look at the brand search being a by-product of having a good brand, and a good brand will have a better site quality score.
Take the second element of the site quality score patent, ‘count searches leading to clicks’ where they measure how many times people search for something related to the website. It’s not clear if this takes a blended phrase into account, such as [brand]+[keyword], i.e. ‘Sisters and seekers hoodies’ or if they are just looking at ‘hoodies’. Though, based on the intricacies of the algorithm, it would be hard to argue a [brand]+[keyword] search is not counted as a measure of quality.
Social media followers are more likely to search for [brand]+[keyword] as they are exposed to a multitude of product images, product types and more. When a follower finds time to do some internet shopping, it’s likely based on user behaviours that they would search for something like ‘sisters and seekers hoodies’ rather than go onto a social app, find the brand and click a link.
Whereas, when it comes to paid traffic, a user may see an ad on television, search or programmatically, usually these ads are structured to create awareness or to match intent and create engagement. If a user has an intent such as to buy ‘red shoes’, the retailers pick these up with text ads, the user goes directly to a relevant landing page, and there is never a need unless researching to go back to search for [brand]+[red shoes]. Awareness ads prove effective at generating the opposite, which is a search based on the brand name alone, to further their research. We see this activity a lot in Google Search Console data. Trends for brand searches follow advertising spends, when clients spend less, we see less brand search.
Therefore, your social media output can create a positive site quality score that is different to that of the paid output. Ultimately, it can help you rank well in Google for non-brand terms. More significantly, what we have seen here is a case whereby extremely good performance in social has an impact that can outweigh years of investment in traditional SEO performance metrics. Pushing well-established brands out of the way with a rather poor SEO game behind it.
So What?
This isn’t specifically what people are calling ‘social search’ where people are using social media to search. We are talking about people using social media as an inspiration to search, and the effect that has on the algorithm.
This has rather large consequences for fashion SEO at the very least and the whole ecosystem around that. The online fashion world is rather dynamic at the moment. Its relatively easy to create a brand and scale up production with a small team than it is in other industries. This creates a more competitive space in an industry where trends, culture and popularity rule.
People in fashion SEO won’t get by, by simply updating content, building internal links and having links from mainstream press coming in any longer. A more strategic, dynamic approach laddering SEO into the wider brand creative and activity, such as PR (not relying on DPR alone) and social media teams, is likely the only path.
There’s a reason we are seeing more trending brands from social media hitting the SERPs hard. It’s easier for them to create their own e-commerce than ever before (Shopify); they create a buzz, and as brands themselves, they become trending. I discussed this with some friends who worked with me across the fashion niche back in the day. They made a salient point about the old timers like Boohoo, ASOS and the like. People won’t lose affinity with the older brands that provide fast fashion, they are still convenient and cheap. They will always be there. Whereas, the newer brands that sell their own clothing and rely on their ability to impact audiences will be forever at the whim of the trends and the next brand that comes along, capturing more attention.
In other words, the social brands are here to stay as a collective, aided by their brand signals. The best thing you can do if you are one of these brands is improve on your SEO prowess, don’t ignore what is often the biggest traffic-driving channel to websites. For the bigger retailers, you are still in competition with the other big retailers, PageRank alone will keep you there when the trending brands drop out and get replaced. Work with your social teams to create campaigns that grow brand search.
Want to chat further? Head over to my LinkedIn, I’m always up for a discussion.
References
AddSauce. Social Commerce Use Case – Adanola. https://www.addsauce.com/case-studies/adanola
Digitaloft. What Does the Latest Google Documentation Leak Mean for Digital PR? https://digitaloft.co.uk/what-does-google-documentation-leak-mean-for-digital-pr/
Glimpse. Adanola – Trending 51% (March 2025). https://meetglimpse.com/trend/adanola/
iMark Infotech Pvt. Ltd. Why Google Prefers Big Brands and Low-Quality Content? https://www.imarkinfotech.com/why-google-prefers-big-brands-and-low-quality-content/
SearchLogistics. 22 Things We Might Have Learned From The Google Search Leak. https://www.searchlogistics.com/learn/seo/algorithm/google-search-leak/
Site Quality Score, Candour Agency – https://searchengineland.com/google-exploit-scoring-classifications-449333