Blend - Fashion App

BLENDing AI and shopping – hyper-personalised fashion retail

Over the past 20 years the Fashion industry has been transformed by two macro-trends: the boom in eCommerce and the rise of social media. 

These shifts have led to the birth of “fast fashion,” a multiplication of trends and the democratisation of trend-setting. In a digital sphere where consumers have access to billions of fashion products at any moment and inspiration and influence coming at them from all sides, consumers are overwhelmed by the sheer amount of choice available and the constant “newness.” 

In this context, product curation, influencer recommendations and personalization are paramount. But despite personalisation having been a fashion industry buzzword for the past decade, no-one has cracked it… yet.  

By combining AI and human curation, BLEND – launching later this year – aims to help customers cut through the overwhelming amount of choice, with hyper-personalised fashion product recommendations and user-generated style videos.

Why do we need personalised fashion recommendations? 

The fashion industry, and fashion eCommerce more specifically, have boomed. Whilst in 2000 online sales accounted for just 3.2% of total apparel sales in the US, they now account for 23% of this $343 Bn market. Over 100 billion items of clothing are available for sale each year, and with eCommerce growing at a rate of 12% this number is only increasing. 

With the growth of social media and eCommerce enablers like Shopify, the barriers to entry in the fashion industry have dropped. As a result, there are constantly exciting new brands entering the market, making it difficult for consumers to discover and keep on top of new brands that might appeal to them. Furthermore, with tech-enabled companies like Shein churning out up to 6,000 new products every day, keeping up with the pace of “newness” is impossible.

With so much choice available to them consumers are overwhelmed and suffering from “decision fatigue.” From BLEND’s interviews with 100+ customers, the resounding feedback was that shopping online is “time consuming,” “annoying” and “sends [them] into a rabbit hole.” Whilst social media sites are the preferred fashion discovery platforms, with 79% of Gen Z’s online purchasing journeys starting here, these sites also add to the problem. The $3Bn fashion influencer marketing industry ensures that these sites are flooded with promotional fashion content, making it difficult for people to get a true sense for the style of an influencer that they follow, or to discover brands that don’t have the budget to run paid ads. 

No more trend gate-keepers

20 years ago fashion trends were set by designers and editors through bi-annual fashion shows and glossy magazine covers. These industry gate-keepers ordained the brand and look of the season. Today, these gates have been smashed open and anyone with a phone and social media account plays an active role in creating and evolving fashion trends.

According to The Business of Fashion (BoF) “social media — namely TikTok — has made anyone an agenda-setter, and today’s trends often start with the masses.” This has flipped the power dynamic from top-down to bottom-up, with brands and designers working in symbiotic relation with consumer, and their influencer intermediaries. The BoF sees this as “an evolution from people looking to magazines for style inspiration,” with the difference being that TikTok users can actually “participate in the creation” of trends, rather than just consuming stylist-sanctified looks. Even fashion criticism, once the dominion of a handful of veteran fashion journalists, has been cracked open with TikTok fashion commentators garnering followings in the hundreds of thousands for their trend analysis.

This shift towards a more democratic and collaborative trend-setting process calls for a new type of retail and fashion discovery; one that prioritises social-first discovery by connecting consumers with peers and influencers whose style they like and trust.

Trend fragmentation and digital “niche-ification”

With trends proliferating as fast as your mobile connection is able to transport them, the fashion industry has had to learn to produce, manufacture and advertise at warp-speed. And with this pace and connectivity trends have become increasingly fragmented and hard to pin down, with many conflicting trends co-existing at the same time. Take for example “the minimalist quiet luxury and maximalist Barbiecore” trends which “are dominating the conversation at the same time.”

These trend waves tend not to converge into mainstream fashion phenomena, instead rippling into a number of sub-trends, constantly re-worked and evolved by active consumer participants. This has parallels with trends in broader social media, which is increasingly moving towards niche, private communities uniting active community members around shared values and experiences, versus the broadcast-style social media of the early 2010s and first-generation influencers.

Facilitated by increasingly sophisticated algorithms and the expansion of available content and data, we have moved from “social media” to “recommendation media.” Uniform, chronological Facebook newsfeeds have been replaced by hyper-personalised, algorithmically-curated video exhibitions, enabling the propagation of increasingly differentiated and specific communities – and the styles and trends that go along with those.

A simple look around you on a busy street provides observational evidence for the co-existence of a multitude of trends and “style-tribes.” Take the humble jean as a case study: during the 2010s skinny jeans proliferated with Topshop’s Joni jeans reaching cult-status. In 2023 low-waist ultra-baggy jeans are seen just as frequently as straight leg or mom-fit jeans, and even skinny jeans are said to be having a comeback.

The choice paradox

In this dizzying whirl of concurrent trends, endless sources of inspiration, and a growth-fuelled obsession with “newness,” it’s no shock that consumers are overwhelmed, struggling to identify their personal style or locate products that help them realise their sartorial aspirations.

It’s time for a new, hyper-personalised fashion retail experience

Every shopper’s tastes are unique, and today’s segmentation-based targeting methods fail to adjust for individual preferences. Instead they rely on demographic generalisations and user-profiling guesswork to target potential new users.

