SEO without Compromise: How a Branded Clothing Online Store Achieved Stable Growth
Niche eCommerce in the fashion industry is a combination of strong brand competition, seasonal demand, and high user expectations. In such an environment, SEO either works systematically or fails.
This case study focuses on ostriv.ua, a website that approached Livepage after a full redesign and architectural restructuring, and which already had experience with another SEO team. Formally, the website was indexed and received traffic. However, behind these numbers were numerous technical issues that were limiting its potential for further growth.
OSTRIV company logo and banner
Our task was not to achieve quick performance spikes but to understand why the system was working inconsistently and rebuild it so the website could scale without the constant risk of traffic drops after every new Google update. This is a case about systematic work, in which SEO became the foundation for stable, predictable growth of the OSTRIV online store.
Stylish eCommerce “behind the scenes” and why no mistake is allowed in SEO
On the outside, fashion websites often look like neat catalogs with strong visuals, brand names, and attractive product listings with prices. However, on the inside, these projects are far more complex than they might seem at first glance.
In this niche, there is no real “margin for error.” If a website is technically unstable or its structure is chaotic, Google quickly shifts attention to competitors whose websites are easier to understand. And that happens regardless of brand strength or product assortment.
Brand competition, seasonality, and dependence on structure
The fashion market isn’t just competition between stores but also competition between brands on the same website. Every brand or category page competes with dozens of similar pages, both within the website’s own assortment and across other eCommerce platforms. At the same time, demand in this niche is highly seasonal. Some brands and categories peak in fall and winter, others in spring or summer. Certain products may temporarily disappear due to logistics issues or limited availability.
For businesses like this, SEO isn’t just another user acquisition channel but part of the overall business architecture. It determines which pages receive traffic, which brands scale, and which ones remain practically “invisible” in search results. Simple optimization isn’t enough here. The whole system needs to work in sync, from server responses to the logic of internal linking.
Overall click dynamics across the entire cooperation period in Google Search Console
We approached the website not as a collection of pages but as a single system, where any mistake in the technical setup, structure, or indexation can directly affect visibility and, ultimately, the client’s revenue.
Starting the cooperation: from SEO audit to full-fledged promotion
The beginning of our cooperation with the project wasn’t a typical entry point into SEO promotion. At first, the client approached us with a request for a full technical audit of the website without specific expectations regarding traffic or ranking growth. The goal was much more practical: to understand the website’s actual condition after the redesign and previous SEO work and to determine whether there was a solid foundation for further growth or if the website required a major rebuild.
It’s also worth noting that at the time we started, OSTRIV was already working with Livepage on the PPC side. Our systematic approach to data, hypothesis testing, and transparent communication was the main reason the client decided to involve us in SEO as well.
Why the SEO approach had to change
The client’s previous experience with SEO didn’t give a sense of control over the situation. Despite ongoing work, the website wasn’t showing stable growth, and any structural changes kept creating new problems. SEO existed almost in isolation from the website’s actual condition and the business processes behind it.
The client needed more than just a list of recommendations. They wanted a clear understanding of:
- What was actually happening with the website?
- Which decisions were harming page visibility?
- Which changes could become an effective channel for attracting traffic and generating new conversions?
That’s why the audit, designed to reveal the real picture without any sugarcoating, became the starting point for deeper, systematic work.
First signs of systemic problems
Even during the initial analysis, it became clear that the website’s issues were systemic. Many of them were the result of earlier internal changes — redesigns, migrations, URL routing updates, and various SEO experiments that had never been fully completed.
We discovered a large number of technical “leftovers” that had accumulated over time:
- Mass redirects within internal linking;
- Incorrectly configured canonical tags;
- Duplicate pages accessible through different URLs;
- Issues with indexation and the sitemap.
Formally, the website was functioning and returning proper 200 OK responses for pages, but for search engines, it was overloaded with conflicting signals.
These findings became a turning point. It was clear that without full technical stabilization, it would make little sense to move on to scaling or optimizing brand and category pages. As a result, the focus of the cooperation shifted from a one-time audit to a full SEO promotion strategy.
Technical “shock”: unpleasant aftermath of redesign and previous SEO mistakes
After conducting a full SEO audit, it became clear that the website had been evolving for a long time without a single logical framework. Structural changes, URL updates, and SEO adjustments made by different teams over time had piled up, but none fully solved the underlying problems. As a result, the website looked modern and polished for users but confusing and complicated for Google.
The core issue wasn’t just individual mistakes. It was the scale of them. Technical flaws were present across multiple levels of the website, from product pages and the blog to brand hub pages and even the sitemap itself.
Mass redirects, duplicates, and indexation issues
The website had become almost completely dependent on redirects. During earlier routing changes, some brand pages had their URLs updated, but internal links, canonical tags, and the sitemap still pointed to the old addresses. As a result, many pages existed as chains of 301/302 redirects, even in cases where they could have returned a 200 OK response.
Redirect chains after the website migration
Duplicate pages further complicated the situation. Some of them appeared after the website migration, while others were created due to incorrectly functioning legacy filters. In particular, pages with the /filter/ path were fully duplicating existing brand or category pages while still being indexed as independent URLs.
Traffic dynamics for pages with the /filter/ path in Google Search Console
For Google, this looked like standard competition between pages from the same website. But in reality, the search engine was forced to decide independently which page should appear in the search results.
Example of resolving duplicate /filter/ pages
Under these conditions, even pages with properly optimized semantic content were losing their rankings or dropping out of the index because technically incorrect but still “live” duplicates existed alongside them.
Seasonal products and the 304 response code solution
Seasonality creates a separate challenge in this niche because part of the product assortment can go out of stock for several months. This raises a key question: what should be done with pages that already have traffic statistics, an established backlink profile, and rankings in search results?
The typical solution in such cases is to remove those pages or return 404/410 response codes. But for this project, that approach would have meant losing the accumulated SEO potential and risking major ranking difficulties when the next season arrived.
That’s why we proposed a less typical but technically justified solution. At the server level, the pages would return a 304 response code for search engine bots, while regular users would still receive a standard 200 OK response.
Scanning product pages with 304 (Not Modified) status
This approach made it possible to preserve seasonal product pages without losing indexation or disrupting internal linking, while also avoiding signals to search engines that large portions of content were suddenly disappearing from the website.
When 200 OK status codes cause more harm than 404
In this particular project, one of the biggest issues was actually the excessive number of pages returning a 200 OK status. The reason was simple: many of those pages had no real value. They were technical leftovers from an old structure, duplicated filter pages, or URLs generated incorrectly.
For Google, however, these pages looked like unique pieces of content that needed to be crawled, indexed, and potentially ranked. As a result, the search engine was spending its crawl budget on technical noise instead of focusing on the pages that actually mattered. This led to unstable indexation.
Example of a typical error in product filtering
One of the key decisions was:
- Reducing unnecessary 200 OK pages;
- Relevant redirects;
- Closing some pages from indexation or removing them entirely;
- Clearing internal links.
Only after these changes did the website start to look like a logical, understandable system from Google’s perspective.
Recovery and stabilization after systemic technical optimization
Once the key technical issues were identified and prioritized, the next step wasn’t just fixing them but bringing the website back to a stable technical state. Our goal was to make sure every page had a clear purpose and status. Only within such a system can SEO work consistently instead of producing short-term spikes.
This stage required patience. Some solutions didn’t produce immediate results, but they laid the foundation for future growth. We deliberately moved from the most critical issues to smaller, less visible details that still played an important role in the overall quality of indexation.
The first priority was cleaning up redirects. We gradually removed unnecessary 301 and 302 redirects from internal linking and replaced them with direct URLs returning 200 OK wherever possible. Redirect chains that had formed after earlier structural changes were also carefully addressed, since they were significantly slowing down how search engine bots crawled the website. At the same time, we reviewed the logic behind canonical tags. Some pages were pointing to URLs that were redirects or duplicates, which created additional confusion for Google.