Hyper-personalisation, and algorithmic curation offer the solution, helping customers identify products and style inspiration that match their unique style, budget and size, without having to do the hard work of filtering and browsing through thousands of irrelevant products.

Personalisation has been a buzzword in the fashion industry for years, but retailers struggle to deliver

Speaking to employees at some of the UK’s leading fashion retailers, it’s clear that data-driven personalisation is something that they struggle with. In industries like entertainment or music, consumers have got used to advanced levels of personalisation – take Spotify’s Discover Weekly, or TikTok’s dynamic content feed – but in the fashion industry, no one has cracked personalisation…yet. According to a survey of more than 100 brands and retailers, only 20% of them said that they customise product recommendations based on a customer’s purchase history, and even fewer are able to personalise according to more granular user behaviour data. 

According to BoF, adoption has been slow due to the complexity of implementing personalisation. For many brands, gathering and interpreting the customer data that they already have on hand is enough of a challenge, making website personalisation a far-flung fantasy. Furthermore, retailers’ tech teams aren’t always trained or specialised in machine learning, and the ‘necessary tech talent can be hard to recruit.’ 

One of the key challenges when it comes to personalisation is tracking, storing and leveraging relevant user data, but when it comes to accessing first-party user data retailers often fall at the first hurdle. Something as simple as a user choosing not to log-in to the retail site makes it challenging to track that user’s session. And with no log-in requirements, the vast majority of a user’s browsing data is lost or unable to be attributed to create a complete customer profile.

Even where a user does log in or is able to be tracked via a “soft login,” minimal data points are actually leveraged to inform product merchandising and recommendations. The more advanced retailers that we spoke to were able to personalise according to past search history and historic purchases. However, user interfaces are optimised for passive browsing, meaning that users leave very few “bread crumbs” to help inform a detailed picture of an individual’s style, preferences and purchasing behaviour.

The less sophisticated of the retailers had no on-site personalisation at all, relying on user data for customer segmentation to inform a degree of segmentation-based tailoring for their email marketing campaigns.

Moreover, with increasingly stringent data-privacy regulations and the end of cookies it’s becoming more and more difficult to access and leverage third-party data, making brands and retailers reliant on their own first and zero-party data sources, and their in-house data capabilities.

Imagine if TikTok and a fashion retailer had a baby – that would be BLEND

Powered by AI and human curation, BLEND aims to be the most personalised, social and entertaining way to shop online. Blending social-media and shopping, BLEND is able to hyper-personalise fashion recommendations à la TikTok to allow users to discover and shop products all from within the same platform. 

Aggregating products from independent designers, second-hand platforms and luxury retailers, BLEND provides a hyper-personalised curation of product recommendations and style videos to help users cut through the overwhelming amount of choice when shopping for fashion online. The interactive and social-first experience, as well as our sequential product layout and fashion-specific data-set allows us to generate a rich understanding of each users’ style, purchasing behaviour and sources of fashion inspiration. Best of all, this doesn’t require us to “know” or leverage any personal or sensitive user data – we don’t need to know about a user’s demographic, location or political views to help them find fashion that is right for them.

Social product discovery

79% of Gen Z’s online purchases originate on social media. Magazines are no longer the style bibles they once were, and the traditional trend-gatekeepers have been overthrown. Instead influencers, creators and consumers are setting the sartorial rules via social platforms. On BLEND, users can explore shoppable influencer style videos to see styling inspiration and authentic product recommendations and reviews. Unlike legacy social platforms, having a huge follower-count isn’t what matters. Any user can become a trend-setter or reach their “style tribe,” no matter how niche, due to our style-based algorithms. With a built-in revenue share scheme, creators are rewarded for their content and curations.

Under the hood – AI-powered hyper-personalisation

BLEND’s hyper-personalised style recommendations are powered by our proprietary data-set and the latest advances in AI. As you interact with the app, BLEND will get better and better at understanding a user’s personal style to help them find items that match their personal style, size and budget. Everyone’s style is dynamic, and so is BLEND. Leveraging sequence to sequence algorithms (part of the tech behind ChatGPT) BLEND is able to adjust in-session to stay up to trend, and to reflect the latest changes in a user’s preferences.

A new business model for fashion retail

BLEND’s hyper-personalisation enables a new business model that isn’t restricted by the same inherent challenge as traditional retailers: with one online retail experience looking much like another, the only thing that differentiates retailers is their product selection and their resultant target consumer. The ability to continue to appeal to these target consumers relies upon their audience continuing to view that retailer as a destination that is well suited to them; where they can reliably, and quickly, find a product that will match their style, size and budget. This poses a business challenge: stay too small and niche, and you are unable to scale or reach profitability. Scale too much, and you lose the specific product curation and therefore the customer appeal that differentiates you.

BLEND is able to side-step this problem, all whilst having the largest inventory of any fashion retailer and appealing to the largest market. BLEND’s hyper-personalised fashion means that any one user will only see the most relevant 1% of our product portfolio, allowing BLEND to perfectly fit the niche and unique taste of an individual whilst still maintaining mass-market appeal. The dream? That BLEND becomes the front-door for every online shopping experience, helping users find the best products to express their unique style. 

Hyper-personalisation en-masse will be the unlock for profitable and scalable fashion retail. 

Sign up here to receive hyper-personalised fashion recommendations on BLEND.

Shopping Cart (0)