List of technical tasks provided to the client after the SEO audit
Another major block of work involved optimizing the sitemap. We removed pages with 3xx and 4xx status codes, as well as service URLs that provided no additional value for search engines. As a result, the sitemap started performing its primary function again — helping search engines quickly discover and reindex the pages that truly matter for the business.
We also optimized large images, fixed incorrect resource requests, and gradually reduced the number of technical issues that were preventing full page rendering. We paid special attention to the clarity of the HTML code. Excess scripts and styles were not only slowing the website down but also making it harder for search engines to analyze the content.
These improvements aren’t always visible from the outside, but they played a major role in stabilizing performance. Search engine bots began crawling the website more efficiently, and pages started returning to the index faster after updates.
Semantics and content: focusing on brand and category pages
Once the website was technically stabilized, it finally became possible to start full-scale work on semantics and content. Before that, any content optimization had a limited impact. Even well-prepared pages could fail to rank due to duplicates, redirects, or incorrect indexation.
We began by reviewing the semantic core that had previously been collected by the former SEO agency. Instead of rebuilding it from scratch, we refined it while taking the existing site structure into account. Some keywords had to be reconsidered, not because they were wrong, but because the website structure didn’t allow them to fully perform. At this stage, it became clear that the main growth potential lay in brand and category pages, rather than in smaller informational sections.
Brand pages as the main growth driver
Brand pages became the logical starting point. First, these pages already had a clear demand — users were directly searching for specific brands. Second, they offered the strongest conversion potential, combining commercial intent with existing trust in the brand.
Optimization started with the most popular brands — CP Company, Premiata, Barbour, and others. For each of these pages, we improved the structure, meta tags, and content. Relevant semantics were collected for every page, after which we prepared detailed SEO briefs for optimization. The texts written by our team, together with the client’s copywriters, were not created just to add “more content,” but to fully and clearly cover the user’s search intent from the perspective of search engine algorithms.
Example of an SEO text and optimized meta tags for a brand page
It was also important that we immediately took seasonality and product availability into account. Before optimizing each brand page, we discussed a simple but critical question with the client: Will the product assortment continue to grow, and does it make sense to scale this page in search results? This approach helped us avoid situations where SEO “pushes” a page higher in search rankings while the page itself cannot deliver business results in the near term.
After optimizing the brand pages, we saw the first movement into the top 10 search results, and growth appeared not only for branded queries but also for related non-branded searches.
Growth in visibility for non-branded queries in Google Search Console
Scaling to less obvious pages
Once the key brand pages were stabilized, the next step was to scale the semantic coverage into less obvious but still promising pages. These included either less popular brands or categories that had previously not been considered a priority from a traffic perspective.
By this stage, we already had a clear understanding of which types of pages performed well in Google within this niche and which ones required structural improvements or stronger internal linking.
At the same time, we began actively optimizing category pages for different genders and seasons, taking into account the real interests of the audience at specific times of the year. As a result, the website gradually stopped relying on just a few major pages. Search traffic began to distribute more evenly across the website, reducing the risk of seasonal drops and improving overall conversion performance.
Growing non-branded queries for category pages in Google Search Console
Improving usability and business interaction with users
Alongside technical stabilization and content optimization, we began paying closer attention to how users actually interacted with the website in real scenarios. You should understand that even a perfectly optimized SEO page won’t deliver results if users struggle to navigate it or take the next step toward a purchase.
The website had already gone through a redesign, but some UI/UX decisions were still unfinished or didn’t fully account for the specifics of this niche. We focused on several key interaction points: product pages, brand pages, and category pages. These are exactly the places where users decide whether to stay on the website, move further into the catalog, or leave for a competitor.
Current product page layout and identification of typical issues
We provided clear recommendations on how product photos should be presented, how color and size variations should be displayed, and how clothing items could be combined into different outfit ideas. We also paid special attention to internal linking blocks. They shouldn’t simply fill “empty space” on a page but genuinely help users discover related products or navigate to relevant brands and categories without unnecessary clicks.
Example of implementing product variations (colors/models)
Together with the client’s marketing team, we also tested several changes within product listings, from the overall look of product mini-cards to the placement of key elements in navigation. Some adjustments focused on small details, such as the placement of the language switcher or sorting controls. However, these details often have a noticeable impact on the overall website usability.
Suggested design for a new product mini-card in listings
AI as a new growth channel: experiments, risks, and results
Alongside traditional SEO, we deliberately started looking at the bigger picture — how user behavior is changing and how new AI systems are influencing the path to purchase. By mid-2025, it had already become clear that search was no longer limited to Google or other traditional search engines. Users increasingly turn to LLM-based systems for recommendations, curated lists, and product advice. As a result, brands are beginning to appear directly in AI-generated responses either as sources or as recommended options.
We fully understood the risks of this approach. There were no clear rules yet, analytics were limited, and there were very few real case studies on the market. But that uncertainty also made the direction promising. For this project, we decided not to wait for the “perfect moment.” Instead, we started testing this new growth channel alongside our core SEO, laying the foundation for future traffic and conversions.
llms.txt file and preparing the website for AI citations
The first step was technically preparing the website to interact correctly with LLM systems. We implemented an llms.txt file, which allows website owners to clearly define rules for how AI models can access and use the website’s content. In fact, it’s similar to robots.txt, but adapted to a new reality where content is not only crawled by search engine bots but also by large language models.
Structured llms.txt file designed for LLM citation
At this stage, it was important for us not just to “add a file” but to carefully determine which pages would provide the most value for AI citations. These included brand hubs, category pages, and key informational sections. Those pages were intended to become the foundation for brand mentions in AI-generated answers, without pushing users toward competitors or third-party aggregators.
Soon after implementing llms.txt, we noticed an increase in AI citations, brand visibility in responses from different LLM systems, and, most importantly, the first visits to the website directly from this channel.
Commercial LLM button: testing a new feature from scratch
The next, more ambitious step was developing and testing a custom commercial LLM button directly on the website. This was an experimental feature that we built entirely from scratch, without ready-made solutions or proven templates.
The main purpose of the button was to help users make decisions faster while browsing the website. It allowed them to quickly:
- Choose a product;
- Understand the differences between brands;
- Receive recommendations based on the available assortment.
At the same time, for the business, this feature created an opportunity to “capture” the user before they leave the website to search for answers in external AI systems, where control over the output is significantly lower.
Interactive commercial AI button on a product page
Testing prompts were the most challenging stage. Early versions of responses either referenced competitors too frequently or failed to present a clear value proposition. We gradually adjusted the prompt logic, focusing more on the commercial side: relevant products, clear price positioning, minimal filler text, and maximum practical value for the buyer.
Within a few months after launch, we saw the first results: noticeable growth in traffic from AI-driven channels, the first conversions, and an increase in brand mentions in LLM responses. For the client, this confirmed that investing in experiments can deliver real business value. For us, it became yet another sign that SEO is gradually expanding beyond traditional search optimization.
Working with an external link profile and improving domain authority
Work on the backlink profile for this website turned out to be just as complex as the technical optimization. The reason is simple: the website has a long history and has been actively developing since 2011. Over time, this resulted in a fairly large external link profile that looked substantial at first glance, but revealed multiple issues during deeper analysis.
In this niche, it’s not just about having relevant links but also about the websites they come from, how the brand is mentioned, and whether those references create misleading signals for Google. Additionally, mentions on external platforms often have the potential to turn users into actual leads, which makes the quality of these placements even more important.
Cleaning up “historical” referral domain spam
The first step was a deep audit of the backlink profile. We discovered a large number of outdated, toxic, and spammy domains, some of which dated back as far as 2011. Another major issue was mass hotlinking spam — external websites were using the project’s images, creating hundreds of unwanted referral links in the process.
Dynamics of domain rating and referring domains according to Ahrefs
The goal here wasn’t to “cut everything off at once,” but to proceed carefully to avoid triggering Google filters or penalties during algorithm fluctuations. The cleanup process was divided into several stages. Toxic domains were gradually disavowed, and additional measures were implemented to block further spam activity. One of those measures was hotlinking protection, which returns 403/404 responses when external domains attempt to load images directly from the website.
The results of this work weren’t immediate. However, over time, we observed a steady decrease in the number of “lost” links from spam sources and a much cleaner overall picture of the backlink profile.
Lost spam anchor links according to Ahrefs
Quality link building and domain rating growth
Alongside the cleanup process, we launched a high-quality link-building strategy. The focus was not on the quantity of links, but on their relevance and contextual value. We worked with sponsored articles prepared around specific search queries and closely aligned with the niche. At the same time, we took into account the criteria Google uses to evaluate the expertise and topical relevance of such content.
Example of controlled link building by the Livepage team
In some cases, the results exceeded expectations. Certain sponsored articles started ranking for the same keywords as the website’s brand or category pages. Instead of competing with them, these pages actually strengthened the website’s presence in search results.
Overall, this had a noticeable effect — an increase in domain authority. The website’s domain rating grew from 37 to 42, which is a meaningful signal of improved trust from search engines in a highly competitive niche. At the same time, we continue to monitor the backlink profile regularly, since spam attacks or unnatural links can appear even without any active actions on our side.
Growth of the website’s domain rating and number of AI citations across different LLM systems
Project results and further cooperation
Despite a difficult start, numerous technical limitations, and the consequences of earlier implementations, the project gradually moved into a stable growth phase. What’s important is that this dynamic wasn’t driven by short-term seasonal spikes. Instead, it became the result of systematic work across all layers of the website: technical optimization, website structure, content, usability, backlink profile, and additional traffic channels.
At the same time, it’s worth noting that we did not observe perfectly linear traffic growth throughout the cooperation period. This is largely because many structural changes on the website, including redirects and URL logic, had already been implemented by the previous SEO team or were part of a previous website migration. These factors technically complicated the process of restoring the previous traffic levels.
Additional factors also influenced the dynamics. The launch of AI Overview and the overall rise in AI usage reduced website traffic in the blog segment (which was not a primary focus of the SEO strategy). Also, shifts in demand occurred due to external circumstances in Ukraine, including renewed missile attacks and power outages.
Dynamics of clicks and impressions over the last 3 months in Google Search Console
The existing traffic became higher quality and more stable in terms of conversions. During the cooperation period, the website showed steady growth in organic page visibility and search rankings. The share of keywords in the top 10 increased from 21% to 49%, while the number of new conversions grew by 127%.
Yearly dynamics of conversions and organic traffic according to Google Analytics
Importantly, this positive trend remains visible even during challenging seasonal periods, when part of the product assortment becomes temporarily unavailable, and competition in search results traditionally intensifies.
Dynamics of keyword share in the top 10 and position changes throughout the year
Another notable development was the emergence of a new traffic acquisition channel:
- Traffic from AI platforms increased by 84%;
- Traffic from conversions grew by 167%.
After implementing llms.txt and the commercial LLM button, the website began consistently receiving mentions in AI-generated responses, along with referral visits and even the first sales coming from this previously untracked promotion channel.
Dynamics of traffic and conversions from AI channels after implementing llms.txt and the AI button according to Google Analytics
Work on the project is ongoing. The current focus is on further optimizing the filtering system, scaling semantic coverage for new brands and categories, continuous technical monitoring, and further development of AI-driven solutions. Some of these initiatives require time to implement, but this consistent approach helps prevent sharp traffic drops and supports long-term growth.
When technical stability, website structure, content, and search engine trust work together as a unified system, a website can continue to grow even in challenging conditions and without aggressive budget increases. If this case sounds familiar — a project with a complex legacy of previous decisions and the feeling that “SEO exists, but doesn’t really work” — the Livepage team knows how to approach such situations systematically. We don’t just fix isolated issues; we work with the entire business logic of a website, ensuring that SEO delivers real, measurable results.
